SilkPerformer® - Borland Technical Publications

User Guide
SilkPerformer ®
2008

Borland Software Corporation
8303 N. Mopac Expressway, Suite A-300
Austin, TX 78759-8374
http://www.borland.com
Borland Software Corporation may have patents and/or pending patent applications covering subject
matter in this document. Please refer to the product CD or the About dialog box for the list of applicable
patents. The furnishing of this document does not give you any license to these patents.
Copyright © 1992-2008 Borland Software Corporation and/or its subsidiaries. All Borland brand and
product names are trademarks or registered trademarks of Borland Software Corporation in the United
States and other countries. All other marks are the property of their respective owners.
March 2008
PDF

Contents
Introduction 1
Overview . . . . . . . . . . . . . . . . . . . . . 1
What's New in SilkPerformer 2008 . . . . . . . . 2
Upgrading to SilkPerformer 2008 . . . . . . . . . 4
Evaluation Version . . . . . . . . . . . . . . . . 4
Performance and Scalability Matrix . . . . . . . . 5
Memory Footprints by Application Type. . . . . . 7
Chapter 1
Getting Started 15
Overview . . . . . . . . . . . . . . . . . . . . 15
Outlining Projects . . . . . . . . . . . . . . . . 16
Modeling Test Scripts . . . . . . . . . . . . . . 21
Validating Test Scripts. . . . . . . . . . . . . . 22
Customizing Test Scripts . . . . . . . . . . . . 24
Establishing Baseline Performance . . . . . . . 25
Defining Workload. . . . . . . . . . . . . . . . 27
Executing Load Tests . . . . . . . . . . . . . . 28
Exploring Test Results . . . . . . . . . . . . . 29
Chapter 2
Configuring SilkPerformer 31
Overview . . . . . . . . . . . . . . . . . . . . 31
SilkPerformer Workbench . . . . . . . . . . . . 32
SilkPerformer Recorder . . . . . . . . . . . . . 39
Application Profiles . . . . . . . . . . . . . . . 41
Proxy Settings. . . . . . . . . . . . . . . . . . 45
Remote Agents . . . . . . . . . . . . . . . . . 48
Client Certificates . . . . . . . . . . . . . . . . 52
Java Configuration . . . . . . . . . . . . . . . 55
Chapter 3
Configuring Projects 59
Overview . . . . . . . . . . . . . . . . . . . . 59
Project Profile Settings . . . . . . . . . . . . . 60
Test Scripts . . . . . . . . . . . . . . . . . . .119
Data Files . . . . . . . . . . . . . . . . . . . .120
Agent Computers . . . . . . . . . . . . . . . .121
Workload Definition . . . . . . . . . . . . . . .121
Outlining Projects . . . . . . . . . . . . . . . .125
Chapter 4
Modeling Test Scripts 127
Overview . . . . . . . . . . . . . . . . . . . 127
Recording Test Scripts . . . . . . . . . . . . 128
Writing Test Scripts Manually . . . . . . . . . 131
Working From Sample Scripts. . . . . . . . . 131
Trying Out Test Scripts . . . . . . . . . . . . 132
Chapter 5
Validating Test Scripts 135
Overview . . . . . . . . . . . . . . . . . . . 135
Visual Analysis with TrueLog Explorer . . . . 135
Log Files. . . . . . . . . . . . . . . . . . . . 136
Report Files . . . . . . . . . . . . . . . . . . 137
Output Files . . . . . . . . . . . . . . . . . . 138
Chapter 6
Customizing Test Scripts 139
Overview . . . . . . . . . . . . . . . . . . . 139
Inserting Function & Transaction Declarations 141
Customizing Workload . . . . . . . . . . . . 142
Customizing Session Handling . . . . . . . . 149
Inserting Parsing Functions . . . . . . . . . . 153
Adding Verifications to Test Scripts . . . . . . 155
Manual Script Editing . . . . . . . . . . . . . 159
Using the System Configuration Manager. . . 160
Serializing and Synchronizing Virtual Users. . 160
Using Custom DLLs in Test Scripts . . . . . . 166
Matrix: Functions, the Recorder, and TrueLog
Explorer . . . . . . . . . . . . . . . . . . . 174
Chapter 7
Establishing Baseline
Performance 177
Overview . . . . . . . . . . . . . . . . . . . 177
Identifying Baseline Performance . . . . . . . 179
Confirming Test Baselines. . . . . . . . . . . 179
Setting Response Time Thresholds . . . . . . 182
Setting Up Server Monitoring . . . . . . . . . 183
SilkPerformer User Guide
iii

Chapter 8
Defining Workload 185
Overview . . . . . . . . . . . . . . . . . . . . 185
Setting Up Workload . . . . . . . . . . . . . . 187
Adjusting Session Times . . . . . . . . . . . . 191
Maintaining Transaction Rates . . . . . . . . . 191
Chapter 9
Executing Load Tests 193
Overview . . . . . . . . . . . . . . . . . . . . 193
Starting Load Tests . . . . . . . . . . . . . . . 194
Monitoring Load Tests . . . . . . . . . . . . . 194
Launching Load Tests Via the Command Line . 205
Chapter 10
Exploring Test Results 209
Overview . . . . . . . . . . . . . . . . . . . . 209
Performance Explorer . . . . . . . . . . . . . 212
Reports . . . . . . . . . . . . . . . . . . . . . 229
Measurements . . . . . . . . . . . . . . . . . 230
Chapter 11
Additional Workbench Features
245
Overview . . . . . . . . . . . . . . . . . . . . 245
Teamwork Features . . . . . . . . . . . . . . 246
Code Completion . . . . . . . . . . . . . . . . 249
Parameter Wizard . . . . . . . . . . . . . . . 252
Java Integration . . . . . . . . . . . . . . . . 256
Converting CSV Data for Use With Silk Performance
Explorer . . . . . . . . . . . . . . . . . . . . 259
Chapter 12
SilkPerformer Plug-Ins 263
Overview . . . . . . . . . . . . . . . . . . . . 263
Checking and Initializing Plug-Ins . . . . . . . 264
Message Output . . . . . . . . . . . . . . . . 264
Borland Optimizeit ServerTrace Plug-In . . . . 265
dynaTrace Diagnostics Plug-In. . . . . . . . . 267
Chapter 13
Component Testing (Java &
.NET) 269
Overview . . . . . . . . . . . . . . . . . . . . 269
.NET Support . . . . . . . . . . . . . . . . . . 269
Java Support . . . . . . . . . . . . . . . . . . 271
Chapter 14
Unit Test Integration (JUnit &
NUnit) 273
Overview . . . . . . . . . . . . . . . . . . . . 273
JUnit Integration . . . . . . . . . . . . . . . . 273
NUnit Integration . . . . . . . . . . . . . . . . 281
Chapter 15
SilkCentral Test Manager
Integration 287
Overview . . . . . . . . . . . . . . . . . . . . 287
Configuring Test Manager Integration . . . . . 288
Opening SilkPerformer Projects from Test Manager
289
Importing SilkPerformer Projects from Test Manager
291
Uploading SilkPerformer Projects to Test Manager
292
Uploading Test Results to Test Manager . . . . 293
Downloading Test Results from Test Manager . 294
Chapter 16
Source Control Integration 295
Overview . . . . . . . . . . . . . . . . . . . . 295
Configuring Source Control Integration . . . . 296
Placing Projects and Files Under Source Control298
Checking Files In . . . . . . . . . . . . . . . . 300
Checking Files Out . . . . . . . . . . . . . . . 301
Getting the Latest File Versions . . . . . . . . 302
. . . . . . . . . Checking In/Out Project Files302
Removing Source Control Linkage . . . . . . . 303
Synchronizing Source Control Status. . . . . . 303
Known Issues and Workarounds . . . . . . . . 304
Chapter 17
Other Supported Environments
307
Overview . . . . . . . . . . . . . . . . . . . . 307
Windows Vista . . . . . . . . . . . . . . . . . 308
Multibyte Support . . . . . . . . . . . . . . . . 308
AJAX Support for Web Application Testing . . . 311
SAP eCATT Support . . . . . . . . . . . . . . 312
Terminal Emulation Support . . . . . . . . . . 314
GUI-Level Testing Support . . . . . . . . . . . 317
SilkPerformer Plug-In for Eclipse . . . . . . . . 325
Remedy Support . . . . . . . . . . . . . . . . 333
Adobe Flex AMF3 Support . . . . . . . . . . . 337
iv
SilkPerformer User Guide

Customizing Oracle Applications 12i Scripts for
Replay . . . . . . . . . . . . . . . . . . . . .349
Chapter 18
Trouble Shooting 353
Overview . . . . . . . . . . . . . . . . . . . .353
Controller Agent Communication Troubleshooting .
354
Error Message Overview . . . . . . . . . . . .356
Index 359
SilkPerformer User Guide
v

vi
SilkPerformer User Guide

INTRODUCTION
Overview
Introduction
About this manual
This manual provides all the information you need to work effectively with
SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 1
What's New in SilkPerformer 2008 2
Upgrading to SilkPerformer 2008 4
Evaluation Version 4
Performance and Scalability Matrix 5
Memory Footprints by Application Type 7
Overview
SilkPerformer is the load testing tool that leads the industry today. It is used to
assess the performance of Internet servers, database servers, distributed
applications, and middleware—both before and after they are developed.
SilkPerformer helps you to quickly and cost-effectively produce reliable, highquality
application solutions.
SilkPerformer creates highly realistic and fully customizable load tests. It does
this through the use of virtual users that automatically submit transactions to
systems under test, in the same way that real users would. Using a minimum of
hardware resources, you can generate tests that simulate hundreds or thousands
of concurrent users. And you can use SilkPerformer's powerful reporting tools
both during and after load tests to analyze the performance of servers and locate
bottlenecks, so that you can maximize the potential of your system.
Questions SilkPerformer can help you answer:
• How many simultaneous users can my server support?
• What response times will my users experience during peak hours?
• Which hardware and software products do I need to ensure optimum
performance from my server?
• Which components are the bottlenecks in my system?
SilkPerformer User Guide 1

INTRODUCTION
What's New in SilkPerformer 2008
• What will the performance impact on my system be if I employ security
technology?
• Which areas of my application perform well and which have bottlenecks
(business transactions, objects, and operations can all be evaluated)?
• Which factors affect performance? What effects do they have? And at
what point will those factors impact service levels?
What's New in SilkPerformer 2008
Following are the significant enhancements and changes that were made to
SilkPerformer 2008.
Enhanced Multibyte Character Set Support
Oracle Forms Multibyte Support
The SilkPerformer Recorder is now able to convert the UTF-8 strings of Oracle
Forms to the system code page, and the replay engine converts them back to
UTF-8, thus enabling multibyte support for Oracle Forms applications. This
includes viewing and customizing multibyte scripts with Silk TrueLog Explorer.
Citrix Multibyte Support
Same as for Oracle Forms, SilkPerformer now supports record and replay of
multibyte strings with Citrix. This includes viewing and customizing multibyte
scripts with Silk TrueLog Explorer.
Note Optical character recognition (OCR) for multibyte characters
is currently not supported.
Multibyte Enabled Project Packages
Project packages containing filenames with multibyte characters can now be
uploaded to and downloaded from SilkCentral Test Manager. The allowed file
name length restriction of about 40 Japanese characters has also been removed.
Enhanced AJAX Support
SilkPerformer now recognizes AJAX requests that come in the form of XML or
JSON, both of which can now be pretty-formatted both in the script and in
TrueLog Explorer. This allows for a much better readability of such requests.
2 SilkPerformer User Guide

INTRODUCTION
What's New in SilkPerformer 2008
Additionally, AJAX requests can be visually customized and verified in
TrueLog Explorer.
See “AJAX Support for Web Application Testing” for full details.
FLEX/AMF3 Support
SilkPerformer now supports FLEX applications based on the new AMF3
protocol. The non-readable AMF3 data is converted to a readable format within
the script, which also enables visual customization and verification via TrueLog
Explorer.
See “Adobe Flex AMF3 Support” for full details.
Remedy Web ARS Support
SilkPerformer now supports Remedy Web ARS versions 6.3, 7.0, 7.0.1, and 7.1.
The proprietary data format of Remedy is translated into an XML-based
representation that makes scripts easier to understand and customizable.
Remedy support is licensed under the Premium Virtual User type.
See “Remedy Support” for full details.
Importing External Monitoring Data Into Silk
Performance Explorer
SilkPerformer offers a command line tool (csv2tsd.exe) that enables you to
convert CSV (Comma Separated Values) files into TSD (Time Series Data)
files. This is useful when you are monitoring external Windows data that, for
security reasons, can not be monitored by SilkPerformer. By converting CSV
data to TSD format, external data can be analyzed alongside internal
SilkPerformer monitoring data within Performance Explorer.
See “Converting CSV Data for Use With Silk Performance Explorer” for full
details.
Technology Updates
• SAPGUI: The SAPGUI for Windows 7.10 client is now supported.
• Oracle Applications 12: Oracle Applications 12, which is based on
Oracle Application Server 10g, is now supported.
See “Customizing Oracle Applications 12i Scripts for Replay” for full
details.
SilkPerformer User Guide 3

INTRODUCTION
Upgrading to SilkPerformer 2008
Upgrading to SilkPerformer 2008
Licensing:
SilkPerformer 2008 requires a SilkPerformer 2008 license.
Project files:
Project files can be upgraded from SilkPerformer versions 4.0 and higher (by
opening them in the current version). Downgrading from the current version to
older SilkPerformer versions is not supported (i.e., has not been tested).
Installation Information:
Parallel installations:
SilkPerformer 2008 (controller and agent) can be installed together with any
older version of SilkPerformer on the same machine; they do not influence each
other.
For complete installation instructions, please refer to the SilkPerformer
Installation Guide.
Evaluation:
During the installation process, you will install one of the following versions:
• Evaluation version: Installing an evaluation version of SilkPerformer
2008 grants you full product functionality for a period of 30 days, limited
to 10 virtual users. To upgrade to the full version at a later time, please
contact your Borland sales representative.
• Full version: Installing a full version of SilkPerformer 2008 requires a
license. Contact your Borland sales representative for further
information regarding licensing.
Evaluation Version
If you have installed the evaluation version of SilkPerformer 2008, you may use
the full product functionality for 30 days, limited to 10 virtual users. The
software will stop operating when the evaluation period expires.
Tip To ease you into the process of working with SilkPerformer
2008, it is recommended that you review the Web Load Testing
Tutorial. This PDF-based tutorial can be found in SilkPerformer's
documentation set, in the Tutorials folder.
To upgrade to the full version of SilkPerformer 2008, contact your Borland sales
representative.
4 SilkPerformer User Guide

INTRODUCTION
Performance and Scalability Matrix
Performance and Scalability Matrix
Max. # of Vusers per
agent
4xPIII, 1GHz, 4
GByte RAM
2xPIII, 1GHz, 2
GByte RAM
1xPIII, 1GHz, 1
GByte RAM
1xPIII, 1GHz, 512
MByte RAM
1xPIII, 600Mhz, 512
MByte RAM
1xPIII, 600Mhz, 256
MByte RAM
Web Business
App.
Table 1: Performance and Scalability Matrix
Web Business App.
with SSL(-35%)
Web low level
(+20%)
5900 3800 7100 4
3600 2300 4300 2
1900 1200 2300 1
800 500 1000 1
800 500 1000 1
300 200 400 1
Note Web Business App uses Page-based Browser-level API
(HTML/HTTP)
Note Web low level uses Browser-level API (HTTP)
Recommended
#NICs
Test Parameters
The numbers in the table above have been determined using a realistic workload
with the following characteristics:
• The page structure of all requested Web pages corresponds to popular
public Web pages.
• The average think-time between pages views for virtual users was 32
seconds.
• A single virtual user was emulated with four concurrent connections.
This means that the number of virtual user simulated corresponds to a realistic
number of concurrent users accessing a popular Web site, resulting in:
Avg. Hits/page: 39
Avg. page size: 130 kByte
Avg. think time between pages: 32 sec
SilkPerformer User Guide 5

INTRODUCTION
Performance and Scalability Matrix
Formula:
The formula for calculating the recommended maximum number of vusers per
agent computer, which is relevant for realistic workload using Web Business
App. = Page-based Browser-level API, is:
Given:
agentCpu: specifies processor speed in MHz (must be >= 100)
agentCpuCount: specifies the number of processors
agentRamMb: specifies the RAM of the agent in MByte. (has to be > 64)
cpuBound = 0;
if (agentCpu >= 100 && agentCpu < 200) cpuBound = 100;
if (agentCpu >= 200 && agentCpu < 300) cpuBound = 450;
if (agentCpu >= 300 && agentCpu < 400) cpuBound = 600;
if (agentCpu >= 400 && agentCpu < 500 cpuBound = 750;
if (agentCpu >= 500 && agentCpu < 600) cpuBound = 900;
if (agentCpu >= 600 && agentCpu < 800) cpuBound = 1200;
if (agentCpu >= 800 && agentCpu < 1000) cpuBound = 1500;
if (agentCpu >= 1000 && agentCpu < 1400) cpuBound = 2000;
if (agentCpu >= 1400) cpuBound = 2500;
ramBound = (agentRamMB-64) * 2.0;
agentCpuFactor = 1.0;
for (i = 2; i

INTRODUCTION
Memory Footprints by Application Type
1000 Mhz P III, 1 GByte
maxVuserPerAgent = (Minimum((1024 - 64) * 2.0, 2000)/100)*100 =
(Minimum(1920,2000)/100)*100 = 1900 Vusers
1000 Mhz P III, 512 MByte
maxVuserPerAgent = (Minimum((512 - 64) * 2.0, 2000)/100)*100 =
(Minimum(896,2000)/100)*100 = 800 Vusers
600 Mhz PIII, 512 MByte
maxVuserPerAgent = (Minimum((512 - 64) * 2.0, 1200)/100)*100 =
(Minimum(896,1200)/100)*100 = 800 Vusers
600 Mhz PIII, 256 MByte
maxVuserPerAgent = (Minimum((256 - 64) * 2.0, 1200)/100)*100 =
(Minimum(384,1200)/100)*100 = 300 Vusers
Memory Footprints by Application Type
The following table outlines the size of the memory footprint associated with
each of the application types supported by SilkPerformer.
Note The listed users are in-process, i.e., 25 users run in a single
process, and 50 users run in a single process, respectively.
Users PerfRun (KB) Per User (KB) Per User (MB)
Web Business Transaction (HTML/HTTP)
1 1,935 1,935 1.89
25 10,323 413 0.40
50 20,400 408 0.40
Web Low Level (HTTP)
1 2,044 2,044 2.00
25 12,136 485 0.47
50 22,732 455 0.44
Flash Remoting
1 1,912 1,912 1.87
25 12,776 511 0.50
50 20,896 418 0.41
SilkPerformer User Guide 7

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
WebDav (MS Outlook Web Access)
1 3,588 3,588 3.50
25 17,724 709 0.69
50 33,364 667 0.65
Remedy Ars Web
1 2,288 2,288 2.23
25 9,688 388 0.38
50 17,340 347 0.34
Email (SMTP/POP)
1 1,904 1,904 1.86
25 5,268 211 0.21
50 8,912 178 0.17
Directory Server (LDAP)
1 2,920 2,920 2.85
25 not tested N/A N/A
50 not tested N/A N/A
FTP
1 2,420 2,420 2.36
25 5,352 214 0.21
50 8,348 167 0.16
Streaming Media (MS, Real)
1 7,188 7,188 7.02
25 30,188 1,208 1.18
8 SilkPerformer User Guide

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
50 67,782 1,356 1.32
TCP/IP based Application
1 1,912 1,912 1.87
25 5,624 225 0.22
50 9,500 190 0.19
Mixed Protocols
1 N/A N/A N/A
25 N/A N/A N/A
50 N/A N/A N/A
TN3270
1 2,224 2,224 2.17
25 7,424 297 0.29
50 13,652 273 0.27
TN5250
1 1,495 1,495 1.46
25 not tested N/A N/A
50 not tested N/A N/A
VT100/VT200
1 2,069 2,069 2.02
25 7,168 287 0.28
50 10,664 213 0.21
SAPGUI
1 30,656 30,656 29.94
SilkPerformer User Guide 9

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
25 110,180 4,407 4.30
50 194,532 3,891 3.80
SAP NetWeaver (Web)
1 2,072 2,072 2.02
25 10,084 403 0.39
50 17,936 359 0.35
Peoplesoft 8
1 1,936 1,936 1.89
25 9,812 392 0.38
50 16,408 328 0.32
Oracle 11i
1 22,636 22,636 22.11
25 34,748 1,390 1.36
50 51,104 1,022 1.00
Siebel 7 Web Client (incl. IE Option Pack)
1 2,256 2,256 2.20
25 11,532 461 0.45
50 20,576 412 0.40
WAP
1 1,848 1,848 1.80
25 5,408 216 0.21
50 9,164 183 0.18
i-Mode
10 SilkPerformer User Guide

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
1 1,936 1,936 1.89
25 10,323 413 0.40
50 20,400 408 0.40
Radius
1 16,104 16,104 15.73
25 21,284 851 0.83
50 26,268 525 0.51
XML/SOAP (recording Web Service client)
1 2,676 2,676 2.61
25 8,460 338 0.33
50 16,436 329 0.32
.Net Framework using Visual Studio .Net Add-On
1 19,964 19,964 19.50
25 N/A N/A N/A
50 N/A N/A N/A
Java Framework
1 13,044 13,044 12.74
25 22,356 894 0.87
50 27,920 558 0.55
COM / DCOM / MTS / COM+ / ADO
1 4,608 4,608 4.50
25 28,160 1,126 1.10
50 51,200 1,024 1.00
SilkPerformer User Guide 11

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
IIOP – Corba / EJB (RMI over IIOP)
1 2,736 2,736 2.67
25 31,696 1,268 1.24
50 61,960 1,239 1.21
Jacada
1 15,556 15,556 15.19
25 24,020 961 0.94
50 31,552 631 0.62
Citrix Metaframe
1 3,900 3,900 3.81
25 9,000 360 0.35
50 11,500 230 0.22
Visual Basic
1 2,100 2,100 2.05
25 4,128 165 0.16
50 6,276 126 0.12
Oracle
1 1,732 1,732 1.69
25 3,696 148 0.14
50 5,792 116 0.11
ODBC
1 5,468 5,468 5.34
25 16,918 677 0.66
12 SilkPerformer User Guide

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
50 29,016 580 0.57
DB2 CLI
1 5,168 5,168 5.05
25 21,952 878 0.86
50 37,468 749 0.73
Tuxedo (all Feat.) + Clarify
1 2,400 2,400 2.34
25 not tested N/A N/A
50 not tested N/A N/A
SilkTest
1 13,000 13,000 12.7
25 N/A N/A N/A
50 N/A N/A N/A
Chordiant CRM
1 not tested N/A N/A
25 not tested N/A N/A
50 not tested N/A N/A
Siebel 6/DB2
1 not tested N/A N/A
25 not tested N/A N/A
50 not tested N/A N/A
Siebel 6/Oracle
1 9,428 9,428 9.21
SilkPerformer User Guide 13

INTRODUCTION
Memory Footprints by Application Type
Users PerfRun (KB) Per User (KB) Per User (MB)
25 66,540 2,662 2.60
50 125,720 2,514 2.46
Siebel 6/SQL Server
1 not tested N/A N/A
25 not tested N/A N/A
50 not tested N/A N/A
GUI-level testing
1 1,788 1,788 1.75
25 5,080 203 0.2
50 not tested not tested not tested
Note GUI-level tests require additional memory resources through
the usage of terminal services sessions and SilkTest:
- 6,300 KB per VUser for Windows Terminal Services
- 13,000 KB per VUser for SilkTest
- 1,092 KB per VUser for SilkPerformer’s Session Manager
14 SilkPerformer User Guide

1
Getting 1Chapter
Started
Introduction
What you will learn
This chapter explains how to get started working with SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 15
Outlining Projects 16
Modeling Test Scripts 21
Validating Test Scripts 22
Customizing Test Scripts 24
Establishing Baseline Performance 25
Defining Workload 27
Executing Load Tests 28
Exploring Test Results 29
Overview
To conduct a SilkPerformer load test, a series of well-defined steps should be
followed:
1 “Outlining Projects”: The basic settings for your load test project are
defined first.
2 “Modeling Test Scripts”: Next, the test script that will prescribe the
actions of the virtual users in your test is created.
SilkPerformer User Guide 15

1 GETTING STARTED
Outlining Projects
3 “Validating Test Scripts”: A trial run of the test script is performed. The
results of the trial run are inspected with the help of TrueLog Explorer to
make sure that the script works properly.
4 “Customizing Test Scripts”: The test is customized to create realistic
tests.
5 “Establishing Baseline Performance”: The baseline performance of the
application under test is identified in a test run with a single user per user
type. This will form the basis for the calculation of the number of
concurrent users and threshold settings for timers later during the full
load test. The results of the baseline test are inspected to confirm that the
baseline reflects the performance required from the tested application;
the results will be used to calculate the number of concurrent users and
appropriate threshold settings.
6 “Defining Workload”: Prior to the execution of a load test, the number of
virtual users per user type, the duration, and the number of agents
involved must be set.
7 “Executing Load Tests”: The full load test is run, using the customized
test and the full complement of agent computers and virtual users.
8 “Exploring Test Results”: Finally, the results of the load test are analyzed
to assess the performance of the target application.
Outlining Projects
The first step in conducting a SilkPerformer load test is to define the basic
settings for the load testing project. The project is given a name, and optionally a
brief description can be added. The type of application to be tested is specified
from a range of choices that includes all of the major traffic available today on
the Internet and on the Web, including the most important database and
distributed applications. See “Available application types” for details.
The settings that are specified are associated with a particular load testing
project. It is easy to switch between different projects, to edit projects, and to
save projects so that they can later be modified and reused.
A project contains all the resources needed to complete a load test. These
include a workload, one or more profiles and test scripts, all the data files that
are accessed from the script, and a specific number of agent computers and
information for server side monitoring. Options for all of these resources are
available directly from the project icon in the corresponding tree list.
Note For details on the memory footprint associated with each
supported application type, see “Memory Footprints by Application
Type”.
16 SilkPerformer User Guide

1 GETTING STARTED
Outlining Projects
Available application
types
• Web browser
- Web business transaction (HTML/HTTP)
- Web low level (HTTP)
- Flash Remoting
- WebDAV (MS Outlook Web Access)
• Internet
- Email (SMTP/POP)
- Directory server (LDAP)
- FTP
- Streaming media (MS, Real)
- TCP/IP based application
- Mixed Protocol
- WebDAV (MS Outlook Web Access)
• Terminal Emulation
- TN3270
- TN5250
- VT100
- VT200+
• ERP/CRM
- SAP
o SAPGUI
o SAP NetWeaver (Web)
- PeopleSoft
o PeopleSoft 8
- Chordiant CRM
- Clarify 8-10 (Tuxedo)
- Oracle
o Oracle Applications 11i
o Oracle Applications 12i
- Siebel
o Siebel 6/DB2
o Siebel 6/Oracle
o Siebel 6/SQL Server
SilkPerformer User Guide 17

1 GETTING STARTED
Outlining Projects
o Siebel 7 Web Client (incl. IE Option Pack)
- Remedy
o Remedy 6.3
o Remedy 7.01
o Remedy 7.0
o Remedy 7.1
• Wireless
- WAP
- i-Mode
- Radius
• Web Services
- XML/SOAP (recording Web Service client)
- .NET Explorer
- Java Explorer
• Database
- Oracle
- ODBC
- DB2 CLI
- ADO
• Application Server/Component Models
- Corba (IIOP)
- EJB (RMI over IIOP)
- EJB (Java Explorer)
- Tuxedo (ATMI)
- Tuxedo (JOLT)
- COM/DCOM/MTS/COM+/ADO
- .NET Remoting
- Jacada
- Oracle
o Oracle Forms 10g
o Oracle Forms 6i
o Oracle Forms 9i
• Terminal Services
18 SilkPerformer User Guide

1 GETTING STARTED
Outlining Projects
- Citrix MetaFrame
• .NET
- .NET Explorer
- .NET Framework using Visual Studio .NET Add-On
• Java
- Java RMI/EJB (recording)
- Java Explorer
- Java Framework
• Frameworks
- .NET Framework using Visual Studio .NET
- Java
- Visual Basic
• Monitoring
- BDL Monitor for Performance Explorer
• Unit Testing
- JUnit
- Java Testing
- NUnit
- .NET Testing
• GUI-Level Testing
- SilkTest
Sample scripts
Sample project files, including scripts, include files, data files, profile files,
workload definitions, and more, are available from the Application type window
on the Outline Project dialog when you are initially prompted to outline a new
project. These files can be used as templates that you can edit manually to meet
your needs, as explained in “Modeling Test Scripts”.
Sample scripts are available for the following application/component/protocol
types:
• ADO
• COM
• Internet
- Cookies
- email
- Form-based file upload
SilkPerformer User Guide 19

1 GETTING STARTED
Outlining Projects
Outlining a project
- FTP
- LDAP
- Mapi
• .NET
- .NET Remoting
- Web Services
• Java Framework
- JDBC
- RMI/IIOP
- RMI
• ShopIT V6.0
• XML
Procedure To define a basic outline for your project:
1 Click the Start here button on the SilkPerformer Workflow bar. If you
have already used this icon in your current project, the icon will be
named Outline Project. The Workflow - Outline Project dialog appears.
2 In the Project name field, enter a name for your project.
3 In the Project description field, enter a description for your project (this
description is for your own project management purposes only).
4 From the Application type tree list, choose the type of application that
you want to use in your load test. See “Available application types” for
details.
Note If you are load testing a Web application, choose the Web
business transaction (HTML/HTTP) option to create simpler scripts
while incorporating advanced functionality.
Choose the Web low level (HTTP) option if you want to put the
highest possible load on your application. These scripts are more
complex than Web business transaction scripts, which require more
effort to customize.
It is recommended that use the Web business transaction (HTML/
HTTP) instead of the Web low level (HTTP) scripts for browser
based applications.
5 Click OK.
An icon for the new project appears in the tree-view area of the main
SilkPerformer window.
20 SilkPerformer User Guide

1 GETTING STARTED
Modeling Test Scripts
6 If you need to add additional resources to the project, right-click the
project icon in the tree list. It is particularly important that all the user
data files (.csv), random data files (.rnd), and .idl files needed by
SilkPerformer are set up for the project.
Modeling Test Scripts
The second step in conducting a SilkPerformer load test is to create a test script
that will prescribe the actions of the simulated users who will be run during the
test. The script is written in SilkPerformer's proprietary scripting language, the
Benchmark Description Language (BDL). See the BDL Function Reference for
full details.
Scripts can be created in different ways depending on the application type. The
easiest way is to use the SilkPerformer Recorder, which automates much of the
process for you. Alternately, you can create a test script manually or edit a
sample script.
• Test script generation with the SilkPerformer Recorder: The standard
method for creating a test script is to use the SilkPerformer Recorder to
capture and record traffic that is representative of the type you need to
simulate in your test. The SilkPerformer Recorder automatically
generates a BDL test script based on the recorded traffic.
• Manual test script generation: A second method of creating a test script
is to manually create a new script from scratch in BDL.
• Sample script reuse: A variant on the second method of creating a test
script is to base the new script on one of the sample BDL scripts that are
provided with SilkPerformer. See “Sample scripts” for a list of available
sample scripts.
Modeling the test
script
Procedure To create a test script:
1 Click the Model Script button on the SilkPerformer Workflow bar.
The Workflow - Model Script dialog appears.
2 In the Script area, define how you want your script to be generated.
- Select the Record option to have the SilkPerformer Recorder capture
and record network traffic, and then automatically generate the
script.
From the Select application profile drop-down list, select the profile
of the application type you want to use. If a profile is not set up for
that application, click the button (...) to the right of the drop-down
list to set one up. See “Application Profiles” for further details.
SilkPerformer User Guide 21

1 GETTING STARTED
Validating Test Scripts
If you are load testing a Web application, you can insert the URL of
the application’s home page in the URL field. Otherwise, enter the
command line options that you want to pass to the client application.
See “Recording Test Scripts”.
- If you want to use one of the sample scripts that accompany
SilkPerformer, select the Sample script option in the Script area. See
“Working From Sample Scripts”.
- Experienced users may want to create their scripts manually. To do
so, select the New script option in the Script area. See “Writing Test
Scripts Manually”Click OK.
Validating Test Scripts
The third step in conducting a SilkPerformer load test is to perform a trial run of
the test script that you created in the previous step. The object of a trial run is to
ensure that your test script is free from errors, and that it accurately reproduces
the interaction needed between the client application and the server. Normally,
this is the traffic that was recorded by the SilkPerformer Recorder in the
previous step.
For a try-out of the test script, options are automatically selected so that you can
see a live display of the actual data as it is downloaded. Log files and a report
file are created so that you can confirm later that the script ran properly. Only
one user is run and the stress test option is enabled so that there is no think time
and no delay between transactions.
TrueLog Explorer can assist you in finding replay errors quickly, customizing
session handling, and adding verifications to your script.
For Try Script runs, the following options are automatically set to the specified
values (see “Replay options” for details):
• The Stress test option is on.
• The Stop virtual users after simulation time (Queuing Workload) option
is off.
• The Virtual user log files (.log) option is on.
• The Virtual user output files (.wrt) option is on.
• The Virtual user report files (.rpt) option is on.
• The Virtual user report on error files (.rpt) option is on.
• The TrueLog files (.xlg) option is on.
• The TrueLog On Error files (.xlg) option is off.
• The Compute time series data (.tsd) option is off.
22 SilkPerformer User Guide

1 GETTING STARTED
Validating Test Scripts
Trying out the test
script
• All logging detail options (Results/Logging and Results/Internet logging
tab) are on.
• The Enable all measure groups (TSD measure groups) option is off.
• The Bandwidth option is set to High Speed (unlimited).
• The Downstream option is set to Unlimited.
• The Upstream option is set to Unlimited.
• The Duplex option is off.
• The Modem compression rate is set to 1.0 x.
Procedure To try out your test script:
1 Click the Try Script button on the SilkPerformer Workflow bar.
The Workflow – Try Script dialog appears.
2 In the Profile drop-down list, the currently active profile is selected (this
is the default profile if you have not configured another).
3 If you want to configure the simulation settings for the selected profile,
click the button (...) to the right of the drop-down list. See “Replay
options” for details.
4 In the Script drop-down list, the first script added to the project is
selected. Use this script, or choose another from the list.
5 Since this is a Try Script, only one virtual user will be run. In the list of
user groups and virtual users, select the user group from which you want
to run a virtual user.
6 To view the actual data that is downloaded from a Web server during the
try-out as it is happening, select the Animated option. If you are testing
anything other than a Web application, you should disable this option.
7 Click Run.
8 The Try Script is run. All recorded think times are ignored during test
runs. The Monitor window opens; giving you detailed information about
the progress of the Try Script.
If you have selected the Animated option, TrueLog Explorer will open.
Here you can view the actual data that is downloaded during the
TryScript.
If any errors occur during the Try Script, TrueLog Explorer will help you
to find the errors quickly and to customize any problematic session
information. Once you have finished examining and customizing your
script with TrueLog Explorer, your script should run without error.
SilkPerformer User Guide 23

1 GETTING STARTED
Customizing Test Scripts
Customizing Test Scripts
Customizing the test
The fourth step in conducting a SilkPerformer load test is customizing your test
script. Script customization has several purposes: to enhance scripts with
randomized user data; to handle session-specific data that might otherwise lead
to errors; and to insert verification functions.
When an uncustomized test script is used, load tests based on that script
repeatedly reproduce the exact behavior recorded on a particular occasion. This
won’t provide accurate results. To realistically emulate the varied traffic of a
large number of different users, transactions in test scripts need to be modified
and, wherever appropriate, randomized.
The behavior of virtual users can be customized in a number of ways. They can
employ different Web browsers and modem speeds. Wherever randomized data
is appropriate, random variables can be employed to provide realistic user data.
In this way, virtual users can input different names, addresses, credit card
numbers, and so on. They can vary in the products they search for, in the
products they order, and in many other ways. When an application under test
uses load balancers, realistic simulation can be achieved by giving each virtual
user a unique IP address. Test scripts can be edited manually to provide
extensive customization.
Procedure To customize a test script:
1 Click the Customize Test button on the SilkPerformer workflow bar.
The Workflow - Customize Test dialog appears.
2 In the Profile drop-down list, the currently active profile is selected (this
is the default profile unless you have configured another profile).
3 If you want to configure simulation settings for the selected profile, click
the button (...) to the right of the drop-down list. See “Replay options”
for details.
4 If you want to add a new profile to your workload, click the second
button to the right of the drop-down list.
5 To define settings for the modem speeds to be emulated in your load test,
click the Modem Speeds button or link. See “Network options” for
details.
6 To configure simulation settings for your Web browser, click the
Browser Types link. See “Browser options” for details.
7 To create realistic functionality for the virtual users in your load test by
means of parameters, click the Create New Parameter Wizard button or
link. See “Parameter Wizard” for details.
24 SilkPerformer User Guide

1 GETTING STARTED
Establishing Baseline Performance
8 To modify user groups and transactions, click the User Group/
Transaction Wizard icon or link. See “Visual user group customization”
for details.
9 If you want to give your virtual users individual IP addresses, click the
System Configuration Manager link. It is recommended that you use this
option if your application uses a load balancer. See “Setting up
individual IP addresses” for details.
10 Once you have finished customizing your test, click the Yes button.
Establishing Baseline Performance
The fifth step in conducting a SilkPerformer load test is to ascertain an
application’s baseline performance (i.e., its ideal performance). This is achieved
by running your newly customized test script with a single virtual user per user
type. The unstressed performance of the application forms the basis for
calculating the number of concurrent users and appropriate threshold settings for
page and transaction times. The measurements typical of a real load test (e.g.,
response times and throughput) are performed during baseline tests.
A secondary reason for running a baseline test is that it serves as a trial run of the
customized test that you’ve created. The objective is to ensure that the
customizations have not introduced errors into the script, and that the script
accurately and fully reproduces the interaction that is required between the client
application and the server.
For baseline runs, the following options are set automatically to the specified
value:
• A Baseline report file is automatically created.
• The Stop virtual users after simulation time (Queuing Workload) option
is on.
• The Random thinking time option is off.
• The Load test description field is set to "BaseLine Test".
• The Display All Errors Of All Users option in the Monitor window is on.
• The Virtual user output files (.wrt) option is on.
• The Virtual user report files (.rpt) option is on.
Identifying a test
baseline
Procedure To find your test baseline:
1 Click the Find Baseline button on the SilkPerformer Workflow bar.
The Workflow - Find Baseline dialog appears.
SilkPerformer User Guide 25

1 GETTING STARTED
Establishing Baseline Performance
Confirming the test
baseline
Analyzing the baseline
report
2 Using the checkboxes on the left of the dialog, select the User Types that
you want to run in the actual baseline test. A “user type” is a combination
of a usergroup, a script, and a profile. One virtual user from each User
Type will be executed.
3 The active profile is selected in the Profile drop-down list (this is the
default profile unless you have configured another).
To add new user types to your load test, press the Add button and select a
combination of script, usergroup, and profile. In actual projects, each
profile can be selected with any usergroup from any script in the project.
4 If you want to configure the simulation settings for the selected profile,
click the button (...) to the right of the drop-down list. See “Simulation
options” for details.
5 Click Run to perform the baseline test.
6 The baseline test is run. The Monitor window opens; giving you detailed
information about the progress of the test. See “Monitoring Load Tests”
for details.
The sixth step in conducting a SilkPerformer load test is to confirm that the test
baseline established by the test in the previous step actually reflects the desired
performance of the application under test. This is done by inspecting the results
of that test in a baseline report. If the results are satisfactory, the baseline can be
accepted and will form the basis for calculating the number of concurrent virtual
users and thresholds of specific timers in future tests. The baseline results will be
used later for comparison with the results of actual load tests.
A secondary reason for running a baseline test is that it serves as a trial run for
the customized test. The baseline confirmation step provides an opportunity to
inspect the measurements generated during the test to ensure that the final load
test contains all the required results.
At this stage, server monitoring is set up so that live monitoring of the server
will occur during load tests, and server-side results information is generated for
exploration and correlation after the test.
Procedure To confirm your test baseline:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
The Workflow - Confirm Baseline dialog appears.
2 To view a report of the results from your baseline test, click the Baseline
Report button or link. In the Summary report that appears, you can
accept the results of the previous test as your baseline by clicking the
Accept Baseline button in the report and clicking Yes on the confirmation
dialog.
26 SilkPerformer User Guide

1 GETTING STARTED
Defining Workload
Note To view the accepted baselines that are currently in effect,
click the Confirm Baseline button. Then click the Accepted Baseline
Report link to view the results. See “Viewing Accepted Baseline
Reports” for details.
3 To set thresholds for future load tests, click the set response time
thresholds button or link. MeasureSetBound statements will be added to
your script accordingly. See “Setting Response Time Thresholds” for
details.
4 To receive online performance data during load tests, you can establish a
monitoring template. See “Monitoring server performance during tests”
for details.
5 If you want to see the output of write statements used in the test script
(assuming their generation has been enabled) click the Virtual User
Output Files link. See “Virtual User Output files” for details.
6 If you are satisfied that you have confirmed the baseline performance of
your application, click Yes.
Note If the baseline test did not identify the correct baseline
performance for your application, or if any errors occurred during
the baseline test, click No. Then click the Customize Test button and
repeat the steps outlined in “Customizing the test”.
Defining Workload
Prior to the execution of a load test, you must select the workload model you
want to use for the test.
You must select one of the following workload models:
• Increasing
• Steady State
• Dynamic
• All Day
• Queuing
• Verification
Note Additionally, the number of concurrent virtual users per user
type, the duration, and the involved agents must be defined.
Specifying load test
workload
Procedure To specify workload for a load test:
1 Click the Adjust Workload button on the SilkPerformer Workflow bar.
SilkPerformer User Guide 27

1 GETTING STARTED
Executing Load Tests
2 Select the Workload model that best fits your needs and click the
Workload Wizard button.
3 Specify the simulation times of the load test and click next.
4 The number of concurrent virtual users is calculated automatically when
you specify the number of user sessions you expect per hour.
In some cases it’s desirable to adjust the session time of user types to a
particular duration. By adjusting a user type’s session time, throughput
can be influenced on a per-virtual-user basis to meet particular load test
requirements.
5 After completing your specifications, the Workload Configuration dialog
appears and allows you to change any of the load test’s parameters.
Executing Load Tests
The eighth step in conducting a SilkPerformer load test is to run a full load test.
Based on the test script, multiple virtual users are run to test the target server. A
large load test requires that an appropriate testing environment be set up in the
local area network, including a full complement of agent computers to host the
virtual users.
It is essential that you carefully set the options for the appropriate type of test to
be used to accurately define the necessary workload, and to enable generation of
the types of test results that will be required to assess the performance of the
server. The TrueLog On Error option helps you to visually inspect replay errors.
With TrueLog On Error, the necessary log files are written to disk only when
errors occur.
Comprehensive information is provided to you while the load test is performed.
This includes real-time information about agent computers, virtual users, and
transactions—in real-time, as they are conducted. Real-time monitoring of the
target server is available as a graph.
If the test is configured as a verification run, the following options will be
automatically set to the specified value:
• A Baseline Report file will be automatically created.
• The Stop virtual users after simulation time (Queuing Workload) option
is off.
• The Virtual user log files (.log) option is off.
• The Virtual user report files (.rpt) option is on.
• The Virtual user report on error files (.rpt) option is on.
• The Compute time series data (.tsd) option is off.
28 SilkPerformer User Guide

1 GETTING STARTED
Exploring Test Results
Running the load test
Procedure To run a load test:
1 Click the Run Test button on the SilkPerformer Workflow bar.
The Workload Configuration dialog appears.
2 If necessary, change the workload configuration for your load test. Then
click Run to begin the test. See “Workload Definition” for details.
3 While the test is in progress, you can monitor its progress and monitor
server activity by viewing the tabular monitor view in SilkPerformer and
the graphical monitor view in Performance Explorer. See “Monitoring
Load Tests” for details.
Exploring Test Results
Exploring test results
The ninth and final step in conducting a SilkPerformer load test is to display and
analyze the results of the test to assess the performance of the server under test.
For detailed information on exploring test results, please refer to “Exploring
Test Results”.
Procedure To explore test results:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 To view a report of the results from your load test, click the Baseline
Report button or link. In such Summary reports, you can accept the
results of any previous test as your baseline. See “Viewing Baseline
Reports” for details.
3 If you want to see the accepted baselines so far, click the Accepted
Baseline Report button to view the results. See “Viewing Accepted
Baseline Reports” for details.
4 If the previous load test produced errors and the TrueLog On Error
option has been enabled, you can inspect the errors with TrueLog
Explorer. See “Visual Analysis with TrueLog Explorer” for details.
5 To view a graphic display of the results from your load test, click the
Performance Explorer button or link. See “Performance Explorer” for
details.
SilkPerformer User Guide 29

1 GETTING STARTED
Exploring Test Results
30 SilkPerformer User Guide

2
Configuring 2Chapter
SilkPerformer
Introduction
What you will learn
This chapter explains how to configure SilkPerformer settings.
This chapter contains the following sections:
Section
Page
Overview 31
SilkPerformer Workbench 32
SilkPerformer Recorder 39
Application Profiles 41
Proxy Settings 45
Remote Agents 48
Client Certificates 52
Java Configuration 55
Overview
The SilkPerformer configuration options provide the opportunity to set global
SilkPerformer settings that are not associated with a particular project.
“SilkPerformer Workbench”: The basic characteristics of a SilkPerformer load
testing configuration are set here—those related to the agent computers used in
tests, the directories where files used in and generated by tests are located, the
layout of the text of test scripts, and how results information is generated.
“SilkPerformer Recorder”: The basic settings used by SilkPerformer's
recording and script generating program, the SilkPerformer Recorder, are
defined here. These include the settings for capturing and recording the network
traffic that will be modeled in your load test scripts. Here, profiles are set up and
SilkPerformer User Guide 31

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
managed for the client applications that are to be used during data capture and
recording. Proxies can be added or removed, and their settings edited. Here you
can set the port for the integrated Web server that is used by the SilkPerformer
Recorder, and enable recording to begin as soon as the client application starts.
“Remote Agents”: SilkPerformer allows for remote agent computers to be
located both in the LAN, out on the Internet, and even behind firewalls—
connected through HTTP and SOCKS proxies. The communication between the
remote agents and the controller can be configured to include secure, encrypted
connections in a distributed secure environment.
“Java Configuration”: Java Framework projects and Java based APIs require a
Java SDK Environment. The Java SDK version, its location, and parameters
such as class path and optional Java Virtual Machine settings can be configured
here.
SilkPerformer Workbench
The basic characteristics of SilkPerformer load testing configuration are set
from the System Settings dialog (Settings/System).
Agent Pool
The pool of agent computers that can be used with the controller computer can
be set up and managed here. New agents can be added to the agent pool, either
from the local network or from somewhere on the Internet, or existing agents
can be removed. A wide range of system options for agent computers can be set,
and their availability for service can be checked.
Directories
You can specify the directories where various types of SilkPerformer files are
located. Default directories are set up during installation; you can specify
alternative directories here. You can set locations for load testing scripts, for
include files, for the various file types containing the results of load tests, and for
user data files, certificate files, and random data files.
Layout
The way the editor behaves when you manually insert text into a script file can
be customized; options can be set for formatting, printing, and how keywords
are highlighted in the SilkPerformer tree view editor.
Control
The options for the multi-machine controller can be set in this area. Settings can
be specified for the time interval within which a controller computer must
communicate with an agent computer or a virtual user; if it fails, a timeout error
will be reported. The frequency with which the controller will ping an agent
computer until it receives a reply is set here, as can the number of virtual users
per process and the number of transaction failures permitted for a virtual user
before the user is terminated.
32 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
Configuring agent
pool settings
Results
Use the Time series data option for merging time series data files (.tsd) on the
agent computers where they were created and to ensure that only the merged
files will be sent to the controller computer. This is especially useful for quickly
processing the results of long load tests as the calculations will then be
distributed to various processors and less traffic will be sent over the network.
The drawback, however, is that time series data files for individual virtual users
are not available on the controller computer.
Note If you do not select this option, you manually have to merge
the time series data files in Silk Performance Explorer in order to get
overall results for the load test.
Workspace
This area is used to specify general settings for the Workbench. Workbench
options include start-up settings, workflow settings, file handling, and the
display of message boxes.
Test Manager
Use this area to set up a connection to SilkCentral Test Manager.
Source Control
Use this area to specify general SCC connection parameters. Click a specific
item for detailed information.
Procedure To specify SilkPerformer agent pool settings:
1 From the SilkPerformer menu bar, select Settings/System.
The System Settings dialog opens at the Agent Pool tab.
2 Use the Agent Pool list to view the computers that are currently
configured as part of the SilkPerformer agent pool.
To add a computer to the default pool of agents associated with this
controller computer:
3 On a LAN network: Click the Add LAN Agent button.
The Add LAN Agent dialog appears, listing all compatible agent
computers that are available on the subnet.
a From the Running agents list of computers in your local network,
select the computer(s) you want to add to your agent pool.
b Click the Add button.
4 On a WAN network: Click the Add WAN Agent button.
The Add WAN Agent dialog appears. Enter connection parameters for the
WAN agent.
a In the Host name or IP address field, enter the computer’s host name
or IP address.
SilkPerformer User Guide 33

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
Configuring directory
settings
b If you want to add a computer from Windows domains outside your
local network, click the browse (...) button.
The Browse for Computer dialog appears.
c Select the computer you want to add to the agent pool and click OK.
d Click OK again on the Add WAN Agent dialog.
5 To remove agent(s) from the default pool of agents associated with this
controller computer, select them in the agent pool list and click the
Remove button.
6 In the ensuing SilkPerformer Message dialog, confirm that you want to
remove the agent(s).
7 To set configuration options for an agent computer, click the Manage
button. This opens the System Configuration Manager, which allows you
to set options for the selected agent computer, including configuring
multiple IP addresses.
8 To validate that a selected agent is available and active, click the Check
Availability button, which will send a remote call to the agent.
Procedure To specify directory settings:
1 From the SilkPerformer menu bar, select Settings/System.
2 Click the Directories tab.
3 Use the File locations area to specify the directories where various
SilkPerformer file types are stored. Default directories are set during
installation.
Click the button (...) to the right of each field to browse the directory
structure and locate appropriate files and directories.
4 In the Projects field, specify the directory where the project files are to
be saved.
5 In the User data files (.pem, .rnd, .csv, .txt, .idl) field, specify the
directory where the system-defined user data files (.csv), certificate files
(.pem), random data files (.rnd files), text files (.txt), and Interface
Definition Language files (.idl) are to be stored. The compiler first
searches for random data files in the directory where the .bdf file is
located, then in this directory. To use a random data file from a different
location, specify a full path for the file in your script file. This location is
used for user data files that are delivered with SilkPerformer.
6 In the Custom user data files (.pem, .rnd, .csv, .txt, .idl) field, specify the
directory where your self-created user data files (.csv), certificate files
(.pem), random data files (.rnd files), text files (.txt), and Interface
Definition Language files (.idl) are located. In contrast to the specified
34 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
Configuring layout
settings
User data files location, this location is used for your own user data files.
As with SilkPerformer pre-defined user data files, custom user data files
in this directory can be shared across multiple projects.
7 In the Include files (.bdh) field, specify the directory where the include
files (.bdh files) are located. The compiler searches for include files first
in the directory where the .bdf file is located, and then in this directory. If
you want to use an include file from a different location, you can specify
a full path in the use statement in your script for your include file.
External function prototypes, internal functions, constants, random
variables, and external function declarations are reusable in different
script files when defined in include files. This location is used for include
files that are delivered with SilkPerformer.
8 In the Custom include files (.bdh) field, specify the directory where your
self-created include files (.bdh files) are located. In contrast to the
specified Include files location, this location is used for your own include
files. As with SilkPerformer pre-defined include files, custom include
files in this directory can be shared across multiple projects.
9 In the Working field, specify the default directory for opening and saving
load test scripts (.bdf files).
Procedure To specify layout settings:
1 From the SilkPerformer menu bar, select Settings/System.
2 Click the Layout tab.
3 Use the Editor area to customize how the editor behaves when you
manually insert text into a script file.
4 In the Tab. size field, enter the number of space characters that equal one
tab character.
5 Select the Auto indent option to automatically insert indents in your test
script. Each time you create a new line in a script file, the line is
automatically indented the same number of spaces as the previous line.
6 Select the Replace tabs with spaces option to automatically insert four
spaces each time a tab is used in a test script. This option makes it easier
to view script files in a text editor.
7 Click the Font & Colors button to select a font and style for
SilkPerformer's script editor. You can select a different color to use for
each type of content in a test script.
The Editor Font and Colors Settings dialog opens.
8 From the Text type list, select the type of text for which you want to
adjust the font, size, style, and color.
SilkPerformer User Guide 35

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
Configuring control
settings
- Normal Text - Identifiers and symbols.
- Keywords - Predefined statements.
- Custom Keywords - Words added to the keyword list by the user.
- Numbers - Whole numbers and floating-point values.
- Comments - Notes in explanation included in a script.
- Strings - Sequence of characters delimited by " characters.
9 From the Font name, Font size, Style, and Color drop-down lists, select
the format options you want to use for the currently selected text type.
10 Click OK.
11 Use the Printing area to specify options for how script files are formatted
when they are printed. Select the Line number option to automatically
number the lines in the printed script file.
12 Select the Wrap marker option to automatically print a marker at the end
of each line that wraps to the next line.
13 Click the Font button to select a font and style for your printed test
scripts. The Printer Font and Colors Settings dialog opens.
14 From the Text type list, select the type of text for which you want to
adjust the font, size, and style. (See step 8).
15 From the Font name, Font size, and Style drop-down lists, select the
format options you want to use for the currently selected text type.
16 Click OK.
17 Use the Custom keywords area to define custom keywords. For
readability, keywords are highlighted in the SilkPerformer tree view
editor.
18 To enter a new custom keyword into the list, click the New button to the
immediate right of the list heading. All keywords entered in the list are
highlighted in SilkPerformer's tree view editor for readability.
19 To remove a custom keyword from the list, select it and click the Delete
button to the far right of the list heading.
Procedure To specify control settings:
1 From the SilkPerformer menu bar, select Settings/System.
2 Click the Control tab.
3 Use the Virtual users area to set options for running the virtual users in
your load test. In the Virtual users per process field, insert the number of
virtual users to be run within each system process.
36 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
4 In the Stop virtual users after __ transaction failures field, enter the
number of transaction failures after which virtual users should be
stopped.
5 Use the Remote agent connection area to specify whether or not to
optimize the controller-agent communication for local or wide area
networks.
Select the LAN option for the controller-agent communication to be
optimized for a local area network. This is especially useful if all of the
agent computers you are going to use for your load tests are located in
the same local area network and all are using a fast connection.
Select the WAN option for the controller-agent communication to be
optimized for a wide area network. This is especially useful if you are
going to use many agent computers that are located somewhere on the
Internet.
Select the Custom option for entering custom parameters for agentcontroller
communication.
Note Use the Custom option only if you run into problems with
LAN or WAN settings.
6 In the Connect field (enabled only when the Custom option is selected),
enter the time, in seconds, in which the controller computer should try to
establish a connection to each remote agent computer. If connection to an
agent computer is not successful within this time interval, SilkPerformer
will report it as an error.
Note It is recommended that you not change this setting.
7 In the Ping Interval field, enter the length of time, in seconds, within
which the controller computer should attempt to retrieve load test status
data from the agent computers. The controller computer requests this
information from the agent computers whenever the contents of the
Monitor window need to be updated. If the agent computer fails while
receiving data during this time interval, SilkPerformer will report a
warning.
Note It is recommended that you not change this setting.
8 Edit the time (in seconds) in the Command field (enabled only with the
Custom option selected), in which the controller computer should try to
send a command to a remote agent computer. If sending the command to
the agent computer is not successful within this time interval,
SilkPerformer will report it as an error.
Note It is recommended that you not change this setting.
SilkPerformer User Guide 37

2 CONFIGURING SILKPERFORMER
SilkPerformer Workbench
Configuring result
settings
Configuring
workspace settings
Procedure To specify results settings:
1 From the SilkPerformer menu bar, select Settings/System.
2 Click the Results tab.
3 Select the Merge time series data on agents option in the Time series
data area for time series data files (.tsd) to be merged in the agent
computers where they were created and to ensure that only the merged
files are sent to the controller computer. This is especially useful for
quickly processing the results of long load tests as the calculations will
then be distributed to various processors and less traffic will be sent over
the network. The drawback, however, is that time series data files for
individual virtual users are not available on the controller computer.
Note If you do not select this option, you will have to manually
merge the time series data files in Performance Explorer to get
overall results for the load test.
4 Click OK to exit the dialog and save your changes.
Procedure To specify workspace settings:
1 From the SilkPerformer menu bar, select Settings/System. Click the
Workspace tab.
2 Select the Open the last opened project file on startup option in the
Workbench start options area to have the last load test project you
worked on opened automatically each time you launch SilkPerformer.
3 Use the Workflow options area to specify the display of the workflow bar.
Select the Always enable all workflow buttons checkbox to enable all
buttons in the workflow bar. By default, workflow buttons are enabled in
sequence based on your progress through SilkPerformer’s workflow.
4 Select the Show simple workflow checkbox to enable simple workflow.
Simple workflow bypasses the baseline concept to make the load testing
process quicker and easier.
5 Check the Enable document file locking checkbox in the Document file
options area to enable document file locking. The document file locking
mechanism allows multiple users (on SilkPerformer Controller
machines) to work simultaneously on the same document files (bdf script
files, bdh include files, data files) without overwriting the files of each
other. Locked document files from other users can still be used read-only.
This option is enabled by default.
38 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
SilkPerformer Recorder
6 Click the Show All Messages button in the Uncheckable message boxes
area to reset the display setting of all uncheckable message boxes to their
default value so that all message boxes will be displayed again as long as
you do not uncheck the corresponding setting.
To specify Test Manager settings, see “Configuring Test Manager Integration”.
To specify Source Control settings, see “Configuring Source Control
Integration”.
SilkPerformer Recorder
The basic settings used by SilkPerformer's recording and script generation
program, the SilkPerformer Recorder, include settings for capturing and
recording the network traffic that will be modeled in your load test scripts.
These settings comprise profiles that are applied to client applications during
data capture and recording. Proxies can be added or removed, and their settings
edited. You can set the port for the integrated Web server that is used by the
SilkPerformer Recorder, and enable recording to begin as soon as the client
application is started.
Application profiles
In configuring the SilkPerformer Recorder, a wide choice of options can be set
for client applications to use during data capture and recording. New application
profiles can be added and named; existing profiles can be edited, removed, or
copied; and applications can be started. Options that can be set include the
application path, working directory, program arguments, and application type
for each application, as well as the method that the SilkPerformer Recorder is to
use for recording traffic.
Proxies
New proxies can be set up on your computer, and existing proxies can be edited
or removed. A wide number of options can be specified for each proxy,
including the protocol type, the listen port, details of the server to which the
SilkPerformer Recorder forwards intercepted traffic, and details concerning the
client certificate that the SilkPerformer Recorder presents to the Web server.
When a SOCKS proxy is being set up, automatic protocol detection can be
disabled for any port you want to specify. Whenever a client application sends
SOCKS traffic to one of these specified ports, the SilkPerformer Recorder
records traffic at the TCP/IP level without trying to detect a familiar protocol
first. A port range can be specified for which the recording of data by the
SilkPerformer Recorder will be suppressed.
Services
Options can be set for the Web server that is integrated into SilkPerformer, and
also for recording with the SilkPerformer Recorder. The port that the integrated
Web server listens to can be specified, and you can use this server to retrieve the
SilkPerformer User Guide 39

2 CONFIGURING SILKPERFORMER
SilkPerformer Recorder
Configuring
application profile
settings
Configuring proxy
settings
Configuring services
root CA certificate that signed the SilkPerformer Recorder server certificate.
The root CA certificate may be downloaded from this Web site and installed into
the Web browser to avoid security warnings by the browser. See “Client
Certificates” for details.
The SilkPerformer Recorder can be set to begin recording function calls
automatically when an application is started from within the SilkPerformer
Recorder.
Procedure To specify application profile settings:
1 From the SilkPerformer menu bar, select Settings/System. Click the
Recorder icon.
2 The Recorder group opens at the Application Profiles tab.
3 Use the Application profile list to view the application profiles that are
set up for recording network traffic, to add new application profiles, and
to edit, remove, or copy existing profiles. You can also start applications
from here.
4 To add, edit, copy, or remove application profiles, or to start an
application, see “Application Profiles”.
Procedure To specify proxy settings:
1 From the SilkPerformer menu bar, select Settings/System. Click the
Recorder icon.
2 Click the Proxies tab. Use the Proxies list to activate proxies, to add
proxies, or to edit or remove existing ones.
3 To activate, add, edit, remove, or copy a proxy, see “Proxy Settings”.
Procedure To specify services settings:
1 From the SilkPerformer menu bar, select Settings/System. Click the
Recorder icon.
2 Click the Services tab.
3 Use the Integrated Web server area to change the port number of the Web
server that is integrated into the SilkPerformer Recorder. The default
address of the server is http://localhost:19100 (where 19100 is the
default port number). You must have the SilkPerformer Recorder
running to use this Web server.
4 In the Port field, enter the port that the integrated Web server listens to.
The default is 19100. Using the integrated Web server, you can retrieve
the root CA certificate that signed the SilkPerformer Recorder server
40 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Application Profiles
certificate. The root CA certificate may be downloaded from this Web
server and installed into the Web browser to avoid security warnings by
the browser. See “Client Certificates” for details.
5 Select the Autostart recording option in the Recording area to start
recording traffic and function calls automatically when SilkPerformer’s
Recorder is started. If this option is disabled, you must use the Start
Recording button on the SilkPerformer Recorder toolbar to begin
recording.
6 Click OK to exit the dialog and save your changes.
Application Profiles
Adding application
profiles
Procedure To add an application profile:
1 From the SilkPerformer menu bar, select Settings/System.
The System Settings - Workbench dialog appears, open at the Agent Pool
tab of the Workbench group.
2 Click the Recorder icon on the left of the dialog. The Recorder group
opens at the Application Profiles tab.
You can also reach this dialog by means of the Model Script button on
the SilkPerformer Workflow bar and the ensuing dialog, or from the
SilkPerformer Recorder.
3 Click the Add button to add a new application profile to the list.
The Application Profile dialog appears.
4 In the Application profile field, enter a unique name for the application
profile. The name is required to identify the application in the
SilkPerformer Recorder window.
5 In the Application path field, enter the name of the application's
executable and the directory where it is located. To locate the executable,
click Browse.
6 Select the Record executable is different from application path option to
enable recording of a process that is different from the specified
application's executable. This may be required if the application is hosted
by another process that powers the communication process.
7 In the Record executable field, enter the name of the application's
executable and the directory where it is located. To locate the executable,
click Browse.
SilkPerformer User Guide 41

2 CONFIGURING SILKPERFORMER
Application Profiles
8 In the Working directory field, enter the working directory of the
application. The working directory is the folder that contains the
application or some related files. Some applications need files located in
other directories, so you may have to specify the folder where these files
are located.
9 In the Program arguments field, enter any number of application
parameters. Passing parameters to an application is useful, for example,
when recording the function calls performed by a command line tool that
has no graphical user interface.
10 From the Application type drop-down list, select the type of application
traffic you want the SilkPerformer Recorder to capture:
- Custom Application to have the SilkPerformer Recorder hook into a
particular function library of the application, and capture and record
all the traffic that the application generates.
- MS Internet Explorer to have the SilkPerformer Recorder redirect
traffic from Microsoft's Internet Explorer to the Recorder proxies.
- Netscape Browser to have the SilkPerformer Recorder redirect
traffic from your Netscape browser to the Recorder proxies.
- Windows NT Service to have the SilkPerformer Recorder capture and
record all traffic generated by a Windows service.
11 Use the Protocol selection area of the dialog to specify the function
library that you want the SilkPerformer Recorder to hook into.
Note The options available here vary based on the selected
application type.
12 Select the Web option to have the SilkPerformer Recorder hook into a
Web application, intercept all function calls, and display results.
When the Web option is selected, you can click the Web Settings button
to open the Web Settings dialog, which enables you to select the method
that the Recorder is to use to capture Web and TCP/IP-based traffic.
- Select the WinSock option to have the SilkPerformer Recorder hook
into the WinSock function library of the client application, wrap all
traffic the application produces in the SOCKS protocol, and redirect
the traffic to a Recorder SOCKS proxy.
- Select the Proxy option to have the SilkPerformer Recorder redirect
all traffic from the selected browser to the Recorder proxies.
Note: This is available only when MS Internet Explorer or Netscape
Browser is selected in the Application type list box.
42 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Application Profiles
- Select the SocksCap option to have the SilkPerformer Recorder use
SocksCap to hook into the WinSock function library of the
application for which you want to record traffic. SocksCap wraps all
the traffic the application produces in the SOCKS protocol and
redirects the traffic to a Recorder SOCKS proxy.
Note: This is available only when SocksCap is installed and Custom
Application, MS Internet Explorer or Netscape Browser is selected
in the Application type list box.
- When the Web option is selected you can also enable any available
Java APIs by selecting the corresponding checkboxes in the field on
the right of the dialog. The following Java APIs are available for
selection:
o Jacada
o Java EJB (JBoss)
o Java EJB (Other)
o Java EJB (WebLogic)
o Java EJB (WebSphere)
o Java JNDI (Naming & Directory Interf.)
o Java RMI (Custom Client-Server)
o Jolt
o Oracle Applications 11i
o Oracle Applications 12i
o Oracle Forms 6i
o Oracle Forms 6i on HTTP/HTTPS
o Oracle Forms 6i on Socket
o Oracle Forms 9i
o Product Manager Sample Extensions
- When a Java API is enabled, you can click the Java Settings button
to open a dialog that enables you to specify Java recording settings,
such as the hooked Java Virtual Machine library or other settings.
Select the Java Virtual machine hooking option to automatically
hook into the Java VM. If this option is enabled, from the JVM DLL
drop-down list you must select the dynamic link library you want the
SilkPerformer Recorder to hook into for recording traffic.
Select the Network only option to only hook class and jar files that
SilkPerformer User Guide 43

2 CONFIGURING SILKPERFORMER
Application Profiles
are downloaded via HTTP.
Select the Manual option if you have manually prepared a startup
script for recording your Java application.
13 Select the TCP/IP option to have the SilkPerformer Recorder hook into a
TCP/IP application, intercept all function calls, and display the results.
14 Select the IIOP option to have the SilkPerformer Recorder hook into the
IIOP function library of the client application, intercept all the function
calls, and display the results.
15 Select the Citrix option to have the SilkPerformer Recorder hook into a
Citrix application, intercept all function calls, and display the results.
16 Select the ODBC option to have the SilkPerformer Recorder hook into
the ODBC function library of the client application, intercept all function
calls, and display the results.
17 Select the DB2 CLI option to have the SilkPerformer Recorder hook into
the DB2 CLI function library of the client application, intercept all
function calls, and display the results.
18 Select the Oracle OCI option to have the SilkPerformer Recorder hook
into the Oracle function library of the client application, intercept all
function calls, and display the results.
Click the OCI Settings button to open a dialog that allows you to select
the OCI client library that is to be hooked into for intercepting function
calls. Select the appropriate library from the drop-down list and click
OK.
19 Select the Tuxedo option to have the SilkPerformer Recorder hook into
the TUXEDO function library of the client application, intercept all
function calls, and display the results.
20 Select the COM option to have the SilkPerformer Recorder hook into a
COM application, intercept all function calls, and display the results.
Click the Interfaces button to open a dialog that allows you to select the
server interfaces that should be used with the COM application. Select
appropriate interfaces from the drop-down list and click OK.
To add a new server interface, select the Add button on the Server
Interfaces dialog.
To delete a selected server interface from the list, select the Remove
button on the Server Interfaces dialog.
21 Select the ADO option to have the SilkPerformer Recorder hook into an
ADO application, intercept all function calls, and display the results.
22 Select SAPGUI option to have the SilkPerformer Recorder hook into
functions of SAPGUI interfaces, intercept all the function calls, and
display the results
44 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Proxy Settings
23 Click OK to add the new profile to the Application Profiles list.
Editing, copying, &
removing profiles
Procedure To edit, copy, or remove an application profile, or to start an
application:
1 At the Application Profiles tab in the Recorder group, select an
application profile.
2 Click the Edit or Copy button. The Application Profile dialog appears.
3 Edit or copy the profile as necessary, based on the instructions detailed
“Adding application profiles”
4 To remove a selected application profile, click the Remove button.
5 Click Yes on the confirmation dialog to remove the profile.
Proxy Settings
Configuring proxy
settings
Procedure To configure proxy settings:
1 From the SilkPerformer menu bar, select Settings/System.
The System Settings - Workbench dialog appears, open at the Agent Pool
tab of the Workbench group.
2 Click the Recorder icon on the left of the dialog. The Recorder group
opens at the Application Profiles tab.
3 Click the Proxies tab.
4 Use the Proxies list to activate proxies, add proxies, edit proxies, or
remove proxies.
5 To activate a proxy, select the checkbox to the left of the proxy name.
6 Click the Add button to set up a new proxy on the controller computer.
You can set up proxies for recording a number of types of traffic, and you
can have a number of proxies active at the same time.
The Proxy Settings dialog appears.
7 Use the Proxy type area to select a protocol type and a port for
forwarding and recording traffic.
8 From the Protocol drop-down list, select a protocol type for forwarding
and recording traffic. You have to set up a proxy for each type of traffic
you want to record. Forwarding and recording traffic with the following
protocols is supported:
- FTP - File Transfer Protocol
- HTTP - Hypertext Transfer Protocol
- LDAP - Lightweight Directory Access Protocol
SilkPerformer User Guide 45

2 CONFIGURING SILKPERFORMER
Proxy Settings
- POP3 - Post Office Protocol
- SMTP - Simple Mail Transport Protocol
- SOCKS - When capturing SOCKS traffic, the SilkPerformer
Recorder automatically detects other protocols like IIOP, FTP, POP3
and SMTP for recording.
- TCP/IP - (Custom Protocol) Custom protocol based on TCP/IP.
9 In the Listen port field, enter the port number to which traffic will be sent
by the traffic-generating client computer. Although any port not in use
can be chosen, avoid standard port numbers (usually smaller than 1500).
10 Use the Remote host area to specify the host name or the IP address, and
the port of the server to which you want the SilkPerformer Recorder to
forward intercepted traffic.
11 Select the Any option to forward intercepted traffic to the host specified
by the traffic-generating client computer.
Select the Proxy/Firewall option to forward intercepted traffic to a
specified proxy or firewall. In this case, the host name and IP address
specified by the traffic-generating client computer are ignored.
Select the Specified option to forward intercepted traffic to a specified
remote host. In this case, the host name and IP address specified by the
traffic-generating client computer are ignored.
12 In the case of a specified remote host, select the Secure checkbox to
record and forward Secure Socket Layer protocol traffic. If this option is
selected, HTTPS traffic will be recorded instead of HTTP traffic, for
example.
13 In the Host name or IP address field, enter the host name or IP address of
a specific server to which you want the SilkPerformer Recorder to
forward intercepted traffic.
14 In the Port field, enter the port number that a specific server—the one to
which the SilkPerformer Recorder forwards intercepted traffic—listens
to.
15 Use the Client certificate area to select a client certificate that the
SilkPerformer Recorder presents to the Web server on behalf of the real
client's certificate. To use a client certificate, you must export it from the
Web browser and place it in the Recorder's certificate store first.
16 From the Certificate drop-down list, select a client certificate that the
SilkPerformer Recorder is to present to the Web server on behalf of the
real client’s certificate.
46 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Proxy Settings
17 Click the View button to view detailed information about the client
certificate currently in use. The information appears in the View
Certificate dialog.
18 Click the Import button to import a new client certificate from a
certificate file (.p12, .pfx, or .pem). To use a client certificate, you must
export it from the Web browser and place it in the SilkPerformer
Recorder's certificate store first.
19 The Open Certificate dialog appears. Use this dialog to locate and select
the new client certificate that is to be imported.
20 To remove the client certificate that is currently in use, click the Remove
button.
21 Click Yes on the confirmation dialog to remove the certificate.
22 When setting up a SOCKS proxy, use the Record always at TCP/IP
protocol level area to specify any number of ports for which to disable
automatic protocol detection. Whenever a client application sends
SOCKS traffic to one of the specified ports, the SilkPerformer Recorder
records traffic at the TCP/IP level without attempting to detect a familiar
protocol first.
23 In the Within port range field, you can enter multiple port numbers and
ranges, separated by commas (,); a range specification consists of the
lower and upper boundary, separated by a hyphen (-).
24 Use the Suppress recording (only forward data) area to specify a port
range for which you want to suppress the recording of data by the
SilkPerformer Recorder. The data will still be forwarded, however.
25 In the Within port range field, enter a port range for which you want to
suppress the recording of data by the SilkPerformer Recorder; the data
will still be forwarded. You can enter multiple port numbers and ranges,
separated by commas (,); a range specification consists of the lower and
upper boundary, separated by a hyphen (-).
26 Click OK.
27 To edit a selected proxy, click the Edit button.
The Proxy Settings dialog appears.
28 Use steps outlined in this section to edit the selected proxy settings.
29 To remove a selected proxy, click the Remove button.
30 Click Yes on the confirmation dialog to remove the proxy.
31 Click OK to exit the dialog and save your changes.
SilkPerformer User Guide 47

2 CONFIGURING SILKPERFORMER
Remote Agents
Remote Agents
Configuring remote
agents, specifying
port settings
Configuring remote
agents, specifying a
password
Configuring the
controller connection
SilkPerformer allows for remote agent computers to be located both in your
LAN or somewhere out on the Internet. The communication between the remote
agents and the controller can be configured in a distributed secure environment
to include secure and encrypted connections.
Communication between remote agents and the controller may be made in a
variety of ways with a choice of proxies, including connections made through
multiple firewalls and through TCP/IP, HTTP, or SOCKS firewalls. These
connections include access through choices for secure or non-secure
connections, with or without encryption.
Procedure To configure remote agents and specify port settings:
1 From the SilkPerformer menu bar, select Tools/System Configuration
Manager.
2 The System Configuration Manager dialog appears; click the Service
Status tab.
3 In the Port and Secure port fields, enter the numbers of the unsecured
port and the secured port. By default these settings are 19200 for the
unsecured port and 19201 for the secured port, respectively.
4 Click OK.
Procedure To configure remote agents and specify a password:
1 From the SilkPerformer menu bar, select Tools/System Configuration
Manager.
2 The System Configuration Manager dialog appears; click the Service
Status tab.
3 To require that a password be used to secure controller/agent
communication, click the Change button in the Password area. By
specifying a password here, the controller must supply a password to be
able to connect to that remote agent.
4 After specifying a password, confirm the password, and then click OK.
Procedure To configure how the controller connects directly to remote
agents:
1 From the SilkPerformer menu bar select Settings/System.
2 The System Settings-Workbench dialog appears; click the Agent Pool tab.
48 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Remote Agents
Connecting through
TCP/IP firewalls
3 In the Agent Pool tab you can choose to configure one of the agents you
have already connected by choosing that agent from the window. Double
click the name and computer icon for the agent you have already
connected. If you have not yet chosen and named an agent to connect,
then click the Add WAN Agent button. A LAN agent may also be added
and configured at this time by choosing the Add LAN Agent button and
choosing the agent from the list. After naming an agent, then proceed as
you would for an agent already chosen, as follows:
4 The Agent Connection Properties or the Add WAN Agent dialog appears,
depending on if you are configuring an already chosen agent or adding a
WAN agent; apart from the name in the name bar of the dialog box, these
two dialog windows are the same. You have a number of options in this
dialog for choosing how the controller connects to remote agents:
- Security - Encryption (SSL): By checking this option, you force the
controller to use a secure connection to the remote agent.
- Security - Authenticate/Password: By checking this option and
supplying a password, the controller will connect to the remote agent
using a password. The password must be identical to the password
you specified in the System Configuration Manager.
- Connection - Connect to port: The controller connects to the agent
using this port if you have turned off the security. The port must be
identical to the Agent connection port you specified in the System
Configuration Manager.
- Connection - Connect to secure port: The controller connects to the
agent using this port if you have turned on the security option. The
port must be identical to the Agent connection secure port you
specified in the System Configuration Manager.
- Firewall - Use HTTP proxy server: The controller connects via an
HTTP firewall to the agent by tunneling the controller/agent
communication in the HTTP protocol.
- Firewall - Use Socks proxy server: The controller connects via a
SOCKS firewall to the agent by tunneling the controller/agent
communication in the SOCKS protocol.
Procedure To configure how the controller connects to remote agents
through TCP/IP firewalls:
1 From the SilkPerformer menu bar, select Tools/System Configuration
Manager.
2 The System Configuration Manager dialog appears; click the Service
Status tab.
SilkPerformer User Guide 49

2 CONFIGURING SILKPERFORMER
Remote Agents
3 In the Port and Secure port fields, enter the numbers of the unsecured
port and the secured port. By default these settings are 19200 for the
unsecured port and 19201 for the secured port, respectively. These port
connections must be changed in order to tell SilkPerformer where it can
go through the firewall if the port is open.
4 If you want to change the port settings, the following examples may be
helpful to you:
Connecting through
HTTP firewalls
Ports to open on the firewall:
- 20000: default connection port to SilkPerformer ACS
- 20020: secure connection to ACS (you only have to open this port if
you want a secure connection).
- 20021, (20022*): for connecting to SilkPerformer Simulation
Controller Agent and System Configuration Manager Agent. (Port
20022* is only to be used if you want the System Configuration
Manager and a Simulation Controller Agent to run simultaneously.)
5 Click OK.
Procedure To configure how the controller connects to remote agents
through HTTP firewalls:
1 From the SilkPerformer menu bar select Settings/System.
2 The System Settings-Workbench dialog opens to the Agent Pool tab.
3 In the Agent Pool tab you can choose to configure one of the agents you
have already connected by choosing that agent from the window.
Double-click the name and computer icon/image for the agent you have
already connected. If you have not yet selected and named an agent to
connect to, click the Add LAN Agent or Add WAN Agent button and
proceed as follows:
a
After adding a LAN agent, double-click the name and computer
icon/image for the agent you have just connected.
b After naming a WAN agent, proceed as you would for an agent
you’d already selected.
4 The Agent Connection Properties or the Add WAN Agent dialog appears,
depending upon whether you are configuring an already selected agent or
adding a WAN agent (apart from the name in the name bar of the dialog,
these two dialogs are the same).
5 In the Firewall area of the dialog, check the Use HTTP proxy server
option.
50 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Remote Agents
Connecting through
SOCKS firewalls
6 You must then enter the address for the firewall in the Address field. The
address and port number to be used can be obtained from your system
administrator.
7 In the corresponding Port field to the right of the Address field, you must
add the port number through which SilkPerformer is to tunnel.
8 Click OK.
Procedure To configure how the controller connects to remote agents
through SOCKS firewalls:
1 From the SilkPerformer menu bar select Settings/System.
2 The System Settings-Workbench dialog opens to the Agent Pool tab.
3 In the Agent Pool tab you can choose to configure one of the agents you
have already connected by choosing that agent from the window.
Double-click the name and computer icon/image for the agent you have
already connected. If you have not yet selected and named an agent to
connect to, click the Add LAN Agent or Add WAN Agent button and
proceed as follows:
a
After adding a LAN agent, double-click the name and computer
icon/image for the agent you have just connected.
b After naming a WAN agent, proceed as you would for an agent
you’d already selected.
4 The Agent Connection Properties or the Add WAN Agent dialog appears,
depending upon whether you are configuring an already selected agent or
adding a WAN agent; apart from the name in the name bar of the dialog,
these dialogs are the same.
5 In the Firewall area of the dialog, check the Use SOCKS proxy server
option.
6 You must then enter the address for the firewall in the Address field. The
address and port number to use can be obtained from your system
administrator.
7 In the corresponding Port field to the right of the Address field, add the
port number through which SilkPerformer is to listen.
8 Click OK.
SilkPerformer User Guide 51

2 CONFIGURING SILKPERFORMER
Client Certificates
Connecting through
multiple firewalls
Procedure To configure how the controller connects to remote agents
through multiple firewalls:
1 To configure a SilkPerformer agent connection so that it uses a SOCKS
proxy and an HTTP proxy in a series, you must indicate the address and
port number for each. The agent will first connect to the SOCKS proxy
and then connect to the HTTP proxy. From the SilkPerformer menu bar
select Settings/System.
2 The System Settings-Workbench dialog opens to the Agent Pool tab.
3 In the Agent Pool tab you can choose to configure one of the agents you
have already connected by choosing that agent from the window.
Double-click the name and computer icon/image for the agent you have
already connected. If you have not yet selected and named an agent to
connect to, click the Add LAN Agent or Add WAN Agent button and
proceed as follows:
a
After adding a LAN agent, double-click the name and computer
icon/image for the agent you have just connected.
b After naming a WAN agent, proceed as you would for an agent
you’d already selected.
4 The Agent Connection Properties or the Add WAN Agent dialog appears,
depending upon whether you are configuring an already selected agent or
adding a WAN agent (apart from the name of the dialog, these dialogs
are the same).
5 In the Firewall area of the dialog, check the Use SOCKS proxy server
option and the Use HTTP proxy server option.
6 You must then enter the address and port number for the firewall in the
corresponding Address and Port fields for each of the selected options.
The address and port number to use can be obtained from your system
administrator.
7 After entering the information in both sets of fields, click OK to
complete the configuration.
Client Certificates
A secure Web server can be configured to refuse clients that do not have a
proper client certificate from a trusted certification authority. In such cases, the
server asks the client to send its certificate before performing any requests. As
long as the client certificate is signed by a trusted CA, the server does not care
what the client is. If the Web server requests a client certificate, it is necessary to
52 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Client Certificates
Installing client
certificates
Using a client
certificate for
recording
export the client certificate from the Web browser and place it in
SilkPerformer’s certificate store to be used for recording and replay.
Procedure To install a client certificate:
1 From the menu bar, select Settings/Active Profile.
2 In the shortcut list on the left side, click the Internet icon.
3 Switch to the Security tab.
4 In the Client certificate area click Import and locate a certificate file of
the formats .p12, .pfx, or .pem (e.g., C:\clientcert\ johndoe.pfx).
Enter the password you used for encryption when exporting the
certificate from your Web browser.
Note: You have to export a client certificate from your Web browser
before the SilkPerformer Recorder can import it.
5 Click OK. This confirmation also creates a certificate file (.pem) in the
SilkPerformer user data directory (e.g., \Working\Data\johndoe.pem).
Note: The certificate's password/pass phrase is set to "Performer"
(without the enclosing comma) when using SilkPerformer's Client
Certificate Import facility.
Each proxy connection set up in the SilkPerformer Recorder can be
associated with a client certificate. The SilkPerformer Recorder presents
the client certificate to the Web server on behalf of the actual client.
Procedure To use a client certificate for recording:
1 From the menu bar, select Settings/System.
The System Settings – Workbench dialog opens.
2 In the shortcut list on the left, click the Recorder icon.
3 In the System Settings - Recorder dialog, click the Proxies tab.
4 Click Add to set up a new proxy connection or select an existing proxy of
your choice to be associated with a client certificate (e.g., SOCKS) and
click the Edit button.
5 Select an appropriate certificate from the Certificate drop-down list in
the Client Certificate area (e.g., John Doe).
If you have not imported any client certificates so far, follow the
procedure for installing a client certificate, as shown above.
6 Click OK.
7 Restart the SilkPerformer Recorder before beginning recording.
The SilkPerformer replay engine presents the client certificate to the Web server
on behalf of the actual client.
SilkPerformer User Guide 53

2 CONFIGURING SILKPERFORMER
Client Certificates
Replay within the GUI
Replay using the
scripting function
Handling client
certificates with the
Web browser
Exporting client
certificates
Procedure To use a client certificate for replay using the SilkPerformer
GUI:
1 From the menu bar, select Settings/Active Profile.
2 In the shortcut list on the left, click the Internet icon
3 Switch to the Security tab.
4 In the Client certificate area select an appropriate certificate from the
Certificate drop-down list (e.g., John Doe).
If you have not imported any client certificates so far, follow the
procedure for installing a client certificate, as shown above.
5 Click OK.
Procedure To use a client certificate for replay using the scripting function:
1 Use the bdl function SslSetClientCert(). For example,
SslSetClientCert(johndoe.pem).
A secure Web server can be configured to refuse clients that do not have a
proper client certificate from a trusted certification authority. If a Web server
requests a client certificate, it is necessary to export the client certificate from
the Web browser and place it in SilkPerformer’s certificate store.
Since the SilkPerformer Recorder works as a proxy between a client and a Web
server, and the replay engine represents the simulated client itself, successful
connection to the server requires a certificate from the client. Therefore the
certificate must be exported by the Web browser and imported into
SilkPerformer.
Note When obtaining a client certificate from a certification
authority, be sure to allow the private keys to be exported with the
certificate. Otherwise, the certificate will be useless for the Secure
Internet Recorder.
Internet Explorer 6 Procedure To export a client certificate from Internet Explorer 6:
1 From Internet Explorer’s menu bar, select Tools/Internet Options.
The General tab of the Internet Options dialog opens.
2 Click the Content tab.
3 In the Certificates area, click the Certificates button.
The Certificate Manager dialog opens.
4 From the list, select the certificate to export then click Export.
Internet Explorer automatically starts the Certificate Manager Export
Wizard. Using the wizard, perform all steps described to export your
54 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Java Configuration
client certificate.
Note: You have to export the private key with the certificate for the
certificate to be usable with the SilkPerformer Recorder.
5 In the Certificate Manager dialog, click Close.
6 Click OK to close the Internet Options dialog.
The certificate file (.pfx) can now be imported by the SilkPerformer Recorder.
Netscape 7 Procedure To export a client certificate from Netscape 7:
1 From the Netscape menu bar, select Edit/Preferences.
The Preferences dialog opens.
2 Select Privacy & Security/Certificates from the menu on the left side.
The Certificates dialog opens.
3 Click the Manage Certificates button.
The Certificate Manager dialog opens.
4 Click the Backup button.
5 In the File Name To Backup dialog, enter a name of the certificate file.
6 In the following dialog, enter the password protecting your Netscape
software security device.
7 Afterwards, fill out the Choose a Certificate Backup Password dialog by
entering the password used for encrypting your certificate.
8 Click OK to close the Netscape dialog.
Java Configuration
General Java
environment settings
Java Environment configuration is managed by SilkPerformer’s system settings
and individual project profile settings. Both system settings and profile settings
offer the same Java configuration options. System settings serve as default
settings for new projects and new profiles within existing projects. However,
during load tests virtual users use the Java Environment configurations of the
profiles to which they are assigned. For detailed information about Java profile
settings, please refer to “Java options”.
Note See “Java Integration” for information on setting up Java
testing projects using SilkPerformer’s Java Framework.
The Java SDK version, its location on your hard drive, and the class path can be
configured using general Java settings. When a class path contains well-known
locations such as a project directory or data directory, these path parts are
automatically replaced with placeholders. Placeholders are useful when
SilkPerformer User Guide 55

2 CONFIGURING SILKPERFORMER
Java Configuration
migrating projects or running virtual users on agent machines where the
directories have different names than on the controller machine.
Advanced settings
Procedure To configure general Java system settings:
1 Select System from the Settings menu.
2 Click the Java group icon.
3 On the General tab, select the version number of the Java virtual
machine from the Version drop-down list.
4 Click the Browse (...) button to the right of the Java home field to select
the installation directory of the Java Virtual Machine.
5 In the Classpath field, specify the class path of the Java Virtual Machine.
By default the class path is set to the system class path.
6 Click the File button to navigate to a class file and add it to the class path
for the Java Virtual Machine.
7 Click the Dir button to navigate to a directory and add it to the class path
of the Java Virtual Machine.
8 (Optional) Click the Up and Dwn buttons to move selected files up or
down incrementally in the class path hierarchy. Reorder the files as
necessary.
9 (Optional) Click the Remove button to remove a selected file from the
class path.
10 Select the Use system classpath checkbox to ensure that the system class
path is used.
11 Click OK to save your settings.
Some Java classes require specific Virtual Machine options to run successfully.
Tuning parameters may also be specified in VM parameters. The JIT (Just-In-
Time) compiler can be disabled when required and non-standard VM DLLs can
be specified. To ensure that the Java environment is set up properly, the
configuration should be tested.
Procedure To configure advanced Java settings:
1 Select System from the Settings menu.
2 Click the Java group icon.
3 Select the Advanced tab.
4 In the Command line options field, enter any command line options that
are to be passed to the Java Virtual Machine.
5 Select the Disable JIT compiler checkbox if you want to disable the Just
In Time (JIT) compiler of the Java Virtual Machine.
56 SilkPerformer User Guide

2 CONFIGURING SILKPERFORMER
Java Configuration
Adding data files
6 Click the Browse (...) button to the right of the JVM dll field to navigate
to the Dynamic Link Library (.dll) that implements the Java Virtual
Machine that is to be used.
7 Click the Check JVM button to verify your Java Virtual Machine
configuration settings.
8 Click OK to save your settings.
When data files have typical Java archive file extensions, such as .jar and .zip,
SilkPerformer Workbench offers to add the files to the class path of the currently
active profile.
SilkPerformer User Guide 57

2 CONFIGURING SILKPERFORMER
Java Configuration
58 SilkPerformer User Guide

3
Configuring 3Chapter
Projects
Introduction
What you will learn
This chapter explains how to configure SilkPerformer load test projects.
This chapter contains the following sections:
Section
Page
Overview 59
Project Profile Settings 60
Test Scripts 119
Data Files 120
Agent Computers 121
Workload Definition 121
Outlining Projects 125
Overview
A SilkPerformer project contains all the resources that are required to complete
a load test. These include a workload, one or more profiles and test scripts, data
files, and a specific number of agent computers. Options for all of these
resources are available directly in SilkPerformer’s project tree view.
“Project Profile Settings”: When a project is set up, profiles can be added,
copied, renamed, or deleted; and options can be specified for recording traffic,
replaying load test scripts, and generating test results. Options for the different
kinds of network traffic that can be simulated are also available.
“Test Scripts”: Load test scripts can be added to or removed from projects, and
scripts can be compiled.
SilkPerformer User Guide 59

3 CONFIGURING PROJECTS
Project Profile Settings
“Data Files”: It is important that all user data files (.csv), certificate files (.pem),
random data files (.rnd), .idl files, and type library files (.tlb, .dll, .exe, .ocx, .olb,
.pkg) required by test scripts be added to projects. Otherwise, they will not be
available to remote agent computers during tests.
“Agent Computers”: Agent computers on which load tests are run can be
added—and existing agents can be removed.
“Workload Definition”: Options for the workloads to be employed in load tests
can be defined.
Once you’ve configured load test properties, you’re ready for the first step in
creating a load test: “Outlining a project”.
Project Profile Settings
A project profile contains many important project-specific settings. A project
may contain as many different profiles as necessary, each with its own settings.
New profiles can easily be added to projects; and existing profiles can be copied,
renamed, or deleted.
In each profile, options can be set for how the SilkPerformer Recorder generates
test scripts from recorded traffic, and for the protocols that are used during
recording. Simulation settings can be defined for script replay. Options for the
results files that are generated during tests can be defined. Options are also
available for the various types of network traffic that can be simulated: Internet,
Web, CORBA/IIOP, COM, TUXEDO, Jolt, Database, Oracle Forms, Citrix,
.NET, ADO, WAP, and Java.
Specified settings are associated with specific load test projects. It is easy to
switch between different projects, to edit projects, and to save projects so that
they can be modified and reused later.
Administering Profiles
Adding profiles
Procedure To add a profile:
1 On the Project tab of SilkPerformer’s tree view window, right-click the
Profiles folder.
2 Select New Profile.
3 The New Profile dialog appears. Enter a name for the new profile and
click OK. The Profiles folder then expands, with the new profile
appearing in the tree view.
Alternative Select Project/New Profile from SilkPerformer’s menu
bar.
60 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Activating profiles
Editing profiles
Renaming profiles
Deleting profiles
Copying profiles
Procedure To activate a profile:
1 In the tree-view area, right-click the icon of the profile that you want to
have set as the active profile.
2 Select Set as Active Profile from the context menu. The active profile
appears in bold text.
Procedure To edit a profile:
1 In the tree-view area, right-click the icon of the profile for which you
want to edit settings.
2 Select Edit Profile from the context menu.
The Profile Settings dialog appears.
3 Edit settings as required. For help on editing settings, see specific help
topics in this guide.
4 Click OK to exit the dialog and save your changes.
Procedure To rename a profile:
1 In the tree-view area, right-click the icon of the profile that you want to
rename.
2 Choose Rename Profile. The Rename Profile dialog appears.
3 Type a new name for the profile.
4 Click OK.
Procedure To delete a profile:
1 In the tree-view area, right-click the icon of the profile you want to
delete.
2 Choose Delete Profile.
3 Click Yes on the confirmation dialog to delete the profile.
Procedure To copy a profile:
1 In the tree-view area, right-click the icon for the profile you want to
copy.
2 Select Copy Profile from the context menu.
3 A new profile then appears in the tree view, with the name Copy of
.
SilkPerformer User Guide 61

3 CONFIGURING PROJECTS
Project Profile Settings
Recording options
General options
Procedure To set general record options:
1 In the Project tab of SilkPerformer’s main tree view, expand the Profiles
folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu. The Profile - [Loadtest] -
Simulation dialog opens.
Alternative Select Settings/Active Profile from the SilkPerformer
menu bar.
4 In the shortcut list on the left side, select the Record category and click
the Script icon.
5 On the General tab, select the Comments option to have the
SilkPerformer Recorder automatically insert comments into the
generated script. Comments are left-delimited by // symbols.
6 Select the Commented functions option to mark in the generated test
script the function calls that are performed to retrieve return parameters
as comments.
7 Select the Include think time option to generate ThinkTime function calls
in the recorded script. These function calls produce delays that simulate
typing time and thinking time in the recorded script to create a more
realistic simulation.
8 Enter a time (in milliseconds) in the Min. think time recorded field to
define a lower limit for delays that will be included as ThinkTime
function calls in the recorded script. Delays less than the time specified
in the Min. think time recorded field will not expanded to ThinkTime
function calls.
9 In the Max. line length field, specify the maximum number of characters
in a script line. This setting can be used to split large amounts of
recorded data into several script lines.
10 (optional) In the Script namespace field, you can enter a prefix for the
names that are declared in the generated script. The user groups,
transactions, timers, and variables defined in the script begin with the
specified string.
11 Select the Record passwords encrypted option to have the SilkPerformer
Recorder automatically encrypt recorded passwords before inserting
them into generated scripts.
62 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Protocol-specific
options
Procedure To set protocol-specific record options:
1 Click the Protocols tab.
2 In the Web comments area of the tab, specify details about automatically
inserted comments (for Web functions) in the generated script. The
settings for these options are only taken into account when the Comments
option (General tab) for script generation is enabled.
- Select the Redirection / Authentication option to have comments
inserted into the generated script when a redirection (HTTP 302,
HTTP 407) from one URL to another is recorded.
- Select the HTTP errors option to have comments inserted into the
generated script when HTTP errors (HTTP 4xx, HTTP 5xx) are
recorded.
- Select the Link, form, custom URL search details option to have
comments inserted into the generated script when links, forms, or
customs URLs are recorded.
- Select the Detailed info for form fields option to have comments
inserted into the generated script when hidden, changed, or filled out
form fields are recorded.
- Select the Custom URL parsing details option to have comments
inserted into the generated script when parsing details of custom
URLs are recorded.
3 In the IIOP area of the tab, select the Generate IIOP Get functions option
to include functions in your test script that will retrieve the return
parameters of CORBA operation calls. You can mark the corresponding
function calls as comments.
4 Use the Database area of the tab to specify how SQL statements are
displayed and whether to include fetched data in the generated test script.
Select the SQL comment option to include SQL statements marked as
comments in the generated test script immediately before the parse
operation.
5 Select the Trim SQL option to display SQL statements in the generated
test script in a more readable format, with line breaks and indentation for
SQL keywords.
6 Select the Fetched data option to include the fetched data, marked as
comments, in the generated test script immediately after the fetch
operation.
7 In the Max. number of rows field, enter the maximum number of data
rows that are to be inserted as comments immediately after fetch
operations.
SilkPerformer User Guide 63

3 CONFIGURING PROJECTS
Project Profile Settings
8 In the Max. column width field, enter the maximum width of columns
inserted as commentary.
Replay options
Simulation options
Procedure To set simulation options:
1 In the Project tab of SilkPerformer’s tree-view, expand the Profiles
folder.
2 Right-click the icon of the profile you want to configure.
3 Select Edit Profile from the context menu.
The Profile - [Profile1] - Simulation dialog opens at the Replay category,
Simulation tab.
Alternative Select Settings/Active Profile from the SilkPerformer
menu bar.
4 Choose the Stress test option to disable the use of wait periods that are
specified in browser-level functions or invoked by calling ThinkTime
functions.
5 Select the Random thinking time option to replace recorded actual
thinking periods that are specified in browser-level functions or invoked
by calling ThinkTime functions, with random values that follow an
exponential distribution.
6 Select the Think time limited to option to define a limit for think times. If
this option is selected, think times will be limited to the specified
maximum value. Enter an upper limit (in seconds) in the provided field.
7 Select the Smooth transaction arrival rate (Queuing Workload) option to
have each virtual user perform its transactions within the specified
simulation interval. With this option selected, the simulation time period
and the total number of transactions to be called determine the mean
arrival interval of the transactions.
When this option is disabled, SilkPerformer calculates the delay of each
transaction arrival randomly.
8 Select the Choose transactions randomly option to force the
SilkPerformer simulation engine to execute all transactions defined in
load test scripts in random order, based on SilkPerformer's queuing
model.
When this option is disabled, simulated virtual users perform their
transactions exactly in the order in which they are defined in test scripts.
64 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Error-handling options
9 Select the Stop virtual users after simulation time (Queuing Workload)
option to cancel users when simulation time expires.
10 Select the Call end transactions for stopped virtual users option to have
stopped virtual users execute end transactions before stopping.
When this option is disabled, virtual users complete their executions
without calling end transactions.
11 Select the Complete current transactions of stopped virtual users option
to allow stopped virtual users to complete active transactions before
stopping. When this option is selected, virtual users check whether to
stop only at the beginning of each new transaction. Thus active
transactions are completed before users stop.
When this option is disabled, virtual users check if they should stop
during think and wait times, while in the queue, and at the beginning of
each transaction.
12 Select the Stop virtual users when errors occur in begin transactions
option to immediately stop virtual users when errors occur during test
initialization (which usually occurs in begin transactions). This setting
overrides the Call end transactions for stopped virtual users and
Complete current transactions of stopped virtual users options.
This setting is useful for tests that involve Connect operations that must
be successful before tests can proceed.
Procedure To set error-handling options:
1 In the Project tab of SilkPerformer’s tree-view, expand the Profiles
folder.
2 Right-click the icon of the profile you want to configure.
3 Select Edit Profile from the context menu.
The Profile - [Profile1] - Simulation dialog opens at the Replay category,
Simulation tab.
4 Click the Errors tab.
5 Use the Severity definition area to modify the severity of the simulation
errors that SilkPerformer reports in its Monitor window.
6 Click the Add button to add a custom severity definition to the list.
The Edit Replay Errors and Warnings dialog appears.
7 From the Module drop-down list, select the module that may cause the
simulation error for which you want to define a custom level of severity.
All errors of the selected severity are displayed below in the list.
8 From the Severity drop-down list, select the default severity of the error
for which you want to define a custom level of severity:
SilkPerformer User Guide 65

3 CONFIGURING PROJECTS
Project Profile Settings
- success - Error is ignored
- informational - Error is ignored, but reported
- warning - Error is ignored, but causes a warning
- error - Error is treated as an error
Note All errors of each selected severity are displayed in the below
field.
9 In the Error code field, enter the code of the error for which you want to
define a custom level of severity. Alternately, you can select the
respective error from the list above the field.
10 In the Description field, enter a description for native errors.
Descriptions of other simulation errors are predefined and cannot be
modified by the user.
11 From the Custom severity drop-down list, select the custom severity for
the specified error:
- success - Error is ignored
- informational - Error is ignored, but reported
- warning - Error is ignored, but causes a warning
- error - Error is reported, but does not affect the simulation
- transaction exit - Current transaction is aborted
- process exit - Simulation of the user is terminated
12 Click OK.
13 To edit the custom severity of an error, select it from the error list and
click the Edit button.
The Edit Replay Errors and Warnings dialog appears.
14 Edit the custom severity of the error as explained earlier in this
procedure.
15 To remove a custom severity definition from the list, select it from the
error list and click the Remove button.
16 Click OK to exit the dialog and save your changes.
Result options
General options
Procedure To set general result options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
66 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the General tab.
- Use the Results files area to specify the types of results files you
want to generate. While the test is running, a separate file is created
for each user in the current agent. Since generating results files will
alter the time measures for your load test, these files should be
created for debugging purposes only.
- Select the Log files (.log) option to generate log files (.log), one for
each user in the current agent. These files contain all the function
calls invoked by the transactions of a specific user.
- Select the Report files (.rpt) option to generate report files (.rpt), one
for each user in the current agent. These files contain the simulation
results for each user.
- Select the Output files (.wrt) option to generate output files (.wrt),
one for each user in the current agent. These files contain the output
of write statements used in the test script. An output file is generated
for a particular user only if that user executes write statements.
- Select the True log files (.xlg) option to create an accurate view of
data downloaded during a load test to be made available by means of
TrueLog Explorer.
- Select the Time series data (.tsd) option to enable generation of the
throughput information used by SilkPerformer's Performance
Explorer. You can specify the interval at which data is computed.
- Select the Client-side measures in real-time option to enable
SilkPerformer to retrieve real-time client-side measurements for use
in Silk Performance Explorer. See the Silk Performance Explorer
User Guide for full details.
7 Use the Timer statistics area to specify time bounds, or limits, for use in
results reporting. Several reports list the percentage of all the
transactions completed within the time bound specified.
SilkPerformer User Guide 67

3 CONFIGURING PROJECTS
Project Profile Settings
Time series data
8 In the Time bound 1 field, enter a first time bound, or limit, for use in
results reporting. Several reports list the percentage of all the
transactions completed within the time bound specified. This time bound
is commonly set to the amount of time that represents the goal for the
completion of the transaction being tested.
9 In the Time bound 2 field, enter a second time bound, or limit, for results
reporting. Several reports list the percentage of all the transactions
completed within the time bound specified. This time bound is
commonly set to the amount of time that should not be exceeded for the
completion of the transaction being tested.
Procedure To set options related to the generation of time series data:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the Time Series tab.
7 Use the Time series data area to specify options for the information used
by SilkPerformer's Performance Explorer.
8 From the Computation interval drop-down list, select the interval at
which you want data to be calculated. This setting is used for all time
series data generated. Keep in mind that the shorter the interval and the
longer the load test, the more data will be generated.
9 Select the Enable all measure groups option to generate high-level
throughput information for all the measure groups shown in the list
below. To choose a particular combination of measure groups for which
you want to generate high-level throughput information, deselect the
Enable all measure groups option, and select your options in the
measure group list.
10 Specify a Detailed Page Timers generation setting (For the whole page,
For documents only, or Disabled).
68 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
11 To define an overview report template for generating overview reports
from Performance Explorer for this load testing project, click the browse
(...) button in the Overview report template portion of the Time Series
tab. From the Select template file dialog, select the overview report
template (.ovt) file that you want to have applied to your SilkPerformer
projects.
The file that you specify here will be used as an overview report template
for the generation of all overview reports for this load testing project.
The template file name will also be preselected on the Overview Report
Template workflow dialog when new overview reports are created
manually.
Monitoring options
Procedure To set monitoring options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the Monitoring tab.
7 In the Monitoring options area, select the Automatically start monitoring
option to automatically launch the monitoring facility of Performance
Explorer when you start a load test. Performance Explorer will then
display the server performance data relevant to the type of server that
you are testing.
- Select the Use default monitoring template option to activate the
default monitor template.
- Select the Use custom monitoring template option to use a custom
monitor template.
8 In the Performance monitor integration area, select the Compute online
performance data option to compute data for additional performance
measurements to be displayed in the Windows NT/2000 Performance
Monitor. This data can be used to view the concurrent users, transaction
throughput, data sent and received, and SQL statements executed.
SilkPerformer User Guide 69

3 CONFIGURING PROJECTS
Project Profile Settings
TrueLog options
Procedure To set TrueLog options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the TrueLog tab.
7 To enable generation of TrueLog files (.xlg) during the recording
session, select the TrueLog files (.record.xlg) checkbox. TrueLog files
contain all traffic that is sent to and received from the server, and the
transactions and function calls that are to be generated in the test script.
Note TrueLog files should be created for debugging purposes only.
8 Select the TrueLog files (.xlg) checkbox to generate a TrueLog file (.xlg)
for each virtual user. These files contain all the function calls that are
invoked by the transactions of a specific virtual user.
9 To enable generation of TrueLog On Error files (.xlg), select the
TrueLog On Error files (.xlg) checkbox.
Note Truelog On Error files can be turned on for even large load
tests where you expect a moderate number of errors without major
impact on replay performance.
10 Select the High-resolution timers for page statistics checkbox to enable
SilkPerformer to use high-resolution timers for page statistics. Note that
you can use the WebSetOption(WEB_OPT_HIGH_RES, x) function to
adjust the resolution of these timers (the default is 100us).
11 The options in the TrueLog On Error area of the dialog are only
available when TrueLog On Error file generation has been enabled
above.
Specify how much information is to be stored in the TrueLog using the
Store TrueLog options.
- Select For one transaction to log the entire transaction in which each
error occurrs. This option requires a significant amount of memory.
70 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
General logging
options
- Select Based on content history to log a minimum of relevant data
(based on data kept in the history) to the TrueLog. If browser-level
calls are used, this setting will have no impact on the size of stored
data. This selection requires minimal memory.
12 From the Generate TrueLog On Error for drop-down list, select the
severity level at which TrueLogs are to be generated (Errors, Severe
errors, or Warnings). The default severity level is Errors.
13 Select the Log all embedded objects option to log all embedded objects
(images, java files, etc) in the TrueLog. Logged objects appear in the
cached documents section of the TrueLog.
14 Select the Exclude cached embedded objects option to only log
embedded objects (images, java files, etc) that have been downloaded
within the scope of the current log in the TrueLog. Objects may not be
logged if they appear in the current log as a cache hit. Such objects may
have been downloaded outside the scope of the current log. They may
also no longer be present in the cache.
15 Select the Exclude all embedded objects option to exclude all embedded
objects (images, java files, etc) from being logged in the TrueLog. Note
that optimal scalability is achieved by selecting this option.
16 Click OK to save your settings.
Procedure To set general logging options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the Logging tab.
7 Select the Log all option to include all available types of information in
the generated log files. These files contain information about timers,
transactions performed by the user, and function calls invoked by the
transactions. However, keep in mind that creating log files can alter the
time measures of the load test.
SilkPerformer User Guide 71

3 CONFIGURING PROJECTS
Project Profile Settings
Internet-specific
logging
8 Use the General area to specify which types of information to include in
log files.
- Select the Transactions option to include in the log file an overview
of the transactions the user performs during the simulation.
- Select the Timers option to include custom timer information in the
log file.
- Select the API calls option to include in the log file an overview of
the SilkPerformer API calls that the user performs during the
simulation.
- Select the Informational option to include additional information—
for example, connection information—in the log file.
- Select the Data option to include in the log file the data that was
exchanged with the server.
- Select the Nonprintable characters option to include nonprintable
characters, represented as dots, in the log file.
Procedure To set Internet-specific logging options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the Internet logging tab.
7 Use the Data sent to server area to specify the level of detail at which
data that SilkPerformer and the Recorder send to the server are written to
the log file.
- Select the HTTP client request header option to write to the log file
all the client request headers sent to the server. This option is only
available when recording HTTP traffic (not for replay).
- Select the Header option to write to the log file all the headers
included in requests that are sent to the server.
- Select the Printable data option to write to the log file all the
printable data included in requests that are sent to the server.
72 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
ARM options
- Select the Binary data option to write to the log file all the binary
data included in requests that are sent to the server.
- In the Max. number of lines field, specify the maximum number of
lines to be written to the log file in the case of lines containing
request data that are sent to the server.
8 Use the Data received from server area to specify the level of detail at
which the data that SilkPerformer and the Recorder receive from the
server is written to the log file.
- Select the HTTP response header to client option to write the client
HTTP response headers (received from the server) to the log file.
This option is only available when recording HTTP traffic (not for
replay).
- Select the Header option to write to the log file all the headers
included in responses that are received from the server.
- Select the Printable data option to write to the log file all the
printable data included in responses that are received from the
server.
- Select the Binary data option to write to the log file all the binary
data included in responses that are received from the server.
- In the Max. number of lines field, specify the maximum number of
lines to be written to the log file in the case of lines containing
response data that are received from the server.
9 Select the Images, applets, and Java classes option to write to the log file
all the images, applets, and Java classes received from the server. If this
option is selected, these items are represented as binary data.
10 Select the TCP/IP connection information option to write additional
information to the log file; for example, when connections to the server
are established or closed.
Procedure To set ARM options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
SilkPerformer User Guide 73

3 CONFIGURING PROJECTS
Project Profile Settings
Recorder logging
options
5 Select the Replay category.
6 Click the ARM tab.
7 Use the Settings area to specify the types of ARM transactions that are to
be reported.
8 Select the Enable ARM for transactions checkbox to enable ARM
transaction reporting for BDL transactions.
9 Select the Enable ARM for custom timers checkbox to enable ARM
transaction reporting for custom timers.
10 Select the Enable ARM for page timers checkbox to enable ARM
transaction reporting for page timers.
Procedure To set SilkPerformer Recorder logging options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Results icon.
5 Select the Replay category.
6 Click the Hook logging tab.
7 Select the Record category.
8 Use the Hooking area to display—in SilkPerformer Recorder’s Log
tab—all the function calls that the client application performs and the
parameters that the client application passes to function calls.
- Select the Function calls option to display—in SilkPerformer
Recorder’s Log tab—all the function calls that the client application
performs. You can also specify whether to display the function
parameters.
- Select the Function parameters option to display—in SilkPerformer
Recorder’s Log tab—all the parameters that the client application
passes to function calls. To enable this option, you must first select
the Function calls option.
9 Click OK to exit the dialog and save your changes.
74 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Real-time client-side
measure options
Procedure To enable real-time client-side monitoring for Silk Performance
Explorer:
1 Go to Settings/Active Profile.
2 Select the Results group icon (in the Replay group).
3 On the General tab, ensure that the Client-side measures in real-time
checkbox is selected.
4 Select the Real-time measures tab.
5 Set the Computation interval to 5 sec.
6 Select the Enable all measure groups checkbox to enable monitoring of
all measure groups. Alternately, you can select specific measure types in
the field below.
7 Click OK to save your settings.
Internet options
Network options
Procedure To set network-related options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Internet icon.
5 Select the Replay category.
6 Click the Network tab.
7 Use the Modem area to specify which type of network device is to be
simulated for Internet access.
8 From the Bandwidth drop-down list, select the bandwidth of the network
device to simulate for Internet access. SilkPerformer causes each virtual
user to transfer data both to and from the server at the rate of the
specified device. Note that, as with a real Web browser, multiple threads
share the selected bandwidth. To use duplex bandwidth, where
downstream and upstream bandwidths are different, select the Custom
option.
SilkPerformer User Guide 75

3 CONFIGURING PROJECTS
Project Profile Settings
9 Check the Duplex option to set a different bandwidth for downstream
and upstream traffic when you select a custom bandwidth of the network
device to simulate for Internet access. Use the Downstream and
Upstream fields to set the desired bandwidth.
10 In the Downstream and Upstream fields, enter values [bits/sec] for the
downstream and upstream bandwidth when you want to use the duplex
option.
11 From the Modem compression rate drop-down list, select the hardware
compression rate of the simulated network device. SilkPerformer causes
each virtual user to transfer data both to and from the server at the
specified compression rate.
12 From the Optimize network settings for drop-down list, select the type of
simulation for SilkPerformer and WinSock buffers. This setting
determines the size of the buffers used to transfer data both to and from
the server. Large buffers provide good performance; however,
SilkPerformer needs more memory to achieve this performance. Small
buffers may be useful on driver machines with less memory. Note that,
when using modem simulation, small buffers will create a more realistic
load on the server.
13 Use the Network area to specify network-specific settings.
14 Select the Client IP address multiplexing option to enable IP address
multiplexing. To assign individual IP addresses to virtual users, you first
have to set up multiple IP addresses for the network device. You do this
with the SilkPerformer System Configuration Manager. Afterwards, you
can assign each of the IP addresses to a different virtual user.
15 Select the Host name cache option for SilkPerformer to use its internal
host name cache. If this option is selected, SilkPerformer stores the IP
addresses of resolved host names and therefore does not need to contact
the DNS server each time the replay engine requires a host's IP address.
16 Select the Passive FTP semantics option to enable passive FTP
semantics, which may be useful when connecting through firewalls or
proxies. This option affects FTP data transfers only.
17 In the Connect attempts field, enter how often a simulated user will
attempt to connect to the server before an error is reported.
18 In the Connect timeout field, specify for how long the simulated user will
attempt to connect to the server before a timeout is reported.
19 In the Send timeout field, specify for how long the simulated user will
attempt to send data to the server before a timeout is reported.
76 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
20 In the Receive timeout field, specify for how long the simulated user will
attempt to receive data from the server before a timeout is reported.
Host options
Security options
Procedure To set host-related options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Internet icon.
5 Select the Replay category.
6 Click the Hosts tab.
7 In the Standardhost area, specify the server you want to use as the
standard host. You can use either a domain name or an IP address. The
name you specify will be used whenever the "standardhost" variable is
included in load test scripts. Optionally, you can use the Port field to
specify a port number to use with the standard host.
8 In the SOCKS4 proxy area, use the Proxy name or IP address and Port
fields to specify the host name or IP address, and also the port number,
for a SOCKS4 proxy through which all Internet traffic will be directed.
This option can be used, for example, when the traffic in a load test has
to negotiate a firewall.
Procedure To set security options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Internet icon.
5 Select the Replay category.
6 Click the Security tab.
SilkPerformer User Guide 77

3 CONFIGURING PROJECTS
Project Profile Settings
Recording options
7 Use the Protocol area to specify protocol-specific settings used for
secure connections to remote servers.
8 From the SSL version drop-down list, select the Secure Socket Layer
protocol version used when establishing a secure connection to the
server.
9 From the Encryption strength drop-down list, select the encryption
strength for data transferred to and from the server over a secure
connection.
10 From the Certificate drop-down list in the Client certificate area, select
the client certificate that SilkPerformer is to present to the Web server on
behalf of the actual client. To use a client certificate, export it from the
Web browser and place it in SilkPerformer's certificate store first.
11 To view detailed information about the client certificate currently in use,
click the View button.
12 To import a new client certificate from a certificate file (.p12, .pfx, or
.pem), click the Import button.
13 Use the following dialog to navigate to and select the new client
certificate.
14 To remove the currently selected client certificate from the drop-down
list, click the Remove button.
15 Select the Force secure connection option to use the Secure Socket
Layer protocol for replaying load tests. This is equivalent, for example,
to replacing all the HTTP schemes with HTTPS in the URL’s of
browser-level functions. It can be used to simplify how load tests are
used to compare performance with and without the use of secure
connections.
Procedure To set recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Internet icon.
5 Select the Replay category.
6 Click the Recording tab.
78 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
7 In the SMTP area, select the Enable SMTP data recording option to
record mail body data exchanged between the client application and the
mail server. To record SMTP traffic, set up an appropriate proxy
connection on the Proxy tab (see “Proxy options” for details).
8 Use the Recorder server certificate area to specify which server
certificate you want the SilkPerformer Recorder to present to the client
on behalf of the actual server.
9 Select the Use custom server certificate option to have the SilkPerformer
Recorder present a custom server certificate to the client instead of the
default Recorder server certificate. Click the browse button (...) to the
right of the field to locate the server certificate you want to use.
10 In the Pass phrase field, enter the pass phrase that is to be used, if the
server certificate you want the SilkPerformer Recorder to use is
protected with a pass phrase.
11 Select the Send root CA during SSL handshake option for the
SilkPerformer Recorder to send the root CA certificate during the SSL
handshake. The root CA certificate may be requested by a client to
authenticate the certificate authority that signed the Recorder server
certificate. Click the browse button (...) to the right of the field to locate
the root CA certificate that you want to use.
12 Click OK to exit the dialog and save your changes.
Web Options
Browser options
Procedure To set browser-related options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Browser tab.
SilkPerformer User Guide 79

3 CONFIGURING PROJECTS
Project Profile Settings
7 Use the Browser type area to specify browser-specific settings. From the
Browser drop-down list, select the Web browser you want to use for your
simulation. The selection you make determines the format of the header
information included in your HTTP requests and the threading model
used for simulation.
8 Select the Limit connections per server option to limit the connections on
a per server basis (as real browsers do), e.g., two connections to one
server, four connections to two servers.
Note It is not recommended to use this option for browser-level
API, as only the Max total connections value is considered.
- In the Max concurrent connections to a HTTP 1.0 server field,
specify the connection limit to every server using HTTP 1.0.
- In the Max concurrent connections to a HTTP 1.1 server field, secify
the connection limit to every server using HTTP 1.1.
- In the Max total connections field, specify the absolute maximum
number of connections, which the emulated browser opens. To
perform an accurate connection simulation this value should be
above the per server setting (above options).]
9 Select the Limit connections per virtual user option to limit the
connections on a per user basis (e.g., four connections to one server, four
connections to two servers).
- In the Max. concurrent connections field, specify the maximum
number of simultaneous connections the client may establish to the
server in the case of custom browsers. Since the server has to
allocate more system resources for each connection, the server's
performance decreases as the number of connections increases. In
the same way as Web browsers, SilkPerformer uses one thread per
connection to send and receive data concurrently. For singlethreaded
(synchronous) operations, set this value to 1. For multithreaded
simulations, choose a value between 2 and 32.
Note Values are pre-defined for each selected Browser.
You can only edit Max. concurrent connections if you select
Custom from the Browser drop-down list.
- Select the Single-threaded option to replay load tests using a single
thread. This setting forces all browser-level functions to execute
synchronously.
10 From the HTTP version drop-down list, select the HTTP version of the
browser to be simulated. This is typically HTTP/1.0.
80 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
11 Select the Keep-Alive semantics option to set Keep-Alive for HTTP/
HTTPS. Keep-Alive is a feature that, when supported by both browser
and server, allows connections to remain intact after a request and
response are completed.
12 Select the Content-Encoding option to enable SilkPerformer to decode
compressed data received from the server.
If this option is enabled, SilkPerformer sends an Accept-Encoding
header to inform the server that encoded data (by using gzip and deflate
compression methods) can be submitted.
13 In the User agent field, specify user agent information sent to the server
in the case of custom browsers. This is typically the browser name,
including platform information. This string is used for the HTTP request
header “User-Agent:” that is sent using SilkPerformer's browser-level
functions.
14 In the Accept field, specify the document types that the server may send
to the client in the case of custom browsers. This string is used in the
HTTP request header “Accept:” that is sent using SilkPerformer's
browser-level functions.
15 In the Additional HTTP headers field, specify one or more additional
headers, using exact syntax in the case of custom browsers. Use a new
line for each header. These headers will be appended to each request that
is sent using SilkPerformer's browser-level functions.
Proxy options
The proxy settings allow you to specify in which way SilkPerformer connects to
the Internet. You can define one of the following methods:
• Direct connection to the Internet: Connect to the Internet without using a
proxy.
• Connect simulated browsers through a proxy server: Connect to the
Internet through a specific proxy server. You can specify separate
proxies for HTTP, HTTPS, FTP, and SOCKS traffic.
• Automatic proxy configuration URL: Connect to the Internet through
dynamic proxy configuration. Dynamic proxy configuration is a
technique used in Web browsers to determine a proxy for each request
sent out from the Web browser to the Internet. Depending on the URL or
the host to connect to, a proxy is dynamically determined.
A PAC (Proxy Auto-Config) file is used for determining the proxy server
for each request. A Proxy Auto-Config file tells the browser to load its
proxy configuration information from a remote JavaScript file, rather
than from static information you enter directly.
For detailed information about the PAC file specification, please refer to
http://wp.netscape.com/eng/mozilla/2.0/relnotes/demo/proxy-live.html.
SilkPerformer User Guide 81

3 CONFIGURING PROJECTS
Project Profile Settings
Procedure To set proxy-related options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Proxy tab.
7 Select one of the following:
- Direct connection to the Internet: Connect to the Internet without
using a proxy.
- Connect simulated browsers through a proxy server: Connect to the
Internet through a specific proxy server.
o Click the Advanced button to open the the Proxy Settings dialog.
o In the Internet access area, configure the Host name or IP
address and Port of the proxy server for each protocol.
Check Use the same proxy server for all protocols to use the
same proxy for all Web traffic. Checking this setting enables
Address and Port fields on the Profile - [Loadtest] - Simulation
dialog.
To change the SOCKS proxy settings, go to the Hosts tab of the
Profile Settings - Internet dialog.
o In the Exceptions area, list any host names, IP addresses, or
subdomains for which you do not want to use the proxies
specified in the Internet access area. This list can contain
wildcards (*) that cause the application to bypass the proxy
server for addresses that fit the specified pattern. To list multiple
addresses and host names, separate them with semicolons.
o Click OK to save your settings.
- Automatic proxy configuration URL: Connect to the Internet through
dynamic proxy configuration. Specify the location of the PAC file to
be loaded.
82 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
8 Click OK to save your proxy settings.
Browser simulation
Procedure To set browser simulation options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Simulation tab.
7 Use the Simulation area to set options for realistic simulation of users
who visit a Web site.
8 Select the Simulate user behavior for each transaction option to order all
virtual users to reset their browser emulation after each transaction.
Depending on the additional option you select, SilkPerformer either
simulates users who visit a Web site for the first time or users who revisit
the site. While users who visit a site for the first time have no persistent
cookies stored and no documents cached, users who revisit a page
typically have closed their browsers between the Web site visits, have
documents cached, and persistent cookies set. Disabling this option lets
the virtual user emulate a Web browser that is not closed until the end of
the load test, thus reusing cached information throughout multiple
transactions.
9 Select the First time user option to generate a realistic simulation of
users who visit a Web site for the first time. Persistent connections will
be closed, the Web browser emulation will be reset, and the document
cache, the document history, the cookie database, the authentication
databases, and the SSL context cache will be cleared after each
transaction. In this case, SilkPerformer always downloads the complete
sites from the server, including all files.
10 Select the Revisiting user option to generate a realistic simulation of
users who revisit a Web site. Persistent connections will be closed, and
the document history, the non-persistent cookie database, the
authentication database, and the SSL context cache will be cleared after
each transaction. In this case, the users do not clear the document cache.
SilkPerformer User Guide 83

3 CONFIGURING PROJECTS
Project Profile Settings
For more details, review the WebSetUserBehavior function in the BDL
Function Reference.
11 Use the User tolerance area to adjust the advanced options of the user
tolerance simulation.
12 Use the slider to adjust the tolerance level of the simulated user.
Click the Customize button to open a dialog that enables you to alter the
individual behavior of the selected user tolerance level.
Advanced User Tolerance dialog:
Use the Loading times tolerance area to specify the user tolerance
regarding the loading times.
- Select the If no data arrives after option to make the virtual user
react if the server does not respond at all in a given time frame.
Also specify the maximum time in seconds a user is willing to wait if
the server does not respond.
- Select the If document is not complete after option to make the
virtual user react if the first document (root) is received, but the
HTML content (frames) is not completely loaded within a given
time frame.
Also specify the maximum time in seconds a user is willing to wait if
the first document (root) is received, but the HTML content (frames)
is not complete.
- Select the If page is not complete after option to make the virtual
user react if the HTML content (documents) is received, but the
embedded objects are not completely loaded within a given time
frame.
Also specify the maximum time in seconds and user is willing to
wait if the HTML content (documents) is received, but the
embedded objects are not completely loaded.
- Select the If the pure image load time exceeds option to make the
virtual user react if the loading time of the embedded objects
exceeds a given time frame (The timer starts, when the last HTML
document is received.
Also specify the maximum time in seconds a user is willing to wait
for the loading of all embedded objects. This time starts, when the
last HTML document is loaded.
- Specify Reload times to define how many times the user presses the
Refresh button when timeouts occur.
84 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
- Select the Continue waiting option if the user is to continue to wait if
the timeout occurs and the page has already been retried the
specified number of times; unlimited or until another timeout occurs.
- Select the Abandon option if the user should press the stop button
(abandon) when a timeout occurs and the page has already been
retried the specified number of times.
- Select the Wait additional option if the user is to wait an additional
amount of time if the timeout occurs and the page has already been
retried the specified number of times. If this time expires too, the
user will press the stop button (abandoning the load).
Also specify the number of seconds that the virtual user is to wait
before pressing the stop button.
Browser
authentication
Use the Error tolerance area to specify user tolerance regarding
errors on a page.
- Select the Retry option to make the virtual user react to errors in the
page.
- Specify how often the user is to press the Refresh button in the case
of errors.
- Select the Page option to have the user press the Refresh button on
any error in the page (e.g., a missing image).
- Select the Documents option to have the user press the Refresh
button only when errors in the document portion of the page. (e.g., a
missing frame) occur.
Use the browser authentication settings if you want all VUsers to login with the
same credentials. If you want each VUser to connect with different login
credentials, we recommend that you specify the credentials with project
attributes (see Project Attributes Overview in the BDL Function Reference for
detailed information) or that you use script customization through data files (see
“Data Files” for detailed information).
Procedure To set browser authentication options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
SilkPerformer User Guide 85

3 CONFIGURING PROJECTS
Project Profile Settings
Browser emulation
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Authentication tab.
7 Use the User authentication area to enter a user name and a password for
user authentication when connecting to a remote server.
8 In the User name field, enter a user name for authentication when
connecting to a protected server.
9 In the Password field, enter a password for authentication when
connecting to a protected server.
10 Use the Proxy authentication area to enter a user name and a password
for proxy authentication login required when connecting to the Internet
through a proxy server.
11 In the Login name field, enter a user name for the proxy authentication
required when connecting to the Internet through a protected proxy
server.
12 In the Password field, enter a password for the proxy authentication
required when connecting to the Internet through a protected proxy
server.
Procedure To set browser emulation options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Emulation tab.
7 Use the Emulation area to set options for realistic emulation of a Web
browser.
8 Select the Accept cookies option to store cookies in SilkPerformer's
cookie database for each virtual user. Note that the cookie database is
non-persistent and exists only as long as a load test is running.
86 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
9 Select the Allow automatic redirections option for SilkPerformer to
automatically parse redirection responses and redirect the request to the
specified server. In addition, if this option is selected, SilkPerformer uses
the user name and password entered on the Authentication tab for
automatic authentication to the server whenever necessary.
In the field to the right of the checkbox, enter the maximum number of
automatic redirections. SilkPerformer automatically parses redirection
responses and redirects the request to the specified server.
10 Select the Automatically load images option for SilkPerformer to
download images from the server. Deselecting this option may be useful
if you are interested in testing dynamic parts of the Web application.
Usually images are static objects that stress the network more than the
server. Note that ignoring images is a trade-off in the realism of the test.
11 Use the Document cache and history area to specify whether to emulate a
document cache. SilkPerformer can emulate a document cache by
keeping local copies of frequently accessed documents, thus reducing the
time connected to the network.
12 Select the Document cache option to have SilkPerformer emulate a
document cache. A document cache is used to keep local copies of
frequently accessed documents and thus reduce the number of hits on the
server.
For more details, review the WebSetDocumentCache function in the
BDL Function Reference.
13 Select the Once per session option to check once per session whether a
page has changed on the Web server since the virtual user last viewed it.
The term of a session is the sequence of starting a Web browser, surfing
the Web, and closing the Web browser. The correlation of a session to
SilkPerformer transactions can be adjusted with the Simulate user
behavior for each transaction option in the Simulation tab.
“Once per session” works as follows: each time you start your Web
browser, it will first look into its cache before sending a request. This
request will either be: a conditional request (if the document resides in
the cache and is requested for the first time in the session); a normal
request (if the document cannot be looked up in the cache); or, a cache
hit (if the document has already been requested in the session).
Select the Every time option to always check if a page has changed on the
Web server since the virtual user last viewed it. Documents with a future
expiration date are not requested until their dates expire.
The “Every time” strategy is slightly different from “Once per session”
caching in that it checks for newer documents each time you reload a
SilkPerformer User Guide 87

3 CONFIGURING PROJECTS
Project Profile Settings
page.
Recording options
Select the Never option to never check whether a page has changed since
the virtual user last viewed it. Selecting this option speeds up the display
of pages the user has already viewed.
14 In the Document history size field, enter the maximum number of pages
to be stored in the history. If this amount is exceeded, the oldest page is
dropped from the history. As long as a page is stored in the history, all
HTML document related information (links, forms, embedded objects,
frames) is kept in memory. As soon as the page is dropped from the
history this information is no longer available. Therefore related
WebPageLink and WebPageSubmit calls do not succeed and subsequent
page-level calls to the same URL do not find the document in the cache
any more.
15 Select the Emulate DNS lookup option to have SilkPerformer emulate a
DNS lookup mechanism for each virtual user, which can be helpful for
DNS-based load balancing. When this option is combined with a user
behavior option, it allows each virtual user to send its individual DNS
query to the DNS server.
For more details, review the WebSetOption function in the BDL Function
Reference.
Procedure To set recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Record category.
6 Click the Recording tab.
7 Use the Browser emulation level area to specify the level at which you
want to record Web traffic for use in your test scripts.
- Select the Page-based browser-level API (HTML / HTTP) option to
record Web traffic at the browser level. If you choose this option the
SilkPerformer Recorder generates scripts that call functions
provided by the page-based browser-level API.
88 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
- Select the Browser-level API (HTTP) option to record Web traffic at
the browser level. If you choose this option the SilkPerformer
Recorder generates scripts that call functions provided by the
browser-level API.
- Select the TCP/IP-level API (TCP/IP / HTTP) option to record Web
traffic on the TCP/IP level. Select this option to capture proprietary
protocols carried by HTTP.
8 Select an Advanced context management level from the combo box.
Click the View Settings button to view or customize the level of
advanced context management.
9 Select the Automatic page detection option to enable the page notion
technology of the Recorder. When you use the page-based Web
functions, this feature attaches embedded documents that are loaded by
JavaScript code into a Web page. When you use the other browser-level
functions, this feature wraps all the requests that load a Web page in
WebUrlBeginPage and WebUrlEndPage function calls and so
automatically creates page timers and synchronizes the downloading of
Web pages for improved accuracy.
10 Select the Convert URL query strings to forms option to convert recorded
URL query strings to Web forms and then include the forms in the
dclform section of the generated test script. You should disable this
option only if the URLs are irregular.
11 Select the Enable persistent-cookie recording for returning-user
simulation option to have the Recorder generate a WebCookieSet
function call at the beginning of the current transaction for each
persistent cookie that you receive from the Web server. During a load
test, each WebCookieSet function then emulates a cookie that is
persistently stored in the virtual user’s cookie database.
Note The WebCookieSet function calls are marked as comments—
so you must remove the comment marks to activate the function
calls.
12 Select the Encapsulate concurrent pages using timers option to
encapsulate concurrent pages with timers. Page timers, normally
included in the page-based browser-level functions or in calls of the
WebUrlBeginPage function, are not scripted. Instead, MeasureStart and
MeasureStop functions are inserted automatically (based on timings).
A series of browser activities is considered “concurrent” if there is no
browser idle period longer than the value of the minimum thinktime
recorded.
SilkPerformer User Guide 89

3 CONFIGURING PROJECTS
Project Profile Settings
Verification options
13 Use the Record additional HTTP headers area to specify additional
HTTP headers you want to record, such as those generated by a custom
HTTP client.
14 Click the Add button to add the name of an additional HTTP header you
want to record, such as one generated by a custom HTTP client you are
using. You must know the exact name of the header you want to record
to use this feature.
15 The HTTP Headers dialog appears. Enter the name of the header you
want to add to the list, or select one from the drop-down list.
16 Click OK.
17 Click the Edit button if you want to edit the selected additional HTTP
header.
18 The HTTP Headers dialog appears. Edit the header as required.
19 Click OK.
20 Click the Remove button if you want to remove the selected additional
HTTP header from the list.
Procedure To set verification options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Verification tab.
7 Use the Recording area to specify options for generating verification
functions when recording a script based on captured Web traffic. Click a
specific item for detailed information.
8 Select the Record title verification option if you want the SilkPerformer
Recorder to automatically generate a title verification function
(WebVerifyHtmlTitle) for each Web page.
90 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
9 Select the Record digest verification option if you want the
SilkPerformer Recorder to automatically generate a digest verification
function (WebVerifyDataDigest) for a response that can be in the form
of one of the listed content types:
- Select the All characters option to include in the digest all of the
characters that are received in HTTP response messages from the
Web server.
- Select the Printable characters option to include in the digest only
the printable characters that are received in HTTP response
messages from the Web server.
- Select the Alphanumeric characters option to include in the digest
only the alphanumeric characters that are received in HTTP response
messages from the Web server.
- Use the Verify data of content type area to list all HTTP response
content types for which to generate a response verification digest.
- Click the Add button to add a new HTTP response content type that
you want to use for response verification.
The Response Content Types dialog appears.
Enter the name of the new content type you want to add to the list, or
select one from the drop-down list.
Click OK.
Click the Edit button to edit the currently selected HTTP response
content type.
The Response Content Types dialog appears.
Edit the currently selected HTTP response content type.
Click OK.
Click the Remove button to remove the currently selected HTTP
response content type from the list.
Click OK to exit the dialog and save your changes.
10 Use the HTML / XML area to enable HTML and XML document
verification when replaying scripts based on Web traffic captured by the
SilkPerformer Recorder.
11 Select the Title verification option to enable SilkPerformer to verify the
title of HTML documents. An operation (depending on the specified
severity: error, warning, informational, custom) will be performed if the
verification fails. The setting used here is valid for both the page-based
browser-level API and the browser-level API (WebSetVerificationEx).
If this option is selected, it will trigger the execution of the
WebVerifyHtmlTitle function in the BDL script.
SilkPerformer User Guide 91

3 CONFIGURING PROJECTS
Project Profile Settings
12 Select the Table verification option to enable SilkPerformer to verify that
a table within an HTML document contains some specific text. An
operation (depending on the specified severity: error, warning,
informational, custom) will be performed if the verification fails. The
setting used here is valid for both the page-based browser-level API and
the browser-level API (WebSetVerificationEx).
If this option is selected, it will trigger the execution of the
WebVerifyTable function in the BDL script.
13 Select the Digest verification tolerance level option to verify the HTML
content of a server response. Use TrueLog Explorer to generate
WebVerifyHtmlDigest function calls.
Enter the maximum difference in character frequency, quantified in
bytes, that may occur between the server response that has been used by
TrueLog Explorer to generate the HTML digest and the response that is
received during replay. An operation (depending on the specified
severity: error, warning, informational, custom) is performed, if the
number of differing bytes exceeds this limit.
14 Select the HTML verification option to enable SilkPerformer to verify
that an HTML document contains some specific text. An operation
(depending on the specified severity: error, warning, informational,
custom) will be performed if the verification fails. The setting used here
is valid for both the page-based browser-level API and the browser-level
API.
If this option is checked it will trigger the execution of the
WebVerifyHTML and WebVerifyHTMLBound functions in the BDL
script.
15 Select the Link checking option to check the validity of links in HTML
documents received while a load test is run. The HTML document is
parsed on receipt, and additional requests are sent to check that links are
valid. An error message is generated if a target pointed to by a link is
invalid. This feature is available only in the page-based browser-level
API.
16 Select the XML verification option to enable SilkPerformer to verify, that
an XML document contains some specific values or attributes. An
operation (depending on the specified severity: error, warning,
informational, custom) will be performed if the verification fails. The
setting used here is valid for both the page-based browser-level API and
the browser-level API.
If this option is checked it triggers the execution of the
WebXmlVerifyNodeValue and WebXmlVerifyNodeAttribute functions
in the BDL script.
92 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
HTTP body
transformation
17 Use the Data area to specify options for HTTP response verification
when replaying a script based on Web traffic captured by the
SilkPerformer Recorder.
18 Select the Data verification option to enable SilkPerformer to verify that
raw data received from the server contains certain specific data (e.g.,
HTML source code containing some specific code). An operation
(depending on the specified severity: error, warning, informational,
custom) will be performed if the verification fails. The setting used here
is valid for both the page-based browser-level API and the browser-level
API.
If this option is checked it triggers execution of the WebVerifyData and
WebVerifyDataBound functions in the BDL script.
19 Select the Digest verification tolerance level option to have
SilkPerformer verify server response data during a load test. Note that
the SilkPerformer Recorder generates the digests needed for verification.
In the field to the right, enter the maximum difference in character
frequency, quantified in bytes, that may occur between the response
message received from the server during recording and the response
message received during replay. If the number of bytes that differ
exceeds this limit, SilkPerformer will report an error. The SilkPerformer
Recorder creates digests only for the content types that are included in
the Verify data of content type list.
Procedure To install an HTTP body transformation:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Web icon.
5 Select the Replay category.
6 Click the Transformation tab.
7 Select a transformation type from the Type drop-down list.
8 Optionally, enter additional parameters for the transformation type.
SilkPerformer User Guide 93

3 CONFIGURING PROJECTS
Project Profile Settings
Note See the SilkPerformer Advanced Concepts book for details
about valid parameters for the Flash Remoting transformation
package that is pre-installed with SilkPerformer.
9 Use the checkboxes to specify if the transformation should be applied to
HTTP request bodies (Transform HTTP requests), HTTP response
bodies (Transform HTTP responses), or both. Select the respective Log
encoded checkboxes to additionally log the encoded HTTP response
bodies.
Terminal Client options
Note SilkPerformer offers tutorials that walk you through the
process of load testing terminal client applications (Load Testing
IBM Mainframe Applications and Load Testing VT 100+
Applications). These tutorials are available in Miscellaneous
Tutorials (/Documentation/
Tutorials/...).
Terminal client
options
Procedure To set terminal client options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
4 In the shortcut list on the left, click the Terminal Client icon.
5 Select the Record category.
6 In the Telnet settings area of the Telnet tab, specify a telnet detection
mode.
- Select Always assume Telnet-mode to instruct the recorder to assume
that each connection is a Telnet connection.
- Select Auto-detect Telnet-mode to instruct the recorder to
automatically detect Telnet on each new connection.
- Select Never assume Telnet-mode to instruct the recorder to assume
that no connections are Telnet connections.
7 In the Command prompt string field, you can specify a command prompt
string that is to be used by the recorder to script
WebTcpipRecvPacketsUntilData functions wherever possible. This
makes recorded scripts more effective for dynamic content.
94 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
8 From the Log level drop-down list in the Terminal properties section of
the dialog, select the appropriate logging level (None, Error, Normal, or
Debug).
9 From the Terminal type drop-down list select the supported terminal
types. When correctly specified, the recorder scripts the appropriate
terminal-type initialization functions and renders the screens to the
TrueLog.
10 In the Configuration string field, initialize the terminal emulator with a
custom configuration string (see your terminal emulator documentation
for available options).
11 Use the Default screen fields to indicate the default appearance of the
terminal screen to the recorder. The recorder will use these values if the
actual values cannot be detected during session initialization.
- Use the rows field to indicate the default height of the terminal
screen. The recorder will use these values if the actual height cannot
be detected during session initialization.
- Use the columns field to indicate the default width of the terminal
screen. The recorder will use these values if the actual width cannot
be detected during session initialization.
- Select the Colors checkbox to specify default color use on the
terminal screen. The recorder will use these values if the use of
colors cannot be detected during session initialization.
12 In the Host code page field, specify the code page that is used by the
server. The code page is used for various data conversions. The available
code pages have been installed with the system. If the required code page
is not listed, you may install it on the system. Ensure that the selected
code page is also available on each agent computer.
13 Click the Advanced tab.
14 Select the Record basic functions option to prevent the recorder from
scripting synchronized receive-functions, for example
WebTcpipScreenRecvPacketsUntilCursor and
WebTelnetScreenRecvRecordsUntilStatus. If enabled, scripted receive
functions will not rely on rendered screen content. This option is enabled
by default when a terminal type is not specified.
GUI-Level Testing options
General options
Procedure To configure GUI-level testing settings for your project:
1 Select Active Profile from the Settings menu.
SilkPerformer User Guide 95

3 CONFIGURING PROJECTS
Project Profile Settings
2 Scroll down and select the GUI-Level Testing group icon.
3 The Session settings section of the General tab is used to configure the
connection between your console session and remote desktop sessions.
Select the Run TryScript on console checkbox to have SilkTest run a
TryScript on the console application. When this option is enabled, you
can view what SilkPerformer does during tests, but you are limited in
what you can do on your computer because SilkTest uses your console.
If disabled, TryScripts will connect to and run SilkTest on remote
desktop sessions.
4 Select the Allow usage of console session on agents checkbox if you do
not have a terminal-server license for all agents, or if your agents run on
Windows Home or Windows Professional editions. SilkPerformer only
uses this option if the connection to a terminal services session fails, or if
no user credentials are provided.
Note When this option is selected, you can start only one virtual
user on each agent (for agents that have no Terminal Server license
and Windows Home/Professional systems).
5 Define a Connection timeout (in seconds) to specify how long
SilkPerformer should wait before a connection is established to remote
desktops. Set this setting to '0' to have no timeout.
6 Complete the Username/Password settings to enable automatic login if
you only require a single set of login credentials. These credentials will
be automatically loaded into StInitSession functions. These credentials
can be overruled by scripted settings, e.g., providing parameters to the
StInitSession function by using project attributes. See “Specifying user
credentials” for more information.
7 The Execution Settings portion of the dialog enables you to configure
settings for the script and the runtime. Select the Log SilkTest errors
checkbox to have errors and warnings that are sent back by SilkTest
logged as informational warnings (this option is enabled by default).
8 Select a TrueLog setting for the remote desktop session. From the
Generate TrueLog file for drop-down list, select None (the default) to not
have TrueLog written. Select On Error to have TrueLog written only
when errors are encountered. Or select Custom when you wish to have
TrueLog written based on your custom settings.
If you select Custom TrueLog settings, your TrueLog settings will be
read from the SilkTest option file, or they can be set via the StSetOption
function.
96 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Note Even if you select OnError or Custom from the TrueLog on
SilkTest drop-down list, SilkTest will only write TrueLog files if you
also specify in SilkPerformer’s settings that TrueLog should be
written. To do this, click the Generate TrueLog Files button or the
Generate TrueLog On Error Files button on SilkPerformer’s
toolbar.
9 Specify an Execution timeout (in seconds) to specify how long the
runtime should wait before test cases execute. Set this setting to '0' to
have no timeout.
10 Results history size (set to 9 by default) specifies how many test case
results SilkTest can save in a .res file.
Note By starting a new load test, the results in the .res file are reset.
11 In the SilkTest option file field, browse to and select a SilkTest option
file. See the Silk Online Help for full details about using option files.
12 Click OK to save your settings.
Oracle Forms options
Note SilkPerformer offers a tutorial that walks you through the
process of load testing Oracle Forms. The Oracle Forms tutorial is
available at (/Documentation/
Tutorials/...).
Oracle Forms options
Procedure To set Oracle Forms options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
4 In the shortcut list on the left, scroll down to and select the Oracle Forms
icon.
5 Select the Replay category.
6 Oracle Forms clients can communicate with servers via Sockets, HTTP,
or HTTPS. Select a Connection mode setting that reflects the server
version and configuration you are testing.
Note This setting can be overridden by the
OraFormsSetConnectMode(connectMode) BDL function call.
SilkPerformer User Guide 97

3 CONFIGURING PROJECTS
Project Profile Settings
- Select Socket mode when the Oracle Forms server is set up for
socket mode.
- Select HTTP mode for Oracle Forms servers that are configured for
HTTP connections.
- Select HTTPS mode for Oracle Forms servers that are configured for
secure HTTP connections.
7 In the Connection timeout field, specify a timeout period (in seconds) for
send and receive connections.
8 To configure additional runtime settings, you can optionally have the
virtual user send a heartbeat message to the server at a specified interval.
Enter an interval (in seconds) in the sec field.
9 Select the Automatically wait for application timers checkbox to direct
the replay engine to wait for application timers to expire after each
function call.
10 The Application timer and window timeout setting specifies the
maximum wait time for expiration of application timers and appearance
of windows. When enabled, the replay engine waits for timers to expire
after each function call. When this option is not enabled, a function
named OraFormsWaitForTimer can be used to make an application wait
for timers to expire. Maximum wait time is specified by this setting. The
function OraFormsWaitForWindow may be used to have an application
wait for windows to appear. This option specifies the maximum wait
time for window appearance.
11 Enter the Forms version you are working with. Oracle Forms clients
send the required version when connecting to the Oracle Forms server.
Create profiles and change this setting when testing applications
deployed on different server versions.
Note This setting can be overridden by the
OraFormsSetInt("INITAL_VERSION", theVersion) BDL function
call.
12 Click the Logging tab. Here you can specify that additional properties
have their values logged during replay.
13 Set the Log level for virtual user logging:
- None - No additional logs will be written (disables logging)
- Error - Only errors that occur during replay will be logged.
- Normal - Oracle Forms Messages will be logged.
- Debug - Additional information will be logged (e.g., events).
98 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
14 In the Additional properties field you can specify other properties for
which virtual users are to log values.
- Click Add to add a property that is to have its value logged during
replay.
- Click Edit to edit a selected property.
- Click Remove to remove a selected property. No further values will
be logged for this property.
15 Click the Measuring tab. Here you can specify a custom measurement
level. By default, all available performance metrics are collected during
test runs.
16 Select the Enable all timers and counters for all controls checkbox to
enable all timers and counters for all actions on controls during replay.
Or, select a specific control from the Control drop-down list and select
the specific timer/counter types that you want to have applied to that
control:
- Enable timers - Measure how long it takes to complete an action on
the selected control.
- Count round trips - Count the round trips for each action on the
selected control.
- Count bytes - Count the bytes sent and received for each action on
the selected control.
- Count messages - Count the messages sent and received for each
action on the selected control.
Note Alternately, you can click the Apply to All Controls button to
apply your timer/counter settings to all controls.
17 Click OK to save your settings.
SAP GUI options
Note SilkPerformer offers a tutorial that walks you through the
process of load testing SAP GUI applications. The SAP tutorial is
available at (/Documentation/
Tutorials/...).
SAP GUI options
Procedure To set SAP GUI options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
SilkPerformer User Guide 99

3 CONFIGURING PROJECTS
Project Profile Settings
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
4 In the shortcut list on the left, scroll down to and select the SAP GUI
icon.
5 Select the Record category.
6 On the General tab you can define options for how the SAP Recorder
records and replays scripts.
7 Login to SAP systems involves several user interactions. With the Script
logon as a single function option enabled, the recorder only scripts one
function that performs all the required login actions.
8 Select the Script low level functions option to use a basic set of scripting
functions instead of object-specific functions.
9 Select the Script timers option to script timers automatically. These
timers are used for measuring SAP API calls.
10 Select Attach to existing SAP session to have the recorder attach to an
existing session rather than create a new session.
11 Select the Record window title verification option to record
SapGuiSetActiveWindow functions in such a way that title verifications
are automatically performed on activated windows during test runs.
12 Now select the Replay category to configure replay settings.
13 In the Replay timeout field, specify a timeout period (in seconds) for
automatic shutdown of SAPFewgsrv (when not otherwise closed at the
end of test runs).
14 Select the Show SAP GUI during replay checkbox to display the R/3
client during test runs. GUI display is only an option with TryScript
executions. This setting is ignored for baseline tests and load tests.
15 Select the Enable client-side scripting option to enable SAP client-side
scripting via the registry. This option disables warnings that raise pop-up
windows when new users start.
16 Select the User new SAP Visual Design option to enable SAP visual
design for the SAP client.
17 Click the Logging tab to specify additional log levels.
Note These settings affect performance and resource utilization.
18 From the Log level drop-down list, select a log level for virtual user
logging:
100 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
- Disable - Virtual user logging is disabled.
- Normal - Default logging.
- Debug - Detailed information is logged, such as additional session
information and errors.
19 Select the Capture screenshots option to generate screenshots for each
activated window. Screenshot capture is only enabled for TryScript
executions. This setting is ignored for baseline tests and load tests.
20 Select the Capture screenshots for each action checkbox to have each
action performed by the script documented with a screenshot. Screenshot
capture is only enabled for TryScript executions. This setting is ignored
for baseline tests and load tests.
21 Select the Log control information in TrueLog option to have additional
control information written to the TrueLog.
22 Select the Log control information for single run in TrueLog option to
have additional control information written to the TrueLog only for try
script and verification tests.
23 Select the Log control information on error option to enable writing of
control information to TrueLogs in the case of errors.
24 Select the Highlight controls checkbox to have controls highlighted as
they are accessed by the script.
25 Select the Measuring tab to specify a custom measurement level. By
default, all available performance metrics are collected during test runs.
26 Select the Enable all timers and counters for all controls checkbox to
enable all timers and counters for all actions on controls during replay.
Or, select a specific control from the Control drop-down list and select
the specific counter types that you want to have applied to that control:
- Count round-trips - The number of round-trips for the action on the
control.
- Count response time - The time it takes a function to return
requested data.
- Count Interpretation time - The time it takes to interpret a request.
- Count flushes - The number of flushes per action.
Note Alternately, you can click the Apply to All Controls button to
apply your counter settings to all controls.
27 Click OK to save your settings.
SilkPerformer User Guide 101

3 CONFIGURING PROJECTS
Project Profile Settings
Citrix options
Note SilkPerformer offers a tutorial that walks you through the
process of load testing Citrix-based applications. The Citrix tutorial
is available at (/Documentation/
Tutorials/...).
Citrix options
Procedure To set Citrix options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
4 In the shortcut list on the left, scroll down to and select the Citrix icon.
5 Select the Replay category.
6 In the Options area you can specify general Citrix options. In the
Synchronization timeout field, enter a default synchronization timeout
(in seconds) that is to be used by all CitrixWaitForXXX functions that
don't specify a timeout value.
7 Select the Force window position checkbox to make the calls to
CitrixWaitForWindowCreation and CitrixWaitForWindowRestore move
windows to the coordinates captured during recording. When this option
is not selected, both functions provide a parameter (bForcePos) that
enables this option for each individual call.
8 Select the Disconnect on transaction end option to disconnect the Citrix
client upon each transaction end (this includes the TInit transaction).
9 You can specify Citrix logging options in the Logging area of the dialog.
When TrueLog generation is enabled, at the end of each synchronization
function a screenshot is made and written to the TrueLog. Select the Log
screen before each user action option to also capture and write a
screenshot at the beginning of each user action.
10 The CitrixWaitForScreen function captures a screen region and checks it
against a specified condition (normally it compares it to a hash value
that's captured at recording). If the function call fails when the Dump
window region on unsuccessful screen synchronization option is selected
(i.e., if the condition still doesn't match and the timeout period has
102 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
expired), the captured screen region will be written to a file in the result
directory. This file can later be examined or compared to what was
captured during recording for error analysis.
11 Select the User RAM disk option to use a different drive for intermediate
storage of images. If the specified drive is a RAM disk, TrueLog
generation performance will improve. Select the drive letter of the RAM
disk from the drop-down list.
12 Click the Simulation tab.
13 In the Mouse area of the dialog you can specify mouse options. In the
Length of time mouse button remains pressed field, enter the length of
time that the virtual user is to hold the mouse button in the pressed state
(for CitrixMouseClick and CitrixMouseDblClick).
14 In the Length of time between the clicks of a double-click field, enter the
length of time that is to pass between the two clicks of a double click (for
CitrixMouseDblClick).
15 In the Mouse speed field, enter the speed at which the mouse is to move
across the screen (pixels per second).
16 In the Keyboard area of the dialog you can specify keyboard options. In
the Length of time each key remains pressed field, enter the length of
time that the virtual user is to hold a keyboard key in the down state (for
CitrixKey and CitrixKeyString).
17 In the Length of time between keystrokes when entering strings field,
enter the length of time that is to pass between the individual key-strokes
specified by the CitrixKeyString function.
18 In the Key repeat time field, enter the time required for a complete key
stroke when simulating repeat functionality (a value of 50 means 20 keys
per second).
19 In the Think times area of the dialog you can specify think time options.
In the Delay after successful synchronization field, enter the length of
time that virtual users are to remain inactive after passing a successful
synchronization point (CitrixWaitForXXX).
20 In the Delay after each user action field, enter the length of time that
virtual users are to remain inactive between actions.
21 Select the Citrix client tab to specify Citrix client options.
22 From the Network protocol drop-down list, select which low-level
network protocol is to be used to locate and connect to the MetaFrame
server (see Citrix client documentation for more information).
23 Select the Use data compression option to compress all transferred data.
This reduces file size, but requires additional processor resources.
SilkPerformer User Guide 103

3 CONFIGURING PROJECTS
Project Profile Settings
24 Select the Use disk cache for bitmaps option to store commonly used
graphical objects, such as bitmaps, in a local disk cache.
25 Select the Queue mouse movements and keystrokes option to queue
mouse and keyboard updates. This reduces the number of network
packets that are sent from the Citrix client to the MetaFrame server.
26 From the SpeedScreen latency reduction drop-down list, select an option
for enhancing user experience on slower network connections. For
slower connections (e.g., WANs) select On. For faster connections (e.g.,
LANs), select Off. Set the mode to Auto to turn latency reduction on and
off based on the latency of connections.
27 From the Encryption level drop-down list, select a level of encryption for
the ICA connection. The MetaFrame server must be configured to allow
the selected encryption level, or higher.
Note Using an encryption level other than the server default or
Basic disables automatic logon to the MetaFrame server.
.NET options
.NET options
Procedure To set .NET options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left, scroll down to and select the .NET icon.
5 Select the Replay category.
6 Select the Route HTTP .NET through SilkPerformer Web engine
checkbox to route all http/https traffic that is generated by your .NET test
driver through the SilkPerformer Web engine. This allows you to take
advantage of the features that SilkPerformer offers for accurate load
testing, such as its high performance Web engine, modem simulation,
detailed timing/throughput information for each transmission, HTTP/
XML TrueLog, and IP spoofing.
104 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
7 The Routed Web service proxy classes field displays the .NET
WebService client classes that are used in your .NET test driver. HTTP/
SOAP traffic of selected classes will be routed through SilkPerformer’s
Web engine, regardless of the setting above. This setting can only be
changed via the SilkPerformer .NET AddIn.
8 Select the One application domain for each virtual user container
process option to share one .NET application domain among all virtual
users running in a virtual user process. This minimizes the administrative
overhead for the .NET Common Language Runtime, but may cause
runtime problems because all objects will be loaded into the same
application domain.
9 Select the One application domain for each virtual user option to give
each virtual user its own .NET application domain. This increases the
administrative overhead of the .NET Common Language Runtime, but
ensures that objects from different virtual users won’t interfere with one
another.
10 Click OK to save your settings.
CORBA/IIOP options
Replay options
Procedure To set replay options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the CORBA / IIOP icon.
5 Select the Replay category.
6 Use the IIOP 1.2 area to specify the default settings for IIOP 1.2 replay.
7 In the Max. chunk size for value type field, enter the maximum size for a
single part of a chunked-encoded value type, that is, a value type that has
been broken into smaller parts.
8 Select the Use fragmentation option to fragment IIOP request and
response messages if they exceed the length set in the Fragmentation
size field.
SilkPerformer User Guide 105

3 CONFIGURING PROJECTS
Project Profile Settings
9 Use the Fragmentation size field to set the fragmentation size for a
chunked-encoded value type. If a request or response message exceeds
this size, the message will be sent using fragmentation.
WChar/WString
options
Procedure To set WChar / WString options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the CORBA / IIOP icon.
5 Select the Replay category.
6 Click the WChar / WString tab.
7 Use the Code set area to specify the settings for the wide-character code
set. In general, a code set determines the encoding rules for characters
and wide characters. Code sets can be byte-oriented or non-byte-oriented
(e.g., ASCII is a byte-oriented code set for single byte characters).
Multibyte characters defined by non-byte-oriented code sets have a byteorientation
(big-endian, little-endian), whereas byte-oriented code sets
do not.
8 Select the Unicode option to use the Unicode code set. Unicode is a nonbyte-oriented
code set for two-byte characters. The byte orientation of
Unicode characters is determined by the IIOP message byte order.
9 From the WChar size drop-down list, select the size of a CORBA wide
character (wchar).
10 Select the Byte-oriented option for a byte-oriented code set to be used for
CORBA wchars and wstrings.
Or, select the Non-byte-oriented option for a non-byte-oriented code set
to be used for CORBA wchars and wstrings. If you select this option, use
the Script byte order drop-down list to select the byte orientation of nonbyte-oriented
CORBA wchar/wstring representation in the BDL script.
The byte orientation can be little-endian (least significant bit first) or bigendian
(most significant bit first).
11 Use the IIOP 1.2 area to specify the settings for IIOP 1.2. The wchar and
wstring encoding in IIOP 1.2 differs from the encoding in previous IIOP
versions. However, not all ORBs use the correct encoding for IIOP 1.2.
106 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Recording options
12 Select the CORBA compliant option if the ORB uses CORBA compliant
wchar/wstring encoding in IIOP 1.2.
13 In the WChar area, select the No length encoding option to use pre IIOP
1.2 encoding. CORBA compliant IIOP 1.2 encoded wchars use a length
indication for the size of the wchar.
14 In the WString area, select the Null-terminated option if the ORB uses
null-terminated wstrings. CORBA compliant IIOP 1.2 encoded wstrings
are not null-terminated.
15 Select the Chars option if the ORB uses the number of chars as wstring
length. CORBA compliant IIOP 1.2 encoded wstrings specify the string
length in bytes (chars).
Or select the WChars option if the ORB uses the number of wchars as
wstring length.
Procedure To set recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon for the profile for which you want to configure
settings.
3 Select Edit Profile from the context menu.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the CORBA / IIOP icon.
5 Click the Recording tab.
6 Select the Record category.
7 Use the Proprietary ORB features support area to specify whether to
consider ORB-specific features when generating load test scripts based
on recorded IIOP traffic. If the appropriate options are selected, the
generated script becomes more readable.
8 When generating a load test script based on recorded IIOP traffic
- Select the Orbix option to consider Orbix-specific features
- Select the OrbixWeb option to consider OrbixWeb-specific features
- Select the VisiBroker C++ option to consider features specific to
VisiBroker for C++
- Select the VisiBroker Java to consider features specific to
VisiBroker for Java
SilkPerformer User Guide 107

3 CONFIGURING PROJECTS
Project Profile Settings
- Select the M3 C++ to consider features specific to BEA Millennium
for C++
- Select the M3 Java to consider features specific to BEA Millennium
for Java
9 From the CORBA version of IDL file drop-down list, select the CORBA
version of the application you are testing. If the application is using a
CORBA version from 2.0 to 2.2 inclusive, select 2.0 from the drop-down
list. If the application is using CORBA 2.3 or CORBA 3.0, select 3.0
from the drop-down list.
10 In the Preprocessor arguments field, enter any number of arguments that
are to be passed to the preprocessor of the .idl file parser. For example,
the arguments may be used to specify the include directory and to
execute some predefined macros.
11 Use the IDL files loaded list to view all the .idl files that are currently
used by the SilkPerformer Recorder to generate high-level load test
scripts. Whenever the corresponding interface is defined in a .idl file, the
Recorder generates type-dependent functions for specifying the
parameters of CORBA operation calls in the load test script.
12 Click OK to exit the dialog and save your changes.
Java options
General options
Procedure To set general Java options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile that you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Java icon.
5 Select the Replay category.
6 Click the General tab.
7 Use the Settings area to set options for the Java Virtual Machine.
8 From the Version drop-down list, select the version number of the Java
Virtual Machine.
108 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
Advanced options
9 In the Java home field, enter the installation directory of the Java Virtual
Machine. This option enables the Java Virtual Machine to be loaded for a
different path than specified in the PATH environment, so that it is
possible to switch between various Java Virtual Machines without
changing the system PATH environment. If necessary, use the browse
(...) button to navigate to and select the installation directory.
10 Use the Classpath field to view the class path specified for the Java
Virtual Machine. By default the class path is set to the system class path.
11 Click the File button to navigate to a class file and add it to the class path
of the Java Virtual Machine.
12 Click the Dir button to navigate to a directory and add it to the class path
of the Java Virtual Machine.
13 To reorder the files, click the Up and Dwn buttons to move selected files
up or down in the class path hierarchy.
14 To remove a selected file from the class path, click the Remove button.
15 Use the System classpath field to view the full path of the system class
path.
16 Select the Use system classpath option to ensure that the system class
path is used.
Procedure To set advanced options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile that you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Java icon.
5 Select the Replay category.
6 Click the Advanced tab.
7 Use the Advanced area to set advanced options for the Java Virtual
Machine.
8 In the Command line options field, enter any command line options that
are to be passed to the Java Virtual Machine.
9 Select the Disable JIT compiler option to disable the Just In Time (JIT)
compiler of the Java Virtual Machine.
SilkPerformer User Guide 109

3 CONFIGURING PROJECTS
Project Profile Settings
Logging options
10 In the JVM dll field, enter the name of any Dynamic Link Library (.dll)
that implements the Java Virtual Machine that is to be used.
11 If necessary, use the browse (...) button to the right of the field to
navigate to and select the Dynamic Link Library that is to be used.
12 Click the Check JVM button to verify your Java Virtual Machine
configuration settings.
13 Click OK to exit the dialog and save your changes.
Procedure To set logging options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile that you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Java icon.
5 Select the Record category.
6 Click the Logging tab.
7 Use the Record settings area to specify that additional properties have
their values logged during record.
8 Set the log level for virtual user logging from the Log level drop-down
list. To disable logging, set the log level to None. The following log
levels are available:
- None: No additional logs will be written.
- Error: Only errors that occur during replay will be logged.
- Normal: Java messages will be logged.
- Debug: Additional information will be logged (e.g., events).
Database options
SQLSTATE-specific
error handling
Procedure To set SQLSTATE-specific error-handling options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
110 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
ODBC-specific error
handling
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Replay category and click
the Database icon.
5 Click the SQLSTATE tab.
6 Use the Simulation exceptions list to define exceptions for specific
ODBC error codes to prevent termination of the simulation. If a database
error occurs, SilkPerformer will automatically terminate the entire
simulation. The list includes all the exceptions that SilkPerformer uses
for handling database-specific errors. The Error field lists the native
error code. The Class field ranks the severity of the error. Class 0 errors
can be handled within a SilkPerformer transaction. Class 1 errors end the
transaction but do not cause the simulation to terminate.
7 To add an error to the list of database-specific error exceptions, click the
Add button.
The Edit Error dialog appears.
8 In the Error field, add the error code for an error to be added to the
exceptions list.
9 In the Class field, add the error class.
10 In the Description field, add a description for the error.
11 Click OK.
12 To edit an error, select it from the list, and click Edit.
The Edit Error dialog appears.
13 Edit the error code, class, and description as required.
14 Click OK.
15 To remove an error from the exceptions list, select it and click the
Remove button.
Procedure To set ODBC-specific error-handling options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
SilkPerformer User Guide 111

3 CONFIGURING PROJECTS
Project Profile Settings
ODBC high-level
options
4 In the shortcut list on the left side, select the Replay category and click
the Database icon.
5 Click the ODBC Database Errors tab.
6 Use the Error definition list to define exceptions for specific database
error codes in order to prevent the termination of the simulation.
7 To add an error to the list of database-specific error exceptions, click the
Add button.
The Edit Error dialog appears.
8 In the Error field, add the error code for an error to be added to the
exceptions list.
9 In the Class field, add the error class.
10 In the Description field, add a description for the error.
11 Click OK.
12 To edit an error, select it from the list, and click Edit.
The Edit Error dialog appears.
13 Edit the error code, class, and description as required.
14 Click OK.
15 To remove an error from the exceptions list, select it and click the
Remove button.
Procedure To set ODBC high-level options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Replay category and click
the Database icon.
5 Click the ODBC High-Level tab.
6 Use the Database area to specify details about the ODBC data source
you will use and how SilkPerformer should interact with the ODBC
driver.
7 From the Data Source Name drop-down list of available ODBC data
sources, select the data source to which you want to connect for load
testing. To set up an additional ODBC data source, use Windows 32-bit
112 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
ODBC Administrator (via the Windows Control Panel. The data source
you select will only be used to replay test scripts with high-level ODBC
functions.
Note that the data source you select here is used only if you use highlevel
ODBC functions that perform an implicit connection to the
database. When you use medium-level ODBC functions, this data source
is not used for connecting. Instead, the connection is performed with an
explicit DB_Connect("connection string") function in the script.
8 In the UserID field, enter the user name that is to be used to connect to
the data source. You must specify a valid user for the DBMS to which
you want to connect. If you use a trusted connection to connect to your
DBMS, leave this field blank.
Note that the user ID you enter here will only be used if you use highlevel
ODBC functions.
9 In the Password field, enter the password that is to be used when
connecting to the data source. You must specify a valid password for the
database user you are connecting. If you use a trusted connection, or if
the user has no password, leave this field blank.
Note that the password you enter here, as well as all remaining options
you choose in this area, are used only if you use high-level ODBC
functions.
10 From the Isolation drop-down list, select the level of isolation for the
entire load test. For a detailed description of isolation levels, consult your
DBMS documentation and see the ODBC functions reference.
You can change the isolation level and any other DBMS specific settings
using the ODBC functions SQLSetConnectOption and
SQLSetStmtOption in your script.
11 From the SQL concurrency drop-down list, select a type of cursor
concurrency to use for the scrollable cursors in your load test script. For
a detailed description of concurrency settings, consult your DBMS
documentation.
12 From the SQL cursor type drop-down list, select a cursor type for the
scrollable cursors in your load test script. For a detailed description of
cursor type settings, consult your DBMS documentation.
13 Select the Autocommit option to commit the database automatically after
each SQL command. Otherwise the database will be committed only
after an explicit commit statement. This setting is effective for all
connected cursors.
SilkPerformer User Guide 113

3 CONFIGURING PROJECTS
Project Profile Settings
14 Select the Reprepare SQL commands option to prepare and execute a
SQL statement each time it is called. Otherwise SilkPerformer prepares
SQL statements only when necessary. This setting is effective for all
cursors used in the load test.
Recording options
Procedure To set recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile that you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, click the Database icon.
5 Select the Record category.
6 Click the Recording tab.
7 In the Oracle OCI area, select the Generate random input for long
placeholders option to replace large amounts of data exchanged between
the client application and the database server with random data. Since the
original data is rarely required for replay, this option helps to prevent
large data files being carried along with load test scripts.
Select the Generate files for long placeholders option to store large
amounts of data exchanged between the client application and the
database server in a file. During replay, the file is used to send to the
database server exactly the same data as to the client application.
8 Click OK to exit the dialog and save your changes.
COM / ADO options
Note SilkPerformer offers two tutorials that walk you through the
process of load testing ADO and COM-based applications (Load
Testing a COM Application and Load Testing an ADO Application).
These tutorials are available in Miscellaneous Tutorials
(/Documentation/Tutorials/...).
Recording options
Procedure To set COM/ADO recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
114 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Record category and click
the COM / ADO icon.
5 On the Settings tab, in the Script options area, select the Generate
warnings option for warnings to be included as comments in the test
script which will be generated when you have finished recording.
Warnings may occur at any time while a COM application is running and
indicate that the application might not be working properly. Select this
option to verify that the application you are going to test works properly
before you run an actual load test.
6 Select the Generate OLE32 functions details option to have
SilkPerformer generate test scripts that mimic COM clients with extreme
accuracy. When you select this option, the generated test scripts pass
exactly the same parameters to all OLE32 functions as the COM client
did; this makes test scripts look more complicated. Disable this option to
replace parameters that are not relevant to load testing with default
values; this will make the test scripts look less complicated and they will
be easier to customize.
Note: This option should be used only by testers who have a
comprehensive knowledge of COM.
7 In the Max. number of table elements field, specify the maximum
number of array elements for which SilkPerformer should generate
access functions in the test script. These functions are generated
whenever an array parameter is passed to or received from a COM
object.
8 Select the Enhanced property access option if you want to make use of
the enhanced property access that allows to get/set a property value with
a single ComSet{Type}/ComGet{Type} call.
9 Select the Enhanced function calls option if you want to make use of the
enhanced function call that allows to call a method that takes no
parameters with a single ComCall statement.
10 In the Log options area you can specify details about logging COM
function calls. Select the Show missing type information option to check
if all interfaces of the COM components under test are available to
SilkPerformer. To record interface functions calls, SilkPerformer needs
descriptions of interfaces or COM classes. These descriptions are
contained in type libraries.
SilkPerformer User Guide 115

3 CONFIGURING PROJECTS
Project Profile Settings
ADO recording
options
11 Select the Show type mismatches option for SilkPerformer to inform you
about mismatches between parameter types that an interface function
expects and actual parameter types. Only late binding clients can cause
type mismatches. A type mismatch happens, for example, if a VBScript
client passes an interface to the server, instead of passing a string value.
SilkPerformer tries to resolve type mismatches automatically.
12 In the Filter (record server types) area, select the Remote out-of-process
server option to capture and record the communication of the client
object with any remote out-of-process server (i.e., a server object that
operates on a different computer than the client object).
13 Select the Local out-of-process server option to capture and record the
communication of the client object with any local out-of-process server
(i.e., a server object that operates on the same computer as the client
object, but in a separate process space).
14 Select the In-process server option to capture and record the
communication of the client object with any in-process server (i.e., the
client and server objects are operating in the same process space in the
computer). In-process servers provide the fastest possible service to the
client. They are typically objects in the same program, or objects loaded
into the same process space ahead of time from an external source.
15 Click OK to exit the dialog and save your changes.
Procedure To set ADO recording options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Record category and click
the COM / ADO icon.
5 Click the ADO tab.
6 In the Script options area, select the Extract SQL statements option to
extract SQL statements which are passed to ADO interfaces to the dclsql
section.
116 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Project Profile Settings
COM/ADO result
options
7 Select the Fetch ADO Recordsets option if you want to make use of the
ADO recordset handling support. AdoFetch statements will be scripted
when the client application is browsing through recordset. This feature
reduces the script size significantly and allows ADO customizing with
the TrueLog.
8 Select the Script ADO Parameters option if you want to make use of the
ADO parameter handling support. AdoParam statements will be scripted
when the client application is using ADO parameters with ADO
commands. This feature reduces the script size significantly and allows
ADO customizing with the TrueLog.
9 Click OK to exit the dialog and save your changes.
Procedure To set COM/ADO result options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Replay category and click
the COM / ADO icon.
5 Click the Results tab.
6 In the Measures area, select the Measure for every COM call option to
get a separate measure for every call to a COM interface method.
7 Click OK to exit the dialog and save your changes.
WAP Options
Note SilkPerformer offers a tutorial that walks you through the
process of load testing WAP-based applications. The tutorial is
available in the Load Testing a WAP Application section of
Miscellaneous Tutorials (/
Documentation/Tutorials/...).
WAP options
Procedure To set WAP replay options:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, expand the Profiles folder.
2 Right-click the icon of the profile you want to configure.
SilkPerformer User Guide 117

3 CONFIGURING PROJECTS
Project Profile Settings
3 Choose Edit Profile.
The Profile - [Loadtest] - Simulation dialog opens.
Alternately, select Settings/Active Profile from the SilkPerformer menu
bar.
4 In the shortcut list on the left side, select the Replay category and click
the WAP icon.
5 On the Phone tab, you can specify WAP-browser-specific settings. From
the Phone drop-down list, select the WAP browser you want to use for
your simulation. The selection you make determines the format of the
header information included in your WAP requests used for load tests.
6 In the Max. outstanding requests field, specify the maximum number of
requests the client can send to the server without receiving a response.
Since the server must allocate more system resources for each request,
server performance decreases as the number of requests increases.
7 From the WAP version drop-down list, select the WAP version of the
browser that is to be simulated.
8 In the User agent field, specify user agent information sent to the server.
This is typically the browser name, including platform information. This
string is used for the HTTP request header “User-Agent:” that is sent to
the server.
9 In the Accept field, specify the document types that the server may send
to the client. This string is used in the HTTP request header “Accept:”
that is sent to the server.
10 In the Additional HTTP headers field, you can specify one or more
additional headers, using exact syntax. Use a new line for each header.
These headers will be appended to each request that is sent to the server.
11 Click the Gateway tab to specify gateway-specific options. A gateway
translates requests from the WAP protocol stack to the WWW protocol
stack.
12 In the Host name or IP address field, enter the name or IP address of the
WAP gateway.
13 Select a Connection type for the gateway:
- Select Connectionless to use an unsecured, connectionless
communication channel to send requests to and receive requests
from the server.
- Select Connection-oriented to use an unsecured, connection-oriented
communication channel to send requests to and receive requests
from the server.
118 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Test Scripts
- Select Secure Connectionless to use a secure, connectionless
communication channel to send requests to and receive requests
from the server.
- Select Secure connection-oriented to use a secure, connectionoriented
communication channel to send requests to and receive
requests from the server.
14 Once you’ve identified a connection type, enter the number of the Port to
which the WAP gateway responds for that particular connection type.
15 Click OK to save your settings.
Test Scripts
Adding test scripts
Compiling test scripts
Removing test scripts
Test scripts prescribe the actions of the virtual users that are run during load
tests. Scripts are written in SilkPerformer's proprietary scripting language, the
Benchmark Description Language (BDL). The usual way to create scripts is to
have the SilkPerformer Recorder generate them from real network traffic that
has been captured and recorded. Scripts can also be generated manually, or
sample scripts provided by SilkPerformer can be used or modified for reuse.
As many load test scripts as necessary can be added to a project, and existing
scripts can be removed. Scripts need to be compiled before they can be used,
and individual scripts in the project can be compiled, or all the scripts associated
with a particular project can be recompiled at once.
Procedure To add a test script:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, right-click the icon for the current project in the tree-view list.
2 Select Add Script from the context menu.
Alternately, select Project/Add Script from the SilkPerformer menu bar.
The Select Script(s) dialog appears.
3 Use the dialog to navigate to, and select, the script to add to the project.
Procedure To compile a test script:
1 Right-click the script you want to compile.
2 From the context menu, select Compile.
Alternately, select Script/Compile from the SilkPerformer menu bar.
Procedure To remove a test script:
1 Expand the folder of the current project in the tree-view list.
2 Expand the Scripts folder.
SilkPerformer User Guide 119

3 CONFIGURING PROJECTS
Data Files
3 Right-click the icon of the script you want to remove.
4 Select Remove Script from the context menu.
5 Click Yes on the deletion confirmation dialog.
Data Files
Adding data files
Removing data files
Data files that are required by SilkPerformer scripts are added to projects by
specifying the directory where the user data files (.csv), certificate files (.pem),
random data files (.rnd files), and type library files (.tlb, .dll, .exe, .ocx, .olb,
.pkg) are located. The compiler searches for random data files first in the
directory where the .bdf file is located, then in the directory specified. If a
random data file from a different location needs to be used, a full path for the file
can be specified in the script file.
Note It is important that all the user data files, random data files,
certificate files, .idl files, and type library files (.tlb, .dll, .exe, .ocx,
.olb, .pkg) needed by a test script are added to the project. Otherwise
they will not be available to remote agent computers during a test.
Procedure To add a data file:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, right-click the icon of the current project.
2 Select Add Data File from the context menu.
Alternately, select Project/Add Data File from the SilkPerformer menu
bar.
The Select Data File(s) dialog appears.
3 Use the dialog to navigate to and select the data file that you want to add
to the project.
Note It is important that all the user data files, random data files,
certificate files, .idl files, and type library files needed by a test script
are added to the project. Otherwise they will not be available to
remote agent computers during a test.
Procedure To remove a data file:
1 Expand the folder for the current project in the tree-view list.
2 Expand the Data Files folder.
3 Right-click the data file you want to remove.
4 Select Remove Data File from the context menu.
5 Click Yes on the deletion confirmation dialog.
120 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Workload Definition
Agent Computers
Adding agent
computers
Removing agent
computers
Agent computers host the virtual users that are run during load tests. As many
agent computers as necessary can be added to a project to support the required
number of virtual users.
Agent computers are assigned to particular projects from the pool of agent
computers available to the controller computer. The maximum number of
virtual users who will run on each agent is computed automatically by
SilkPerformer; Alternately, this number can be set manually. New agents can be
added to the agent pool from the local area network or from other Windows
domains; they will then be available to be added to the project. Agents can be
checked for availability, and agents can be removed from the pool.
Procedure To add an agent computer:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, right-click the icon of the current project in the tree-view list.
2 Select Add Agent from the context menu.
Alternately, select Project/Add Agent from the SilkPerformer menu bar.
The Add Agent(s) From Agent Pool To Project dialog appears.
3 Select an agent computer from the list and click OK to add it to the
project.
To add agents that don’t appear on the list, click the Configure Agent
Pool button and proceed as described in “Configuring agent pool
settings”.
Procedure To remove an agent computer:
1 Expand the folder of the current project in the tree-view list.
2 Expand the Agents folder.
3 Right-click the icon of the agent computer you want to remove.
4 Select the Remove Agent option.
5 Click Yes on the deletion confirmation dialog.
Workload Definition
A full range of options for the workloads to be employed in load tests is
provided. You determine the workload type to be used: queuing, increasing,
steady state, or dynamic workload. For each user group, numerous settings can
be set, including the number of virtual users in the group; simulation times,
warm-up times, and measurement times; and the allocation of virtual users to
SilkPerformer User Guide 121

3 CONFIGURING PROJECTS
Workload Definition
Adding workload
definitions
Editing workload
definitions
agent computers. The start time for load tests can be set, either relatively to the
current time (e.g., two hours from now), or absolutely (at a specific time of day).
All the configuration options for workload are available from the editing version
of the workload definition. From the execution version, it is also possible to run
load tests.
To define additional workloads for your load test and save them for future use,
expand the Workloads section in the tree view, right-click an existing workload
and select New Workload or Copy Workload from the context menu. If you
create a new workload model, the baseline will have to be accepted for the
model. You can copy previous accepted baseline results from an existing
workload model by copying and renaming an existing workload.
Procedure To edit the workload definition:
1 In the Project tab of the tree-view area of the main SilkPerformer
window, right-click the current project in the tree-view list.
2 Select Edit Active Workload from the context menu.
Alternately, select Settings/Active Workload from the SilkPerformer
menu bar.
Alternative In the Project tab of the tree-view area of the main
SilkPerformer window, right-click the workload you want to edit
and select Edit Workload, or double-click the workload.
3 The Workload Configuration dialog appears in the version for workload
editing only.
4 Use the Workload area to configure the workload that you will use in
your load test.
5 Choose a workload model: Increasing, Steady State, Dynamic, All Day,
or Queuing options.
- Select the Increasing option to use the increasing workload model.
At the beginning of a load test, SilkPerformer simulates not the total
number of users defined, but only a certain percentage of them. Step
by step, the workload is increased until all the users specified in the
user list are actually running.
This workload model is especially useful when you want to find out
at which load level your system crashes or does not respond within
acceptable response times or error thresholds.
- Select the Steady State option to use the steady-state workload
model in your load test. In this model, the same number of virtual
users is employed throughout the test. Every virtual user executes
the transactions defined in the load testing script. When work is
finished, the virtual user starts again with executing the transactions.
122 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Workload Definition
There is no delay between transactions. The test finishes when the
simulation time that was specified has been reached.
This workload model is especially useful when you want to find out
about the behavior of your tested system at a specific load level.
- Select the Dynamic option to use the dynamic workload model in
your load test. You can manually change the number of virtual users
in the test while it is being run. The maximum number of virtual
users to be run is set; within this limit, the number can be increased
or decreased at any time during the test. No simulation time is
specified, and you must bring the test to an end manually.
This workload model is especially useful when you want to
experiment with different load levels and to have the control over the
load level during a load test.
- Select the All Day option to use the all day workload model in your
load test. This workload model allows you to define the distribution
of your load in the most flexible way. You can assign different
numbers of virtual users to any interval of the load test. Each user
type can use a different load distribution. Therefore, you can design
very complex workload scenarios such as workday workloads or
even weekly workloads. You can adjust the load level during a load
test for intervals that have not yet started executing.
This workload model is especially useful when you want to model
complex, long lasting workload scenarios in the most realistic way
possible.
- Select the Queuing option to use the queuing workload model in
your load test. In this model, transactions are scheduled following a
prescribed arrival rate. This rate is a random value based on an
average interval calculated from the simulation time and the number
of transactions you have specified in your script (dcluser section:
number of transactions per user). The load test will finish when all of
the virtual users have completed their prescribed tasks.
Note that the test may take longer than the simulation time that was
specified, because of the randomized arrival rates. For example, if
you specify a simulation time of 3000 seconds and want to execute
100 transactions, then you will get an average transaction arrival rate
of 30 seconds.
This workload model is especially useful when you want to simulate
workloads that use queuing mechanisms to handle multiple
concurrent requests. Typically, application servers like servlet
engines or transaction servers, which are receiving their requests
from Web servers and not from end users, can be accurately tested
by using the queuing model.
SilkPerformer User Guide 123

3 CONFIGURING PROJECTS
Workload Definition
6 Use the figure at the top of the dialog to see a graphical representation of
the selected workload model. The figure uses the actual data of the user
group selected in the workload list below.
7 Select the Balance load across agents option to automatically distribute
the virtual users on the agent computers. This option is especially useful
in order to not overload the agent computers.
8 Use the workload list to configure the user groups that are to be run in
your load test.
9 In the User Type (Script / Usergroup / Profile) columns, select the user
groups that you want to run in your test. All of the user groups from your
test scripts for the current project are listed here. Also you can choose the
Profile that should be used for this user.
10 In the Max. Vuser column, specify the maximum number of virtual users
to be run in each user group.
11 To specify that your load test begin at a set period after the time given in
the Start Time area, enter a time value (in seconds) in the Start Time
Offset [s] column.
12 In the Simulation Time column, enter the duration (in seconds) for which
each user group will be run in your load test.
13 In the Warm-up Time column, enter the duration (in seconds) for the
period at the start of a load test during which no measurements will be
performed.
14 In the Measurement Time column, enter the duration (in seconds) for
which measurements will be performed during the load test.
15 If you selected the Increasing workload model, you have three additional
columns:
- Start Users: Specify the number of virtual users to be run at the
beginning of the load test.
- Add Users: Specify the number of virtual users to be added to the
current number of virtual users after each interval defined in the
Increase after column (see below).
- Increase after: Enter the duration of the interval after which virtual
users are to be added to the current number of virtual users.
16 In the Auto-assign column, select the assignment option for each user
group for which you want SilkPerformer to automatically compute and
assign virtual users to agent computers. To assign virtual users manually
to agent computers for a particular user group or groups, leave this
option unselected.
124 SilkPerformer User Guide

3 CONFIGURING PROJECTS
Outlining Projects
Running load tests
17 Use the Agents column to check SilkPerformer's automatic assignment of
virtual users to agent computers, or to manually assign them yourself.
In either case, click the browse (...) button to the left of the Agent column
for the relevant user group.
- For a user group using the Auto-assign option, the Automatic Agent
Assignment for Script / Usergroup appears, where you can view
SilkPerformer's assignment of virtual users to agent computers.
- For a user group not using the Auto-assign option, the Manual Agent
Assignment for Script / Usergroup appears.
18 In the Assigned Vusers column of the Agent Assignment for Script /
Usergroup dialog, enter the number of virtual users from the relevant
user group to run on each agent. All of the users from the user group
must be assigned to agents.
19 Click OK to return to the Workload Configuration dialog.
20 In the Start Time area, set the start time for your load test in hours,
minutes, and seconds. Select the Relative option to have your test start
after the time period you have specified has elapsed. Select the Absolute
option to have the test start at that specific time.
21 In the Loadtest Description field, you can optionally enter a description
of the load test. This feature is provided for project management
purposes only.
22 Click the User distribution Overview button to view the assignment of
virtual users to the agent computers that are currently available.
23 Click OK to save your changes.
Procedure To run a load test:
1 Click the Run Test icon on the SilkPerformer Workflow bar, or select
Run/Run Test from the SilkPerformer menu bar.
2 The Workload Configuration dialog appears in the version from which
tests can be run.
3 If you need to edit the workload definition, follow steps outlined in the
preceding section.
4 Click Run to execute the test.
Outlining Projects
The first step in conducting a SilkPerformer load test is to define the basic
settings for the load testing project. The project is given a name, and optionally a
SilkPerformer User Guide 125

3 CONFIGURING PROJECTS
Outlining Projects
Outlining a project
brief description can be added. The type of application to be tested is specified
from a range of choices that includes all the major traffic that is encountered in
e-business today on the Internet, including the most important database and
distributed applications.
Procedure To outline a load testing project:
1 Click the Outline Project button on the Workflow bar.
2 The Workflow - Outline Project dialog opens.
3 In the Project name field, enter a name for the project.
4 In the Project description field, you can optionally enter a description for
the project. This feature is provided for project management purposes
only.
5 From the Application type list, select the type of application on which
you want to run your load test.
6 Click OK.
126 SilkPerformer User Guide

4
Modeling 4Chapter
Test Scripts
Introduction
What you will learn
This chapter explains how to model load testing scripts in SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 127
Recording Test Scripts 128
Writing Test Scripts Manually 131
Working From Sample Scripts 131
Trying Out Test Scripts 132
Overview
The second step in conducting a SilkPerformer load test is to create the test
script that will prescribe the actions of the simulated users who will be run
during the test. The script is written in SilkPerformer's proprietary scripting
language, the Benchmark Description Language (BDL).
Scripts can be created in two ways. The easiest way is to use the SilkPerformer
Recorder, which can automate much of the process. An alternate way, for
experienced users, is to create test scripts manually.
• “Recording Test Scripts”: The standard method of creating a test script is
to use SilkPerformer's recording engine, the SilkPerformer Recorder, to
first capture and then record a representative amount of real traffic of the
kind that needs to be simulated in the load test. The SilkPerformer
Recorder then automatically generates the BDL test script from the
recorded traffic.
SilkPerformer User Guide 127

4 MODELING TEST SCRIPTS
Recording Test Scripts
• “Writing Test Scripts Manually”: A second method of creating test
scripts is to manually create them from scratch in BDL.
• “Working From Sample Scripts”: A variant on the second method of
creating test scripts is to base a new script on one of the sample BDL
scripts that is provided with SilkPerformer.
Recording Test Scripts
The standard, and easiest, method of creating a load test script is to use the
SilkPerformer Recorder, SilkPerformer's engine for capturing and recording
traffic and generating test scripts.
The SilkPerformer Recorder can record the following types of Internet traffic
(secure and non-secure):
- Hypertext Transfer Protocol (HTTP)
- File Transfer Protocol (FTP)
- Post Office Protocol (POP)
- Simple Mail Transfer Protocol (SMTP)
- Lightweight Directory Access Protocol (LDAP)
- Internet Inter-ORB Protocol (IIOP)
- SOCKS
The Recorder can capture function calls of the following APIs (see “Available
application types” for the complete list):
- Oracle Call Interface (OCI)
- Open Database Connectivity (ODBC)
- TUXEDO
- Jolt
- COM
- WinSock
- Oracle Forms
- Oracle Applications
- Citrix
- .NET
- ADO
- WAP
128 SilkPerformer User Guide

4 MODELING TEST SCRIPTS
Recording Test Scripts
- Java
- SAPGUI
- Remedy
- Flash
First, the SilkPerformer Recorder captures and records a representative amount
of real traffic between a client application and the server to be load tested. When
recording is complete, the SilkPerformer Recorder then automatically generates
the test script from the recorded traffic. The script is written in SilkPerformer's
proprietary scripting language, the Benchmark Description Language (BDL).
During the recording phase, you define transactions. A transaction is a discrete
piece of work that will later be assigned to a virtual user in a load test and for
which separate time measurements will be made. You should create a new
transaction only for a piece of work that has no dependencies on another piece
of work. Individual time measurements can be made for any action or series of
actions that happen during recording.
Advantages
Using the SilkPerformer Recorder has a number of advantages:
- Recorded traffic and function calls include a lot of static
information. Replay of static information ordinarily does not lead to
reliable benchmark results. Because recorded scripts can easily be
customized, you can add random functions to generate dynamic and
realistic workloads.
- Transactions defined during recording modularize the script inTO
clearly laid out sections. It is easy to modify workloads based on
different statistics or requirements using the randomization functions
embedded in SilkPerformer.
- Timers set during recording provide user-defined granularity for
measuring any part of a transaction.
- Development time and costs can be reduced considerably. There is
no need to change the front-end application to simulate new
behaviors. It is sufficient to modify the script and repeat the
simulation until the requirements are met. The front-end application
has to be implemented only once; the script then becomes its
prototype and reproduces its behavior.
Recording test scripts
Procedure To record a test script:
1 Open the project in which you want to work, or start a new project.
2 Click the Model Script button on the SilkPerformer Workflow bar.
The Workflow - Model Script dialog appears.
SilkPerformer User Guide 129

4 MODELING TEST SCRIPTS
Recording Test Scripts
3 In the Script area, select the Record option.
4 From the Select application profile drop-down list, select the profile for
the client application that you plan to test in your load test.
If a profile has not yet been set up for the application you want to use,
click the button to the right of the drop-down list to set one up.
To add a new profile, see “Adding application profiles”.
5 If you are load testing a Web application, enter its URL in the URL field.
6 Click OK.
The SilkPerformer Recorder dialog opens in minimized form, and the
client application starts.
7 To see a report of the actions that happen during recording, maximize the
Recorder dialog by clicking the Change GUI size button.
The maximized Recorder then opens at the Actions tab.
8 Using the client application, conduct the kind of interaction with the
target server that you want to simulate in your load test. The interaction
is captured and recorded by the Recorder. A report of your actions and of
the data downloaded appears in the Actions tab.
9 Insert transactions and timers into the test script during the recording
phase. You can create as many transactions and timers as you want.
10 To insert a transaction, click the New Transaction button. A transaction
represents a piece of work that can be assigned to a virtual user.
11 In the ensuing dialog, enter a name for the transaction and click OK.
The new transaction appears in the Actions log.
12 To insert a timer, click the New Timer Session button. A timer is a userdefined
measurement period in a load test. You should create timers for
each component of a transaction for which you want to analyze
performance.
In the ensuing dialog, enter a name for the timer and click OK.
13 To end recording, click the Stop Recording button. The test script is
generated from the recorded traffic.
14 In the ensuing dialog, name your test script and specify where to save it.
15 In the SilkPerformer Recorder message that appears, confirm that you
want to close the Recorder.
The new script appears in a script editor window.
130 SilkPerformer User Guide

4 MODELING TEST SCRIPTS
Working From Sample Scripts
Writing Test Scripts Manually
Writing a new test
script
For those who do not want to use the standard method of generating load test
scripts (using the SilkPerformer Recorder), an alternate method is to manually
create new scripts from scratch.
SilkPerformer test scripts are written in the program's proprietary scripting
language, the Benchmark Description Language (BDL). To generate scripts
manually, you will need to be familiar with BDL, which is a high-level language
resembling Pascal. You should be able to create prototypes of user requests, and
define the typical components of a load testing script, including modules,
functions, workload definitions, transactions, Web forms, and more.
Procedure To write a new test script:
1 Open the project in which you want to work, or start a new project.
2 Click the Model Script button on the SilkPerformer Workflow bar.
The Workflow - Model Script dialog appears.
3 In the Script area, select the New script option.
A SilkPerformer script template opens in a new script editor window.
4 Use Benchmark Description Language (BDL), SilkPerformer's
proprietary scripting language to write your script. Refer to the BDL
Function Reference for detailed information about the Benchmark
Description Language.
5 Save your new script.
6 Follow the steps in the SilkPerformer Workflow to use the new script in
your load test.
Working From Sample Scripts
Sample script reuse is a timesaving possibility for those who want to generate
SilkPerformer load test scripts manually (and bypass the standard method,
which uses the SilkPerformer Recorder).
SilkPerformer test scripts are written in the program’s proprietary scripting
language, the Benchmark Description Language (BDL). Testers who want to
generate scripts manually will need to be familiar with BDL, which is a highlevel
language resembling Pascal. A number of sample BDL scripts are
provided with SilkPerformer, and testers can base the new script on one of these.
To work from a sample script you need to be familiar with how to create
prototypes of user requests, and how to define the typical components of a load
SilkPerformer User Guide 131

4 MODELING TEST SCRIPTS
Trying Out Test Scripts
testing script, including modules, functions, workload definitions, transactions,
Web forms, and so on.
Using a sample test
script
Procedure To re-use a sample test script:
1 Open the project in which you want to work, or start a new project.
2 Click the Model Script button on the SilkPerformer Workflow bar.
The Workflow - Model Script dialog appears.
3 In the Script area, select the Sample script option.
4 Click OK.
The Open dialog appears
5 Browse to the Samples folder.
6 Browse to the folder of the specific application type you plan to test.
7 Select a sample script and click Open.
8 In the ensuing dialog, confirm that you want to add the script to the
current project.
9 The script is added to the project, and opens in a new script editor
window.
10 Follow the steps in the SilkPerformer Workflow to use the sample script
in your load test.
Trying Out Test Scripts
The third step in conducting a SilkPerformer load test is to perform a trial run of
the Benchmark Description Language test script that was created in the previous
step. The object of a trial run is to ensure that the test script is free from error,
and that it accurately reproduces the required interaction between the client
application and the server. Normally, this will be the traffic that was recorded by
the SilkPerformer Recorder in the previous step.
For this try-out of the test script, options are automatically selected so that you
can see a live display of the actual data downloaded. Log files, TrueLog files,
report files, output files, and error files are created so that you can also check
later that the script is working properly. Only one user is run, and the stress test
option is enabled so that there is no think time and no delay between
transactions. However, at this stage, the measurements typical of a real load test
are not performed.
For Try Script runs, the following options are automatically set to these
specified values:
- The Stress test option is on.
132 SilkPerformer User Guide

4 MODELING TEST SCRIPTS
Trying Out Test Scripts
Trying out test scripts
- The Stop virtual users after simulation time (Queuing Workload)
option is off.
- The Virtual user log files (.log) option is on.
- The Virtual user output files (.wrt) option is on.
- The Virtual user report files (.rpt) option is on.
- The Virtual user report on error files (.rpt) option is on.
- The TrueLog files (.xlg) option is on.
- The TrueLog On Error files (.xlg) option is off.
- The Compute time series data (.tsd) option is off.
- All logging detail options (Results/Logging and Results/Internet
logging tab) are on.
- The Enable all measure groups (TSD measure groups) option is off.
- The Bandwidth option is set to High Speed (unlimited).
- The Downstream option is set to unlimited.
- The Upstream option is set to unlimited.
- The Duplex option is off.
- The Modem compression rate is set to 1.0 x.
Procedure To try out a test script:
1 Click the Try Script button on the SilkPerformer Workflow bar.
The Workflow - Try Script dialog appears.
2 The currently active profile name is displayed in the Profile field. To edit
the script replay settings for this profile, click the button to the right of
the field. Note that appropriate settings for script try-outs are
automatically selected by SilkPerformer.
3 From the Script drop-down list, select the script that is to be tried out.
4 Select the Animated option for a live view of the data downloaded during
the test.
5 To configure project attributes, click the Project Attributes button. See
Project Attributes Overview in the BDL Function Reference for details
on these attributes.
6 Click Run to execute your test script try-out.
7 During the trial run, monitor the test progress in the Monitor window,
and view downloaded data using TrueLog Explorer.
SilkPerformer User Guide 133

4 MODELING TEST SCRIPTS
Trying Out Test Scripts
134 SilkPerformer User Guide

5
Validating 5Chapter
Test Scripts
Introduction
What you will learn
This chapter explains how to validate test scripts.
This chapter contains the following sections:
Section
Page
Overview 135
Visual Analysis with TrueLog Explorer 135
Log Files 136
Report Files 137
Output Files 138
Overview
SilkPerformer offers several means of evaluating test scripts after the execution
of TryScript runs.
Visual Analysis with TrueLog Explorer
TrueLog Explorer guides you through the process of visually analyzing Try
Script runs with its Workflow bar. The steps are as follows:
1 Examine the Virtual User Summary report.
2 Find replay errors.
3 Compare a replay TrueLog with the TrueLog that was produced during
recording of the application.
SilkPerformer User Guide 135

5 VALIDATING TEST SCRIPTS
Log Files
TrueLog Explorer provides an accurate view of the actual data that is
downloaded during load tests. This enables you to verify that the data specified
to be received from the server under test was, in fact, received. In the case of
Web application testing for example, TrueLog Explorer shows the actual Web
pages that are received during tests. Through the use of TrueLog Explorer’s
animated mode, live monitoring of downloaded data is available—data is
displayed as it is received during tests.
TrueLog Explorer provides different views of received data, enabling you to:
- Inspect the rendered HTML code, as a real user would see it.
- Inspect the native HTML code.
- Inspect a difference table that shows the differences between replay
and recorded TrueLogs.
Viewing Web pages
accessed during tests
Procedure To view the Web pages that were accessed during a load test:
1 Checking the Animated checkbox before running a Try script operation
opens TrueLog Explorer automatically.
Alternately, TrueLog Explorer can be accessed by selecting Results/
Explore TrueLog from the SilkPerformer menu bar.
2 A Virtual User Report is shown automatically when the first node of the
tree view at the left side of the window is checked. In the report you will
find detailed information about the Try Script run in tabular form.
3 Use the tree view area on the left of the window to expand and collapse
folders as necessary to locate and view data downloaded during the test.
4 To view a downloaded Web page, select its URL or link description in
the tree view area.
5 To locate replay errors, click the Analyze Test button on the workflow
bar and select Find Errors. A new window appears that helps you
navigate from one error to the next. HTML errors often occur one or two
steps before they are apparent to end users. With TrueLog Explorer you
can easily examine the Web pages that were displayed preceding errors.
Log Files
Any errors that are encountered in load tests are reported in log files (assuming
log file generation has been enabled). Log files contain a record of all the
function calls that are invoked by the transactions of each virtual user in load
tests. These files contain session and error information, including details for
each type of event executed during a simulation, the API calls in a transaction,
and the data exchanged with the server, as well as other information.
136 SilkPerformer User Guide

5 VALIDATING TEST SCRIPTS
Report Files
However, generating log files alters the time measurements of load tests.
Therefore, these files should be used for debugging purposes only, and should
never be generated for full load tests.
Viewing log files
Procedure To view log files:
1 Select File/Open/Virtual User Log File from the SilkPerformer menu
bar.
2 Locate the appropriate file (or right-click a virtual user in the
SilkPerformer Monitor window and select Show Log File). The log file
will then display.
Report Files
Report files contain the same information as baseline reports for one specific
virtual user. This report is based on XML/XSL and includes the most important
test results in tabular format. It comprises the following sections:
- General Information
- Virtual User
- Summary Tables
- Transaction Response Times
- HTML Page Timers
- Web Form Measurements
General information
User types
Summary tables
Transactions
The general information section includes administrative information in a tabular
form.
Administrative information includes the SilkPerformer version info, the project
name, a description of the project, the date and time of the load test, the
workload definition, the workload model and the number of errors that occurred.
For the virtual user, a section is available with details of the measured response
times. The summary line shows the duration of the test, the session time, the
session busy time, the average page time, the number of transactions executed
successfully, cancelled, failed and the number of errors, if any, that occurred.
This section contains summary measurements in a tabular form, that is,
aggregate measurements. The first table provides general information, such as
the number of transactions that were executed and the number of errors, if any,
that occurred. All the following tables provide summary information relevant to
the type of application that was tested.
This section contains summary measurements in a tabular form, that is,
aggregate measurements for all transactions of the specific user. For every
SilkPerformer User Guide 137

5 VALIDATING TEST SCRIPTS
Output Files
Page timers
Web forms
Viewing report files
transaction, the transaction response time and the transaction busy time are
displayed.
The transaction response time is measured from the beginning to the end of the
transaction, including all think times.
The transaction busy time is the transaction response time without any think
time included.
This section contains summary measurements in a tabular form for every
accessed Web page. For every page, the download time of the whole page, the
HTML-document download time, the server busy time, the amount of data
downloaded with this page, and the amount of data downloaded for embedded
objects are displayed.
This section contains summary measurements in tabular form for every used
Web form. For every form, the round trip time, the server busy time, the number
of HTTP hits, the amount of request data sent and the response data received are
displayed.
Procedure To view report files:
1 If TrueLog Explorer opens, the Virtual User Report File will be
displayed automatically.
2 You can access report files by selecting File/Open/Virtual User Report
File from the SilkPerformer menu bar and then locating the appropriate
file, or by right-clicking the line for a virtual user in the SilkPerformer
Monitor window and selecting Show Report File from the context menu.
The report file is then displayed.
Output Files
Viewing output files
Output files contain the output of write statements that are used in test scripts.
An output file is generated for a particular user only if write statements are
executed by that user.
However, generating output files alters the time measurements of load tests.
Therefore, these files should be used for debugging purposes only, and should
not be generated for full load tests.
Procedure To view output files:
1 Select File/Open/Virtual User Output File from the SilkPerformer menu
bar.
2 Locate the appropriate file (or right-click a virtual user in the
SilkPerformer Monitor window and select Show Output File from the
context menu). The output file will then display.
138 SilkPerformer User Guide

6
Customizing 6Chapter
Test Scripts
Introduction
What you will learn
This chapter explains how to customize your test scripts with session handling,
verification functions, and more.
This chapter contains the following sections:
Section
Page
Overview 139
Inserting Function & Transaction Declarations 141
Customizing Workload 142
Customizing Session Handling 149
Inserting Parsing Functions 153
Adding Verifications to Test Scripts 155
Manual Script Editing 159
Using the System Configuration Manager 160
Serializing and Synchronizing Virtual Users 160
Using Custom DLLs in Test Scripts 166
Matrix: Functions, the Recorder, and TrueLog Explorer 174
Overview
The fourth step in conducting a SilkPerformer load test is to customize the test
script. If a test script is used that has not been customized, the load tests will
repeatedly reproduce the behavior recorded on that one particular occasion, and
this will not provide very accurate results. To realistically emulate the varied
traffic of a large number of different users, the transactions in the test script need
to be modified and, where necessary, randomized.
SilkPerformer User Guide 139

6 CUSTOMIZING TEST SCRIPTS
Overview
Web server applications often generate session IDs at runtime that are necessary
for further client requests. If a session ID needs to be customized, TrueLog
Explorer assists you by inserting the right BDL functions into the script.
Virtual users can be furnished with different personal data, such as names,
addresses, credit card numbers, and so on. They can vary in the products they
search for information about, in the products they order, and in many other
ways. SilkPerformer provides extensive facilities for customizing virtual users.
- “Adding Verifications to Test Scripts”: The content of HTML page
titles, HTML text and HTML tables can be verified during the
execution of a load test. In order to add verification functions to the
script, TrueLog Explorer allows you to visually identify and then
mark the text to be verified in the respective HTML page. The BDL
script will be automatically updated to reflect your verifications.
- “Inserting Parsing Functions”: Similar to the verification functions,
the parsing functions can be used to parse content returned from a
Web server and check if the values meet your testing criteria.
- “Defining modem speeds”: Virtual users can be customized to use
the bandwidth associated with any of the major connection types in
wide use among consumers today.
- “Defining browser types”: A wide choice of Web browsers, and of
the functionality they embody, is available to be assigned to virtual
users.
- Parameter Wizard: Realistic and varied personal data can be created
for the virtual users in a test script by means of parameters, which
can be easily generated using this wizard. For details on the
Parameter Wizard, please refer to TrueLog Explorer User Guide.
- “Customizing Workload”: The Customize User Groups dialog
enables you to add new user groups, copy user groups, rename user
groups, and delete user groups.
- “Manual Script Editing”: A common way to customize the behavior
of virtual users is to manually edit the test scripts using BDL, the
Benchmark Description Language.
- “Using the System Configuration Manager”: When the application
being tested uses load balancers, realistic simulation can be provided
by giving virtual users individual IP addresses.
140 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Inserting Function & Transaction Declarations
Inserting Function & Transaction Declarations
Inserting function
declarations
SilkPerformer offers GUI tools that assist you in inserting function and
transaction declarations into test scripts.
Procedure To insert a function declaration into a test script:
1 With the test script into which you want to insert a new function open,
select Script/New Function from the SilkPerformer menu bar.
The New Function dialog appears.
You can also access this dialog by right-clicking within the body of the
script and selecting New Function. A function is used to encapsulate
some tasks that a script requires. Functions can access and change data
declared outside themselves, either through variables or parameters.
Optionally, a function can return a value of a specific type.
2 In the Name field, enter a name for the function you want to insert into
your test script. The name must begin with a character and may consist
of any number of letters of the alphabet, digits, and underscores.
3 From the Return type drop-down list, select the data type of the
function's return value. Choose one of the following:
- none - for the function to return nothing
- boolean - for the function to return either true or false
- float - for the function to return any positive or negative floatingpoint
value or zero
- number - for the function to return any positive or negative whole
number or zero
- string - for the function to return a sequence of characters
4 In the Parameters list, insert the parameters that you want to pass to the
function. For each parameter, enter a name in the Name column and
select the data type from the Type drop-down list:
- boolean - for the parameter to contain either true or false
- float - for the parameter to contain any positive or negative floatingpoint
- value or zero
- number - for the parameter to contain any positive or negative whole
number or zero
- string - for the parameter to contain a sequence of characters
SilkPerformer User Guide 141

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
Inserting transaction
declarations
5 Use the Preview area to see how the function declaration looks when it is
inserted into your script. The declaration contains the parameters you
insert into the list as well as the return type. Keep in mind that it is up to
you to add the function code after the function declaration has been
inserted into the script.
6 Click OK.
The new function is placed in your script, and appears in the script editor
window.
Procedure To insert a new transaction declaration:
1 With the test script into which you want to insert a new transaction open,
select Script/New Transaction from the SilkPerformer menu bar.
2 The New Transaction dialog appears.
3 You can also access this dialog by right-clicking in the body of the script
and selecting New Transaction.
4 In the Name field, enter the name of the transaction you want to insert.
The name must begin with a character and may consist of any number of
letters of the alphabet, digits, and underscores.
5 Click OK.
The new transaction is placed in your script, and appears in the script
editor window.
Customizing Workload
“User groups” are groups of virtual users who share similar, though not
identical, behavior. The members of user groups perform slightly different
transactions with the goal of simulating real-world client behavior while
maintaining load balance.
User groups are used in conjunction with parameterized user input data and
profiles (browser types, connection speeds, etc) to simulate typical end-user
activity.
User groups are defined by the “transactions” that they call and the frequency at
which they call those transactions. For example, you might have a user group
called “Searcher” that performs a TLogon transaction once, a TSearch
transaction five times, and a TLogoff transaction once. You might also have a
user group called “Updater” that performs the TLogon transaction once, the
TSearch transaction three times, the TUpdate transaction twice, and the TLogoff
transaction once. A third user group called ”Inserter” might perform the TLogon
transaction once, the TSearch transaction once, the TInsert transaction three
times, and the TLogoff transaction once. Collectively, these user groups can be
142 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
used to simulate real-world client activity while maintaining load balance (i.e.
transactions are distributed in such a way that all users don’t perform the same
transaction at the same moment).
By assigning different profiles to a user group and a script you can define new
“user types,” which are unique combinations of scripts, user groups, and
profiles.
When multiple users have the same workload profile, they can be combined into
a user group. The workload description of a user group contains the name of the
user group, the transactions that are to be invoked by the user group, and the
frequency at which the transactions are to be invoked (TransCount). User group
workload descriptions are included in the workload sections of SilkPerformer
test scripts (UserDcl).
Visual user group
customization
The Customize User Groups dialog enables you to add new user groups, copy
user groups, rename user groups, and delete user groups. The Customize User
Groups dialog also enables you to insert new transactions, copy transactions,
remove transactions, and rename transactions. Transactions can be added and
removed for each user group. The transaction renaming process renames
transactions in all user groups to which they have been assigned.
The Customize User Groups dialog is available via SilkPerformer’s Script menu
(Script/Customize User Groups). This menu option is only available when a
BDF script is open and active. The Customize User Groups dialog works only
on the active BDF script.
Procedure To customize a user group or transaction:
1 Open and activate the BDF script you want to customize.
2 Select Customize User Groups from SilkPerformer’s Script menu.
3 Right-click a user group or transaction node to open a context menu that
lists the available operations:
User Group Operations:
- Add a new user group
- Copy a user group
- Rename a user group
- Delete a user group
Transaction Operations:
- Insert a transaction
- Copy a transaction
- Rename a transaction
- Remove a transaction
SilkPerformer User Guide 143

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
4 Select the desired action from the context menu and follow the prompts.
5 Begin and End transactions can be defined for each user group by
selecting/deselecting the checkboxes in the Begin/End column.
6 Entries in the Count column, which represent the number of times that
transactions are to be called, can be edited directly by clicking them.
7 When the Choose transactions randomly checkbox is selected and a user
group contains at least two transactions that are neither “begin”
transactions nor “end” transactions, the Random Order column shows
how the transaction distribution will be weighted, by percentage. For
example, if two non-begin and non-end transactions are present, each
transaction will have a random order distribution of 50%. If three nonbegin
and non-end transactions are present, the transactions will have
random order distributions of 33%, 33%, and 34%.
8 Click OK when you’ve completed user group and transaction
customization. The modified data will then be written to the dcluser
section (and dcltrans section if necessary) of your test script.
User Profiles
Defining browser
types
Defining modem
speeds
Virtual users in a load test can be customized to use any of a wide choice of Web
browsers, and of the functionality they embody. The most popular browsers in
use today can be emulated, with a wide range of versions of Microsoft Internet
Explorer and of Netscape Communicator. Lesser-known browsers are also
available, as are browsers serving mobile phone users. Custom browsers can be
defined, using different versions of threading technologies and of HTTP.
Modems are the means that home users typically use to access the Internet.
Since a user’s modem is usually the slowest link in the network communication
chain, modems need to be simulated in load tests.
Virtual users can be customized to use the bandwidth associated with any of the
major connection types in wide use among consumers today—standard voiceline
modems, ISDN, Digital Subscriber Line, and mobile phone. Custom
settings can be defined where connections use different bandwidth settings for
downstream (from the server to the client) and upstream (from the client to the
server) traffic.
User Types
Setting up individual
IP addresses
Procedure To set up individual IP addresses:
1 Click the Customize Test button on the SilkPerformer Workflow bar.
The Workflow - Customize Test dialog appears.
144 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
2 Click the System Configuration Manager link.
The System Configuration Manager opens.
3 Click the IP Address Manager tab.
You can also access this by selecting Tools/System Configuration
Manager from the SilkPerformer menu bar, and clicking the IP Address
Manager tab.
4 From the Adapter drop-down list, select the network adapter for which
you want to configure IP addresses. The list contains all the network
adapters to which IP is bound on the computer in question.
5 Use the IP addresses area to find out information about the current
configuration of the network adapter, and to set up IP addresses for the
adapter.
6 To add new IP addresses to the network adapter configuration, click the
Add button.
The Add IP Addresses dialog appears.
7 In the From IP address field enter the IP address you want to add to the
network adapter configuration. If you specify that you want to add
multiple addresses, the System Configuration Manager generates
sequential IP addresses, starting with this address.
8 In the Subnet mask field, enter the subnet mask number for the IP
addresses. This number, combined with the IP addresses, identifies
which network the computers are on.
9 To add multiple IP addresses, either select the Number option and enter
the number of addresses you want in the field; or select the To IP address
option.
In the latter case, specify the highest IP address you want to add to the
network adapter configuration. The System Configuration Manager
generates sequential IP addresses, starting with the address you type in
the From IP address field and ending with this address.
10 Select the Do not add conflicting addresses option to only add IP
addresses to the network adapter configuration that do not cause
conflicts.
11 When enabling the Allow CIDR addresses checkbox, the verification of
the IP address and subnet mask pair is disabled. Classless Inter-Domain
Routing (CIDR) has been adopted as a solution to the scaling problem in
the Internet. CIDR is not as restrict as the "normal" IP-Routing.
12 Click OK.
13 To edit a selected IP address, click the Edit button.
The Edit IP Addresses dialog appears.
SilkPerformer User Guide 145

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
14 Edit the IP address and the subnet mask.
15 Select the Do not add conflicting addresses option to automatically
remove the modified IP address from the network adapter configuration
if it causes any conflicts.
16 Click OK.
17 To save the network adapter configuration to a file, click the Save button.
By default, the file is stored in the System Configuration Manager home
directory.
18 To load a network adapter configuration from a file, click the Load
button and locate the appropriate file.
19 To remove IP addresses, click the Remove button. The Remove IP
Addresses dialog appears.
Note Be cautious about removing multiple IP address at once.
20 In the From IP address field, enter the IP address you want to remove
from the network adapter configuration. If you elect to remove multiple
addresses, the System Configuration Manager deletes all successive IP
addresses, starting with this address.
21 In the Subnet mask field, enter the subnet mask number for the IP
addresses you want to remove.
22 To remove multiple IP addresses, select either the Number or the To IP
address option.
Select the Number option to specify the number of IP addresses you want
to remove from the network adapter configuration, and enter the number
of addresses in the field. The System Configuration Manager deletes that
number of successive IP addresses, starting with the address you type in
the From IP address field.
23 Select the To IP address option to specify the highest IP address you
want to remove from the network adapter configuration. The System
Configuration Manager deletes all successive IP addresses, from the
address you type in the From IP address field to this address.
24 Click OK.
25 If you want to check whether IP routing to a specified computer works
for all configured IP addresses, enter the name or IP address of the
computer in the Host name or IP address field in the Network area.
26 Click Check. A report from this check will be provided at the bottom of
the dialog.
27 Click OK to exit the dialog and save your changes.
146 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
Working with Web
forms
Web forms are defined in the dclform section of test scripts.
Form declaration syntax
The form declaration is specified by a form name (for example, FORM_
NAME) which can be any valid identifier string and must be followed by a
colon. The left side string specifies the name of the field or parameter. An empty
string is allowed for nameless URL parameters, such as image map coordinates.
The right value specifies the data value of the field or parameter. These can be
global constants, global variables and random variables.
The example below illustrates the required syntax for Web form declarations:
dclform
FORM_NAME:
"field-name" := R-Value;
Web form string encoding
All strings used in the forms declaration will be URL encoded when sent to the
server. This means that unsafe characters will automatically be converted into a
hexadecimal sequence (%xx). Parameters added directly to the URL will not be
encoded (for example, WebUrl("http://host/cgi-bin/doit?MyUnSafeMessage&-
?%?-?", 0.0). This allows applications to use non-standard URL parameters.
Web form script example
The following example shows a portion of a SilkPerformer script for simulating
a form submission. Random variables rsEmail and rsPlatform with two different
random functions (RndFile and RndInd) are defined in the random variables
section. With each subsequent call to WebFormPost and WebFormGet, the
random variables are refreshed and contain new random values. These random
values are used to generate the URL encoded form string automatically sent
within the request to the Web server. WebFormPost adds the form content string
at the end of the HTTP header while WebFormGet adds it at the end of the URL
separated by a “?”.
dclrand
rsEmail : RndFile
("elname.rnd", 20);
rsPlatform : RndInd("Windows NT" = 0.3;
"Windows 2000" = 0.6;
"other" = 0.1);
dcltrans
transaction TWebFormPost
begin
WebFormPost("http://www.comp.com/cgi/FormMail.pl", CGI_
CUST_SUPPORT, nWait);
WebFormGet("http://www.comp.com/cgi/form", CGI_CUST_
SUPPORT, nWait);
end TWebFormPost;
dclform
CGI_CUST_SUPPORT:
"email" := rsEmail,
SilkPerformer User Guide 147

6 CUSTOMIZING TEST SCRIPTS
Customizing Workload
"sitetype" := "business",
"sitelang1" := "english",
"platform" := rsPlatform;
Working with cookies
SilkPerformer supports record and replay for cookies used in Web applications.
Cookies enable Web servers to store and retrieve state information on the client
computer. A server, when submitting an HTTP object, may also send a
“cookie,” a state object that the client will store for a specified amount of time.
The cookie includes a description that defines a range of URLs for which it is
valid. The next time a client reconnects from within this range of URLs, the
contents of the cookie will be submitted to the server in addition to the request.
Cookies may either be persistent or non-persistent. Persistent cookies are stored
in a cookie database on the client computer and are deleted only after a specified
expiration date occurs. Non-persistent cookies are maintained only during a
single Web browser session and are deleted when the client Web application is
closed. To create as realistic a simulation as possible, only persistent cookies are
recorded (and thus replayed) by the Internet Recorder.
Cookie database functions
SilkPerformer creates a non-persistent cookie database for each virtual user.
During a load test, cookies are automatically received into and sent from the
database. To allow you to evaluate or modify cookies during a load test,
WebCookie functions are provided for users to set cookies into, and access them
from, the cookie database.
Function
Description
WebCookieGet Returns cookies for the specified URL and all its
parent URLs.
WebCookieSet
Sets a cookie in SilkPerformer’s cookie database
for the specified URL.
WebCookieListFirst Retrieves the first cookie in the cookie database.
WebCookieListNext Retrieves the next cookie in the cookie database.
WebCookieResetDb Deletes all cookies from the cookie database.
Table 2: Cookie database functions
Before you use these functions, you should be familiar with cookies as outlined
at http://www.netscape.com/newsref/std/cookie_spec.html.
Important notes
Keep in mind the following when you work with cookies with SilkPerformer:
- The SilkPerformer cookie database is non-persistent and therefore
only exists while a load test is running.
148 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Customizing Session Handling
- Each virtual user has a cookie database of its own and therefore the
contents of the cookie database varies from virtual user to virtual
user.
- Although cookies are generally created at server side, you can
simulate already existing cookies using the WebCookieSet function.
Customizing Session Handling
Session handling for
Web applications
Web server applications often generate information at runtime that is necessary
in order to identify further client requests. In the response to the browser, the
server may include a unique string, commonly known as the Session ID. This
string is returned by the browser to the server as a part of each subsequent
request, allowing the server to identify the unique Web session of which the
request is a part. Generally, Session IDs refer to the method the Web server
application uses to identify individual users and to associate this identification
with the state of the user session information that the application has previously
had with those users.
Session IDs can be sent to the client in a number of ways. Most often you will
find them included in cookies, or inside HTML as part of URL's used in hyper
links or embedded objects, or in hidden HTML form fields. Session IDs are sent
back to the server in cookies, URL's, and HTTP post data.
Example
Session information inside cookies:
From the server:
Set-Cookie: SessionID=LGIJALLCGEBMIBIMFKOEJIMM; path=/
To the server:
Cookie: SessionID=LGIJALLCGEBMIBIMFKOEJIMM
Session information in the URL's of HTML links:
From the server:
...
Enter Shop
...
To the server:
GET /ShopIt/acknowledge.asp? SessionID =
LGIJALLCGEBMIBIMFKOEJIMM HTTP/1.1
SilkPerformer User Guide 149

6 CUSTOMIZING TEST SCRIPTS
Customizing Session Handling
Session information in hidden form fields:
From the server:
...
Currently we only accept Credit Cards
...
To the server:
POST /ShopIt/kindofpayment.asp HTTP/1.1
...
SessionId=LGIJALLCGEBMIBIMFKOEJIMM&name=Jack&paymentButt
on=Submit
Value of customized
session handling
When to use
customized session
handling
Assuming that a WebPageUrl call in your script uses a URL that contains a
session ID as part of the query string of the URL. When replaying the script, this
hard coded static session ID is sent to the server. Because the session ID does
not correctly identify your replayed session—it still identifies the recording
session—the replay won’t work correctly. By not replacing static session IDs in
your script with dynamic values that had been generated at runtime, your Web
application will usually generate errors such as “We are sorry, your session has
expired. Please return to the login screen and try again”.
The good news is that with SilkPerformer, session customization is not often
necessary and, even if manual session customization is needed, SilkPerformer's
TrueLog Explorer will guide you through the process.
Most often when you record a script, it will work without any modifications
necessary for the customization of session handling. SilkPerformer therefore
uses multiple, highly sophisticated methods that prevent the user from manually
handling hard coded session IDs:
• Cookie Management: When your server uses cookies to exchange
session information, SilkPerformer will automatically handle dynamic
session ID values. Because SilkPerformer exactly emulates the cookie
management of a browser, it will send cookies the same way a browser
sends them to your server. No manual interaction is needed for state
management.
• Page-based browser level API (Page-level API): Using the Page-level
API when recording (which is the default setting) will generate scripts
that mainly generate context-full Web API function calls such as
150 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Customizing Session Handling
WebPageLink and WebPageSubmit. Context-full Web API calls are
working on the HTML level, not the HTTP level, and thus they do not
use URL's as parameters. For all context-full API calls, manual session
customization is not needed. The page-level API is used if you choose
the application type Web business transaction (HTML/HTTP) in the
Outline Project dialog. Therefore it is strongly recommended that you
use SilkPerformer's page-level API instead of the low-level browserlevel
API.
Due to the heavy use of client side Java Script for dynamically generating
HTML, the SilkPerformer recorder sometimes loses the HTML context and
scripts context-less Web API calls. Context-less Web API calls, such as
WebPageUrl and WebPageForm, contain URL's as parameters. In these rare
cases, your script may contain hard coded session IDs. You will find them in the
URL parameters of Web API calls and in the form fields declared in the
DCLFORM section of your script.
Example
Context-full script (there is nothing to customize):
transaction TMain
begin
WebPageUrl("http://lab38/ShopIt/"); // first call is
always context-less
WebPageLink("Join the experience!");
WebPageSubmit("Enter", SHOPIT_MAIN_ASP001);
WebPageLink("Products");
end TMain;
dclform
SHOPIT_MAIN_ASP001:
"SessionID" := "" , // hidden value:
"LGIJALLCGEBMIBIMFKOEJIMM"
// recognized as a hidden form
// field, the value is taken from
// the actual HTML form field.
"name" := "Jack", // changed
"New-Name-Button" := "" ; //unchanged
value: "Enter"
Scripts with context-less functions (static session data that needs to be
customized is included in the DCLFORM section):
transaction TMain
begin
WebPageUrl("http://lab38/ShopIt/"); // first call is
always context-less
SilkPerformer User Guide 151

6 CUSTOMIZING TEST SCRIPTS
Customizing Session Handling
WebPageUrl("http://lab38/ShopIt/main.asp", NULL,
SHOPIT_MAIN_ASP001);
WebPageForm("http://lab38/ShopIt/main.asp", SHOPIT_
MAIN_ASP002);
WebPageUrl("http://lab38/ShopIt/products.asp");
end TMain;
dclform
SHOPIT_MAIN_ASP001:
"from" := "welcome";
Customizing session
handling
Identifying Session
IDs
SHOPIT_MAIN_ASP002:
"SessionID" := "LGIJALLCGEBMIBIMFKOEJIMM",
"name"
:= "Jack",
"New-Name-Button" := "Enter";
The following basic steps can be used to customize session handling. Each of
these steps is detailed in the following sections:
1 Identify the session ID that needs to be customized.
2 Search for the first response (Web API call in the script) in which the
session ID is sent from the server to the client.
3 Parse out the session ID from the found response of the Web API call
into a variable.
4 Replace all occurrences of the hard coded session ID in the script with
this variable.
Without the support of SilkPerformer's TrueLog Explorer functionality, even
the first step of this process, that is, identifying the session relevant data, could
be challenging.
As previously mentioned, session IDs are dynamic values generated by Web
application to maintain the state of a Web application. Because of their dynamic
nature, session IDs differ between each user session. Thus they can also differ
between the recorded session and the replayed session. By comparing record
and replay logs for differences, you can find these dynamic values.
Finding session IDs can be easily done using SilkPerformer's TrueLog Explorer.
Choosing Customize Session Handling from the workflow bar will
automatically open the corresponding record TrueLog file of the opened replay
TrueLog. By going through your record and replay log step by step, you can
look for differences in the HTML responses between the record and replay
scripts. The HTML Source Differences tab shows you an overview of all the
differences between two corresponding HTML documents.
The differences found between the recording session and the replay session
scripts are the first source to consider for locating possible session IDs. Not all
of the differences that you find in the returned data of your recording session and
152 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Inserting Parsing Functions
the returned data of your replay session are relevant for session customization.
Only if you find differences between the record and replay scripts, will you then
have found a possible candidate for session customization.
More precisely:
If the differing phrase in the server response data of the recording session can
also be found in the URLs of Web API calls or in the form fields inside the
DCLFORM section of your script, you have identified a possible session ID that
can be customized.
Parsing and replacing
session IDs
Now that you have identified the session ID and the Web API call in the BDL
script which returns the session ID, you must parse out the session ID from the
response to the Web API call. SilkPerformer offers numerous parsing functions
that enable you to parse out values from server responses and store the result in a
variable in your script (for details see WebParse* functions in the BDL Function
Reference). The next step is to replace all occurrences of the hard coded session
ID in the script with this variable.
SilkPerformer 's TrueLog Explorer also supports this task so that no manual
interaction with your script is needed. TrueLog Explorer will generate the
appropriate parsing and replacement statements automatically without the need
to manually change a single line of code in your BDL script. It automatically
recognizes relevant session data and shows it in the view HTML Source
Difference. There, you simply can select the line that contains the session ID and
customize it by double clicking on the row. A dialog, in which you can check
the parameters of the parsing function, will be shown. Accepting the dialog will
generate the parsing and replacement statements that are needed to customize
the session handling for the selected session ID.
Inserting Parsing Functions
Similar to the verification functions, parsing functions can be used to parse
content returned from a Web server and check if the values meet your testing
criteria. Different from the verification functions, which basically check the
occurrence of a specified input value, the parsing functions parse the content of
a server response and return the parsed value in a BDL variable. Typically, you
will use parsing functions for the following tasks:
- Parse session IDs and replace static session IDs in the script with
parsed, dynamic session IDs to maintain state information in Web
applications. This is one of the main fields of application of parsing
functions.
- Build enhanced verifications into your script with parsing functions,
which cannot be done with verification functions. For example, if
you want to verify that the value of column 2 of row 3 in an HTML
SilkPerformer User Guide 153

6 CUSTOMIZING TEST SCRIPTS
Inserting Parsing Functions
table is equal to the sum of the values of column 2 of row 1 and
column 2 of row 2. By using parsing functions to parse out the three
values and compare them in our script, you can build this enhanced
verification check.
- Conditional execution of parts of your testing script which depends
on the data returned from the server. For example, an HTTP request
returns an HTML page which includes: "Results: items
found". You want to execute different actions depending on the
value of . Let's say you want to exit the transaction if =
0, and you want to go to link "Details" if = 1, and you want to
go to link "Next" if is greater than 0. To accomplish this, you
need to parse the value of from the HTML page and script the
actions, depending on the value you have parsed out.
SilkPerformer offers two means to enhancing your test scripts with parsing
functionality:
- Apply parsing functions visually in TrueLog Explorer without the
need to write a single line of BDL code. TrueLog Explorer will
automatically generate parsing functions in your script.
- Directly enhance your script by coding verification functions
manually.
The following parsing functions are available with SilkPerformer:
HTML content parsing
Response data
parsing
HTML content parsing functions parse out portions of the rendered, visible
HTML content. You can apply HTML content parsing functions in the HTML
view of TrueLog Explorer.
- WebParseHtmlBound
- WebParseHtmlTitle
- WebParseTable
- WebParseResponseTag
- WebParseResponseTagContent
Data parsing functions parse out portions of the response data returned from the
server. In cases in which an HTML document is returned from the server, the
complete source code of the HTML document is parsed. You can apply data
verification functions in the source view of TrueLog Explorer.
- WebParseDataBound (this function replaces the obsolete
WebParseResponseData function)
- WebParseResponseHeader
See “Matrix: Functions, the Recorder, and TrueLog Explorer” for more
details.
154 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Adding Verifications to Test Scripts
Adding Verifications to Test Scripts
Value of verification
checks
Often application errors do not cause erroneous HTTP responses. Instead, the
application responds with incorrect data values or with error messages
incorporated in the HTML content, such as “A Servlet Exception
occurred&ldots;" or "Server Too Busy&ldots;”. A simple check of the HTTP
status code will not uncover this class of errors. Therefore, application errors are
often overseen if you do not build additional checks into your test script.
Verification functions help you to check for application errors that are not
simple, standard HTTP errors. With verification built into your test scripts, your
test evolves from being a simple load test to a becoming a combined load and
functionality test. You can use these scripts without major performance
penalties, even during large load test scenarios. They will, therefore, be able to
detect a new class of application errors—errors which only occur under load,
but which would be overlooked by a simple load test script without additional
verification checks.
SilkPerformer offers three means of enhancing your test scripts with verification
functionality:
• Automatically let the Recorder generate verification functions during the
recording session.
• Apply verifications visually via TrueLog Explorer without the need to
write a single line of BDL code. TrueLog Explorer will automatically
generate verification functions into your script.
• Directly enhance the script by coding verification functions manually.
All three approaches are based on SilkPerformer's verification functions. The
following verification functions are available with SilkPerformer.
HTML content
verification
HTML content verification functions verify the rendered, visible HTML
content. You can apply HTML content verification functions in the HTML view
of TrueLog Explorer.
- WebVerifyHtml
- WebVerifyHtmlBoundEx
- WebVerifyHtmlTitle
- WebVerifyTable
See the BDL Function Reference for details.
Data verification
Data verification functions verify the complete response data returned from the
server. In cases in which an HTML document is returned from the server, it
provides the complete source code of the HTML document. You can apply data
verification functions in the Source view of TrueLog Explorer.
- WebVerifyData
SilkPerformer User Guide 155

6 CUSTOMIZING TEST SCRIPTS
Adding Verifications to Test Scripts
Automatically
generating
verifications during
recording
- WebVerifyDataBoundEx
- WebVerifyDataDigest
See the BDL Function Reference for details.
See “Matrix: Functions, the Recorder, and TrueLog Explorer” for more
details.
HTML title verifications and response data digest verifications can be
automatically generated during a recording session. To enable automatically
generated verifications during recording, go to the Menu Settings/Active Profile/
Record/Web/Verification/Recording and check the Record title verification and
the Record digest verification checkboxes.
Title Verification
Title verification is especially useful if you want a simple and automatic means
of checking if your application returns the correct Web pages. It is not useful if
the title (the content of the HTML tag ) of all of your HTML pages is
not set or they all have the same title.
The recorder can automatically generate a title verification function for each
Web Page-level API call or each browser-level API call when retrieving an
HTML document.
Example script generated by SilkPerformer's Recorder with Record title
verification enabled:
transaction TMain
begin
WebVerifyHtmlTitle("ShopIt - Greetings");
WebPageUrl("http://bi_dell/shopit/", ...);
WebVerifyHtmlTitle("ShopIt - Confirm your name");
WebPageLink("Join the experience!", ...);
end TMain;
Title Verification can also be generated visually for selected Web pages with
TrueLog Explorer.
Digest Verification
Digest verification is useful if you want to verify the complete content of a
rather static HTML page. Digest verification is not useful if your HTML page is
dynamic or contains session information in links, embedded objects, and hidden
form fields and therefore changes with each call.
The function call WebVerifyDataDigest checks if the data that SilkPerformer
receives during replay differs from the data that the Recorder captured during an
earlier recording session. As does the Recorder, SilkPerformer generates a
digest containing information about the occurrence of each character and
compares the results with the digest generated by the Recorder. Using this
function, it is possible to check if SilkPerformer receives the same data as the
client application did during recording. The recorder automatically generates a
digest verification function for each Web Page-level API call or each browser-
156 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Adding Verifications to Test Scripts
level API call retrieving a document with a content type specified for digest
verification (by default: text/html). To enable digest verifications during replay,
go to Settings/Active Profile/Replay/Web/Verification/Data and select check the
Digest verification option.
For details about digest verification, see WebVerifyDataDigest in the BDL
Function Reference.
Example script generated by SilkPerformer's Recorder with Record digest
verification enabled:
const
gsVerDigest_ShopIt_Greetings := "\h014200000B&ldots;"
"\h160005000A&ldots;"
"\h030001000A&ldots;"
"\h0900410020&ldots;";
transaction TMain
begin
...
WebVerifyDataDigest(gsVerDigest_ShopIt_Greetings,
162);
WebPageUrl("http://bi_dell/shopit/", "ShopIt -
Greetings");
...
end TMain;
Visually generating
verifications with
TrueLog Explorer
With SilkPerformer, you can apply verifications visually in TrueLog Explorer
without the need to write a single line of BDL code. You can apply verification
checks in the Rendered HTML view and the HTML Source view of TrueLog
Explorer. Just mark the text you want to verify, and choose the verification
method you want to use. You can choose the verification method from four
different places:
• Script-Menu,
• Add Verifications dialog,
• Context menu of the Rendered HTML view and the HTML Source view,
or
• Tool bar.
TrueLog Explorer automatically generates the appropriate verification function
into your script.
HTML verification
HTML verifications check rendered, visible HTML content. They verify the
textual content that you see in your browser. Verifications that you apply in the
Rendered HTML view will generate the following HTML verification
functions:
SilkPerformer User Guide 157

6 CUSTOMIZING TEST SCRIPTS
Adding Verifications to Test Scripts
• WebVerifyHtmlTitle: Checks the title (the content of the HTML tag
) of the selected HTML page (no selection of text is needed to
apply a title verification). This function is generated when you select
Verify Page Title from the menu or when you select Verify the page title
from the Add Verifications dialog.
• WebVerifyHtml: Checks the text you have marked in the Rendered
HTML view of the selected HTML document. TrueLog Explorer will
automatically detect the number of occurrences of the marked text in the
selected HTML document. Selecting Verify Selected Text from the menu
or the Add Verifications dialog will open up the Insert Content
Verification dialog that generates the function.
• WebVerifyHtmlBoundEx: Checks the text you have marked in the
Rendered HTML view. TrueLog Explorer will automatically detect
unique boundaries that identify the position of the marked text. Selecting
Verify Selected Text from the menu or the Add Verifications dialog will
open up the Insert Content Verification dialog. The function is generated
when checking the Require boundary strings checkbox.
• WebVerifyTable: Checks the text you have marked within a cell of an
HTML table in Rendered HTML view. TrueLog Explorer automatically
identifies the HTML table, as well as the row and the column of the cell
you have marked. Selecting Verify Selected Text in an HTML Table from
the menu or the Add Verifications dialog will open the Insert Table
Verification dialog that generates the function.
Note If the HTML document is contained within a set of frames, the
text will only be verified against the frame that contains the selected
text, not the whole page.
Response data
verification
Response data verification functions can be applied to the complete response
data returned from the server. Using TrueLog Explorer, you can visually apply
them by marking the text in the HTML Source view. Depending on the selection
you choose, TrueLog Explorer generates the following data verification
functions:
• WebVerifyData: Checks the text you have marked in the HTML Source
view of the selected HTML document. TrueLog Explorer will
automatically detect the number of occurrences of the marked text in the
HTML source. Selecting Verify Selected Text from the menu or the Add
Verifications dialog will open the Insert Content Verification dialog that
generates the function.
• WebVerifyDataBoundEx: Checks the text you have marked in the HTML
Source view. TrueLog Explorer will automatically detect unique
boundaries that identify the position of the marked text. Selecting Verify
158 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Manual Script Editing
Selected Text from the menu or the Add Verifications dialog will open up
the Insert Content Verification dialog. This function is generated by
checking the Require boundary strings checkbox.
Enabling verification
checks during replay
Verification functions can be enabled/disabled in the profile settings dialog
Profile/Replay/Web/Verification:
- HTML/Title verification: Enables/disables the function
WebVerifyHtmlTitle.
- HTML/HTML verification: Enables/disables the functions
WebVerifyHtml and WebVerifyHtmlBoundEx.
- HTML/Table verification: Enables/disables the function
WebVerifyTable.
- HTML/Link checking: Enables/disables link checks for all Web
Page-level API calls. You can select this option to check the validity
of links in HTML documents received during a load test. For each
link included in an HTML document, additional requests are sent to
check that the link is valid.
- Data/Data verification: enables/disables the functions
WebVerifyData and WebVerifyDataBoundEx.
- Data/Digest verification: enables/disables the function
WebVerifyDataDigest.
You can overrule the profile settings in your script with the BDL function
WebSetOption.
Specifying verification
and parsing functions
All parsing and verification functions must be specified before the Web API
calls that will offer the response data that is to be parsed/verified. You can
specify multiple parse/verification functions before a Web API call. The order
of the parse/verification functions is not relevant (Exception:
WebParseDataBound and WebVerifyDataBound using the flag WEB_FLAG_
SYNCHRON).
Example BDL:
WebVerifyHtml("Proper equipment leads to a successful
trip", 1, ...);
WebPageLink("ShopIt");
Manual Script Editing
An effective means of customizing the behavior of virtual users is to manually
edit the test scripts. SilkPerformer scripts are written in the program's
proprietary scripting language, the Benchmark Description Language (BDL).
SilkPerformer User Guide 159

6 CUSTOMIZING TEST SCRIPTS
Using the System Configuration Manager
To edit scripts manually you need to be familiar with BDL, which is a high-level
language resembling Pascal.
You must be familiar with how to create prototypes of user requests, and how to
define the typical components of a load testing script, including modules,
functions, workload definitions, transactions, Web forms, and more. Extensive
help on those topics is available in the BDL Function Reference.
Using the System Configuration Manager
The System Configuration Manager enables a wide range of options to be set for
any agent computer from the controller computer. The agent software and the
launcher service can be configured, started, stopped, locked, and unlocked from
here.
For load testing, some of the most important functions of the System
Configuration Manager are those that allow agents to be easily optimized for
large load tests. The number of current TCP/IP connections that can be
established from the computer to the server being tested can be increased; this is
affected by a simple slider control. Also, if the application under test uses a load
balancer, individual IP addresses can be allotted to the virtual users.
The System Configuration Manager also enables information about the
hardware, operating system, and memory being used by a selected agent to be
viewed; reports can be written, settings can be changed, and the agent computer
can be started to make the changes effective.
Serializing and Synchronizing Virtual Users
There are several SilkPerformer functions that allow you to either serialize or
synchronize transactions in your test scripts in to better control the timing of
loads in your simulation.
Both serialization and synchronization are implemented using high-level
functions that call pre-existing Windows NT functions.
Serialization enables you to test the effects of transactions executed one at a
time. This is useful, for instance in connecting to a server that is not able to
handle a large number of concurrent logins. Serialization uses a function called
CreateMutex, which creates a token (semaphore) that only one user can possess
at a time. When the user completes the transaction, the token is released and
made available to another user.
Synchronization is useful in testing the effects of a large number of
simultaneously called transactions. It allows you to create stress on your system
to test specific concurrency issues. Synchronization works by employing an
160 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Serializing and Synchronizing Virtual Users
Function overview
additional user called an Event Starter, which functions like the starting gun at
the beginning of a race. Using the CreateEvent function, you can use the Event
Starter user to send an Event object after a specified period of time. You can set
up the users in the simulation to wait for this event and thus perform a specified
transaction simultaneously when the event is released.
The following functions are used in serializing and synchronizing transactions
and users in SilkPerformer scripts.
Function
Description
CloseHandle
Closes an open object handle
previously opened with CreateEvent,
CreateEventEx, CreateMutex, or
CreateMutexEx.
CreateEvent
Uses the Win32 CreateEventEx
function to create a named event
object.
CreateEventEx
Creates a named or unnamed event
object.
CreateMutex
Uses the Win32 CreateMutexEx
function to create a named mutex
object. Mutex objects are used to
serialize access to critical sections
inside transactions for multiple
concurrent users.
PulseEvent
Provides a single operation that sets
the state of a specified event object
to signaled and then resets it to
nonsignaled after releasing the
appropriate number of waiting
threads.
ReleaseMutex
Releases ownership of the specified
mutex object.
WaitFor
Waits for a number of users to call
this function.
WaitForSingleObject
Returns when a specified object is in
the signaled state or when the
timeout interval elapses.
Table 3: Functions for serializing and synchronizing virtual users
SilkPerformer User Guide 161

6 CUSTOMIZING TEST SCRIPTS
Serializing and Synchronizing Virtual Users
Note Refer to the SilkPerformer Benchmark Description Language
Reference manual and online Help for complete details and
examples of all SilkPerformer functions.
Serialization
Serialization functions
Serializing users and transactions requires the use of a token or mutex object.
Named mutex objects are created by calling the CreateMutex function. The
users to be synchronized have to wait until they receive the mutex object. This
guarantees that only one user executes a transaction at a time.
After a user has completed the transaction, the user must release the token by
calling the ReleaseMutex function. This makes the mutex object available for
other users.
The following functions are used in serializing transactions and users:
- CreateMutex
- CreateMutexEx
- ReleaseMutex
- WaitForSingleObject
For details on these functions, refer to the BDL Function Reference
Serialization sample
script
The following example is excerpted from a sample script called Mutexlogin.bdf,
located in the Samples\Database folder in the SilkPerformer installation
directory. The script consists of three transactions. TMain waits until the user
possesses the token, and then connects to the DBMS; Selling executes the
database transaction; and CleanUp closes the connection to the database and
releases all the resources.
Initially, a mutex object called MyConnectMutex is created to serialize access to
critical sections inside transactions for multiple concurrent users. Calling the
WaitForSingleObject function effects a delay until the mutex object is in the
signaled state. This means that from that point forward, only the current
transaction is executed, and this condition remains until the token is released.
During execution time a connection to a database system is established. After
successfully connecting, the ownership of the mutex object is released. After
these steps have been completed successfully, the transaction is committed. In
this example, only connecting to the database system is serialized.
benchmark MutexLogin
use "dbapi.bdh"
var
gHdbc1 : number; // handle for database connection
gHMutex : number; // mutex handle
dcluser
user
Seller
transactions
TMain : begin;
Selling : 50;
162 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Serializing and Synchronizing Virtual Users
CleanUp : end;
dcltrans
transaction TMain
var
str: string;
begin
// SYNC: serialize connect transaction with mutex object
// (enter critical section)
gHMutex := CreateMutex("MyConnectMutex");
WaitForSingleObject(gHMutex, INFINITE);
gHdbc1 := DB_Connect("dsn=sqs_purple;uid=u1;pwd=u1");
Print("Connected!", 1, 1);
// SYNC: leave critical section and pass control to
other user
ReleaseMutex(gHMutex);
end TMain;
// process database transaction
transaction Selling
begin
...
end Selling;
transaction CleanUp
begin
DB_Disconnect(gHdbc1);
end CleanUp;
Synchronization
Synchronization
functions
Synchronizing users and transactions requires at least one event object and an
additional user to start the events. The event starter user has to be defined in the
user declaration section of the script. This user functions to control the sequence
of events by setting the state of corresponding event objectS to signaled and by
resetting them to nonsignaled after releasing the appropriate number of waiting
users. SilkPerformer provides the PulseEvent function for this purpose.
The other users call the WaitForSingleObject function to wait until the specified
event is in the signaled state. Using the WaitFor function, users are able to
synchronize themselves without the use of event objects.
The following functions are used in synchronizing users and transactions:
- CreateEvent
- PulseEvent
- WaitFor
- WaitForSingleObject
For details on these functions, refer to BDL Function Reference.
Synchronization
sample script
In the following example, the users create an event object called StarterEvent
and establish a connection to an FTP server. The Starter user sets the state of the
event object StarterEvent to signaled every ten seconds and resets it after the
SilkPerformer User Guide 163

6 CUSTOMIZING TEST SCRIPTS
Serializing and Synchronizing Virtual Users
appropriate number of waiting users process the transaction. In this way, all the
Uploader users are forced to execute the TUpload transaction at the same time.
The Uploader's wait until the event object is in the signaled state, and then
execute their Upload transaction.
dcluser
user
Starter
transactions
TInit : begin;
TStartEvent : 20;
user
Uploader
transactions
TInit : begin;
TUpload : 20;
var
hEventGo : number; // handle for starting event
hFTP : number;
dcltrans
transaction TInit
begin
hEventGo := CreateEvent("StarterEvent");
WebFtpConnect(hFTP, "standardhost", WEB_PORT_FTP,
"username", "password");
end TInit;
transaction TStartEvent
begin
wait 10.0;
// every 10 seconds the starter event forces all
// seller users to execute the selling transaction
// at the same time
PulseEvent(hEventGo);
end TStartEvent;
transaction TUpload
begin
WaitForSingleObject(hEventGo, INFINITE);
WebFtpPut(hFTP, 4096, "test.dat", true);
end TUpload;
Rendezvous
functionality
SilkPerformer includes the following rendezvous function:
- WaitFor
This function is used to create a delay until the specified virtual users reach a
certain state, and then to proceed with all the users simultaneously. This tests the
ability of the system to handle a number of concurrent transactions by making it
possible to generate very specific stress on the system. For example, in a
transaction that stores a number of files on an FTP server, you may want your
164 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Serializing and Synchronizing Virtual Users
script to wait until all the users are logged on, and then to store all the files at the
same time.
Script example
Multiple agents
const
USERS := 5;
dcluser
user
Buyer
transactions
TBuy : 10;
dcltrans
transaction TBuy
const
TIMEOUT := 60;
begin
// wait for users rendezvousing at this point
if not WaitFor("Rendezvous", USERS, TIMEOUT) then
write("timeout"); halt;
end;
// proceed with transaction
print("rendezvous", 1, TEXT_RED);
end TBuy;
In this example, the script waits until five virtual users (Buyer) are at the line in
the script containing the WaitFor function, and it then proceeds with the
transaction.
The functionality described above is available only on a single agent computer.
SilkPerformer provides an additional number of functions that extend the use of
serialization and synchronization across multiple agents.
Serialization across multiple agents is made possible by the
GlobalResourceEnter and GlobalResourceRelease functions.
GlobalResourceEnter defines a resource that can be occupied by only a limited
number of users. The caller has to wait until the number of users who occupy the
resource is less than the maximum number of users who are allowed to occupy
the resource. The GlobalResourceRelease function releases a resource that was
occupied with the GlobalResourceEnter function. If there are users still waiting
for the resource, the next user in the queue resumes the simulation.
Synchronization across multiple agents is made possible by the GlobalWaitFor
function. This function defines a checkpoint (rendezvous point) and blocks the
calling user. The caller waits until a specified number of users (including the
caller) have reached or passed the checkpoint, or until a specified timeout
occurs. The checkpoint is identified by name and is visible to all the users,
including those running on various remote agents. A waiting user continues
when the specified timeout is reached or the given number of users have called
the function for a checkpoint. A user who has timed out is also regarded as
having passed the checkpoint.
SilkPerformer User Guide 165

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
The functions GlobalVarGet, GlobalVarInc, and GlobalVarSet are used,
respectively, to get, increment, and set the values of global integer variables that
are accessible for all the users on all the agents.
Using Custom DLLs in Test Scripts
In SilkPerformer scripts, it’s possible to call functions that are provided in
Dynamic Link Libraries (DLLs). This feature allows you to implement
functions in any programming language that can generate DLLs and then utilize
such external functions in SilkPerformer load test scripts.
Declaring and calling
external functions
Two files are typically required to create a DLL:
• A source file that contains the implementation of one or more functions
• A definition file that specifies which functions are to be exported
When developing a DLL using a common C/C++ compiler, the source file (.c,
.cpp) and the definition file (.def) must be included in the project.
In BDL, all external functions called from a script must be declared before they
can be used. This must be done in the external functions section of
SilkPerformer scripts.
Note When a DLL file isn’t located in the SilkPerformer installation
directory or a directory included in the path specification, it’s
necessary to add the DLL file to the project’s data files.
Please see the External functions sections of the BDL Function Reference for
more information and an exact syntax description.
Function prototypes
In the external functions section of scripts, after the name of a DLL has been
declared, any number of prototypes of DLL functions can be defined. Each
function that is to be used in a SilkPerformer script must be declared; these
declarations are case-sensitive. Additionally, identifiers that specify the names
of all functions that are to be used in BDL scripts must be defined.
When parameters are passed to external functions, it’s necessary to indicate
whether they are passed into functions by reference or by value (default).
Parameters can be declared with BDL data types or basic C data types (long,
short, char, float, double, or pointer).
The following table shows the data-type mapping between formal parameters of
native functions exported by DLLs, formal parameters of external functions
166 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
declared in the DLL sections of BDH files, and actual parameters (BDL data
types).
C Data Type DLL Section (BDH) BDL Data Type
char char Number
unsigned char unsigned char Number
short short Number
unsigned short unsigned short Number
long long, number, boolean Number, Boolean
unsigned long long, number Number
char* string String
*(Pointer) long, number Number
float float Float
double double Float
Ptr (C data type
representing a pointer to
any data type)
string
String
Table 4: Data-type mapping
Script example
In the following example, the external functions DemoSum and DemoDiff are
declared and then called in the TMain transaction.
dll "DemoFunctions.dll"
// function without parameters and return value
"DemoOutput"
function Output;
// function declared using C data types
"DemoSum"
function Sum(in long, in long): long;
// function declared using BDL data types
"DemoDiff"
function Diff(in number, in number): number;
// DLL is not located in the SilkPerformer directory
dll "C:\\Debug\\Functions.dll"
dcltrans
transaction TMain
SilkPerformer User Guide 167

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
var
x, y, nSum, nDiff: number;
begin
x := 4; y := 5;
nSum := Sum(x, y); // external function calls
nDiff := Diff(x, y);
Output();
end TransMain;
Passing parameters
Simple types
Any type of parameter can be passed to external functions. While using simple
data types is easy, handling arrays and structured data types requires knowledge
about the alignment and size of the specific data types in use.
Simple data types allowed as input and output parameters are
C Data Type
BDL Data Type
Char, unsigned char Number
Short, unsigned short Number
Long
Number
Double
Float
Float
Float
Char*
String
Table 5: Simple data types
C source code
Note The unsigned long data type is converted to long when used in
SilkPerformer scripts.
When a parameter is passed to a function by value, it’s declared simply by
specifying its type. If a parameter is passed by reference (for example, if it’s
used as an output parameter), a pointer to the data type must be defined.
The example below illustrates how function parameters are declared in C and
how corresponding external functions must be declared in BDL. In addition, an
example of an external function call is provided.
In the example below, the external function T_PARAM receives the following
input parameters: inChar, inShort, inLong, inDouble, inFloat, inUShort, and
inUChar. A corresponding output parameter belongs to each: outChar, outShort,
outLong, outDouble, outFloat, outUShort, and outUChar. These parameters
must be declared as pointers because during function calls references to the
variables are passed to the functions. Additionally, a string called inoutString is
declared as an input and output parameter.
168 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
BDL external function
declarations
BDL external function
calls
char* T_PARAM( char inChar, // in
short
inShort,
long
inLong,
double
inDouble,
float
inFloat,
unsigned short inUShort,
unsigned char inUChar,
char*
inoutString, // in/out
char* outChar, // out
short*
outShort,
long*
outLong,
double*
outDouble,
float*
outFloat,
unsigned short* outUShort,
unsigned char* outUChar )
{
char sBuffer[1000];
// copy the values of the input parameters to the
// output parameters and the return value
return sBuffer;
}
To use the above-defined T_PARAM function in BDL, the function must be
declared in the external functions section of a SilkPerformer script. This is
illustrated in the below example. In this case, the parameters are defined as
formal data types. The comments attached to each parameter indicate the
corresponding BDL data type.
"T_PARAM" function t_param(
in char,
// in number
in short,
// in number
in long,
// in number
in double,
// in float
in float,
// in float
in unsigned short,
// in number
in unsigned char,
// in number
inout string,
// inout string
out char,
// out number
out short,
// out number
out long,
// out number
out double,
// out float
out float,
// out float
out unsigned short,
// out number
out unsigned char): string(100); // out number
The following example illustrates how to call the T_PARAM function. In this
example, random values are assigned to the variables and passed to the function
as input parameters.
dcltrans
transaction TMain
SilkPerformer User Guide 169

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
var
n1, n2, n3, n4, n5, n6, n7, n8, n9, n10 : number;
f1, f2, f3, f4 : float;
s1, sRet : string;
begin
n1 := 127; n2 := 32000; n3 := 2000000;
n4 := 64000; n5 := 255;
f1 := 12345.12345; f2 := 12.99;
s1 := "Teststring";
sRet := t_param(n1, n2, n3, f1, f2, n4, n5, s1,
n6, n7, n8, f3, f4, n9, n10);
end TMain;
Arrays
C source code
examples
Fixed boundaries
Variable upper bounds
An array is an ordered list of variables of the same type. Arrays of the Boolean,
number, float, and string data types are available in BDL. Passing array
parameters to external functions is more complex than handling simple data
types.
It’s possible to pass arrays with either fixed boundaries or variable upper
bounds. When an upper bound is variable, it must be specified by an additional
parameter.
Note Lower bounds are always zero (0) in C and one (1) in BDL.
The two examples below show a function that manipulates the contents of a
string and a number array.
In the first example, an array with fixed boundaries (1 - 20 in BDL) must be
passed to the function.
#define ARRAYSIZE 20
void T_ARRAY( char as1[ARRAYSIZE][101],
long an1[ARRAYSIZE],
char as2[ARRAYSIZE][101],
long an2[ARRAYSIZE] )
{
// copy the values of the input arrays to
// the output arrays for demonstration purpose
for (i=0; i

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
void T_ARRAY( char as1[][101],
long an1[],
char as2[][101],
long an2[],
long array_upper_bound )
{
// copy the values of the input arrays to
// the output arrays for demonstration purpose
for (i=0; i

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
Structured Data Types
When using structured data types, it’s important to pay attention to the
alignment and size of structure elements. The table below shows the size of all
conventional data types that can be used within function declarations in
SilkPerformer scripts.
Data Type
Size
Char, unsigned char 1
Short, unsigned short 2
Long 4
Double 8
Float 4
Char[x]
x
Table 6: Conventional data type sizes
C source code
For more information, refer to the Introduction to data structure functions
sections of the BDL Function Reference.
The example below clarifies structure composition by showing a data structure
that contains all the simple data types.
// declare a demo data structure
#pragma pack(1) // alignment to 1-byte boundaries
typedef struct {
char myChar; // size: 1 byte pos: 1
short myShort; // size: 2 bytes pos: 2
long myLong; // size: 4 bytes pos: 4
double myDouble; // size: 8 bytes pos: 8
float myFloat; // size: 4 bytes pos: 16
unsigned short myUShort; // size: 2 bytes pos: 20
char myString[10]; // size: 10 bytes pos: 22
unsigned char myUChar; // size: 1 byte pos: 32
} TESTSTRUCT;
#pragma pack()
char* T_STRUCT(TESTSTRUCT * s)
{
char sBuf[10000];
// manipulate values of data structure and generate a
// return value for demonstration purpose
...
return sBuf;
}
172 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Using Custom DLLs in Test Scripts
BDL external function
declarations
BDL external function
calls
To handle structured data types, BDL uses strings of appropriate length and
number of type-dependent Get and Set functions. For each structured data type
passed to a function, an inout string parameter must be defined. These strings
must be exactly as long as the data structures defined in the DLLs.
The first position in a BDL string that can hold a structured data type is one (1).
When variables are aligned to one-byte boundaries, the starting position of the
next element is derived by adding the current position to the size of the current
element. For the data structure declared in the example above, a 32-byte string is
required.
"T_STRUCT"
function t_struct(inout string(32)): string(1000);
In the following example, arbitrary values are assigned to all the elements of the
data structure using the corresponding Set functions. Next, the external function
t_struct is called to manipulate the data structure.
dcltrans
transaction TMain
var
n1, n2, n3, n4, n5 : number;
n6, n7, n8, n9, n10 : number;
f1, f2, f3, f4 : float;
sStruct
: string(100);
sRet
: string(1000);
begin
/* one-byte alignment */
SetChar (sStruct, 1, 127);
SetShort (sStruct, 2, 32000);
SetLong (sStruct, 4, 2000000);
SetDouble(sStruct, 8, 12345.12345);
SetFloat (sStruct, 16, 12.99);
SetUShort(sStruct, 20, 64000);
SetString(sStruct, 22, "123456789", 10);
SetUChar (sStruct, 32, 255);
sRet := t_struct(sStruct);
end TMain;
SilkPerformer User Guide 173

6 CUSTOMIZING TEST SCRIPTS
Matrix: Functions, the Recorder, and TrueLog Explorer
Matrix: Functions, the Recorder, and TrueLog Explorer
HTML content
Response data
Verification/
Parsing Usage Description
WebVerifyHtml WebVerifyData Verification TrueLog Explorer's
Visual verification,
scripting
WebVerifyHtmlBound WebVerifyDataBound Verification TrueLog Explorer's
Visual verification,
scripting
WebVerifyHtmlTitle - Verification TrueLog Explorer's
Visual verification,
automatically
generated by the
Recorder, scripting
WebVerifyTable - Verification TrueLog Explorer's
Visual verification,
scripting
- WebVerifyDataDigest Verification Automatically
generated by the
Recorder
WebParseHtmlBound WebParseDataBound Parsing TrueLog Explorer's
Visual parsing,
scripting
Checks if a
specified string
occurs in the
HTML content/
response data.
Checks if a
specified string
occurs in the
HTML content/
response data
within given
boundaries.
Checks if a
specified string
occurs in the title of
the HTML
document.
Checks if a
specified string
occurs in an HTML
table.
Checks if the
response data
differs from the
response data that
the Recorder
captured during an
earlier recording
session.
Parses the HTML
content/response
data within given
boundaries.
WebParseHtmlTitle - Parsing Scripting Parses the title of an
HTML document.
Table 7: Functions, the Recorder, and TrueLog Explorer
174 SilkPerformer User Guide

6 CUSTOMIZING TEST SCRIPTS
Matrix: Functions, the Recorder, and TrueLog Explorer
HTML content
Response data
Verification/
Parsing Usage Description
WebParseTable - Parsing Scripting Parses the content
of a cell in an
HTML table.
- WebParseResponseTag Parsing Scripting Parses the value of
the specified
HTML attribute.
Also applicable for
XML response
data.
- WebParseResponseTagContent Parsing Scripting Parses the content
of the specified
HTML attribute.
Also applicable for
XML response
data.
- WebParseResponseheader Parsing Scripting Parses HTTP
response headers.
- WebParseResponseRedirect Parsing Scripting Parses the URL of a
redirection
response for URL
encoded
parameters.
Table 7: Functions, the Recorder, and TrueLog Explorer
HTML Content
Response Data
HTML content verification functions and HTML content parsing functions can
be applied to the rendered, visible HTML content. You can apply HTML
content verification and parsing functions in the HTML view of TrueLog
Explorer. The HTML content verification and parsing functions allow you to
verify and/or parse the textual content that you see in your browser.
Data verification and data parsing functions can be applied to all of the response
data returned from the server. In cases in which an HTML document is returned
from the server, this includes the complete source code of the HTML document.
You can apply data verification and parsing functions in the source view of
TrueLog Explorer.
SilkPerformer User Guide 175

6 CUSTOMIZING TEST SCRIPTS
Matrix: Functions, the Recorder, and TrueLog Explorer
176 SilkPerformer User Guide

7
Establishing 7Chapter
Baseline
Performance
Introduction
What you will learn
This chapter explains how to establish baseline performance with
SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 177
Identifying Baseline Performance 179
Confirming Test Baselines 179
Setting Response Time Thresholds 182
Setting Up Server Monitoring 183
Overview
The fifth step in conducting a SilkPerformer load test is to ascertain the baseline
performance (i.e., to determine the ideal performance of the application under
test). The now customized test is run with just one user per user type, and the
results from this unstressed performance of the application form the basis for
calculating the number of concurrent users per user type and for setting the
appropriate boundaries for the HTML page response times and transaction
response times. Additionally, the required bandwidth for running the load test
will be calculated from the baseline results. The measurements typical of a real
load test are used, and report and output files are generated.
If you have set the option Random thinking time for your load test, it will be
ignored for the baseline test. The think times do not influence the transaction
SilkPerformer User Guide 177

7 ESTABLISHING BASELINE PERFORMANCE
Overview
response times and you receive accurate results when comparing different
baseline tests.
A secondary function of the baseline-finding test is to serve as a trial run of the
customized test that was created in the fifth step. Here the objective is to ensure
that the customization has not introduced new errors into the script, and that the
script will accurately and fully reproduce the interaction that is intended
between the client application and the server.
For baseline runs the following options are automatically set to these specified
value:
- A Baseline report file will automatically be created.
- The Stop virtual users after simulation time (Queuing Workload)
option is on.
- The Random thinking time option is set off.
- The Loadtest description field is set to "BaseLine Test".
- The Display All Errors Of All Users option in the Monitor window is
on.
- The Virtual user output files (.wrt) option is on.
- The Virtual user report files (.rpt) option is on.
Specifying user types
Output files
By adding different profiles to your load test project, you can define different
user types, which provide a unique combination from which to create a load
testing script with a user group contained in the script and a load testing profile.
With this combination, it is very easy to assign different behaviors to one user
group contained in the script.
Before you can define user types, you must add the profiles to your load test
project. If you want to emulate 3 different modem speeds during your load test,
you must add a profile for each modem speed first.
The Find Baseline dialog allows you to define many user types (i.e., a
combination of a script, a user group, and a profile).
You can run a baseline test to establish the baseline performance of your load
test for the selected user types. The different user types can be run at the same
time. For the baseline test, only one virtual user per user type is executed; the
think times contained in the script will not be randomized for comparison
reasons.
Output files contain the output of write statements used in test scripts. An output
file is generated for a particular user only if write statements are executed by that
user.
However, generating output files alters the time measurements of a load test.
Therefore, these files should be used for debugging purposes only, and should
not be generated for full load tests.
178 SilkPerformer User Guide

7 ESTABLISHING BASELINE PERFORMANCE
Confirming Test Baselines
Identifying Baseline Performance
Finding the test
baseline
Procedure To find the test baseline:
1 Click the Find Baseline button on the SilkPerformer Workflow bar.
The Workflow - Find Baseline dialog appears.
2 Add different user types (unique combinations of a script, a user group,
and a profile).
3 Select the user types that you want to execute simultaneously.
4 Click Run to execute your test.
5 During the baseline test, monitor the test progress in the Monitor
window.
Confirming Test Baselines
The sixth step in conducting a SilkPerformer load test is to confirm that the test
baseline established by the test in the previous step actually reflects the desired
performance of the application under test. These measurements are taken into
account in calculating the number of concurrent virtual users, setting acceptable
thresholds for the load test, and for calculating the required bandwidth.
This can be achieved by inspecting the results from the test in the baseline
report. If the results are satisfactory, they can be stored for further processing.
• “Viewing Baseline Reports”: A detailed XML/XSL based report that
provides you with a summary table, transaction response-time details,
timers for all accessed HTML pages, Web forms and errors that
occurred. This information is available for all user types involved in the
baseline test.
• “Viewing Accepted Baseline Reports”: All accepted baseline reports for
this project so far. All accepted results for every user type are stored and
can be inspected with this report.
• “Setting Response Time Thresholds”: Thresholds for response times can
be set by using the accepted results from previous baseline tests. With
these thresholds, the analysis of future load tests will be very easy,
because the boundaries are set separately for every user type for every
Web page, form, and transaction response time.
• “Setting Up Server Monitoring”: A server-monitoring template can now
be set up. The template will then be used for live monitoring during the
load test, so that server-side results information can be generated for
exploration and correlation after the test.
SilkPerformer User Guide 179

7 ESTABLISHING BASELINE PERFORMANCE
Confirming Test Baselines
Baseline Reports
Once a baseline test is complete, a baseline report can be displayed. The
baseline report can be displayed for any load test you want to use as a baseline.
If, for example, you run a load test with 100 users and you want to use the
results of this test as your baseline, treat this test as you would a baseline test.
The baseline report is based on XML/XSL and includes the most important test
results in a tabular form. It comprises the following sections:
• General information
• User Types
- Summary tables
- Transaction response times
- HTML Page Timers
- Web form measurements
• Accept Results Button
General Information
The general information section includes administrative information in a tabular
form.
Administrative information includes the SilkPerformer Version information, the
project name, a description of the project, the date and time of the baseline test,
the workload definition, the workload model, and the number of errors that
occurred.
User Types
For each user type involved in the baseline test run, a separate section is
available with details on the measured response times. In the summary line, the
number of virtual users (1 for a baseline test), the duration of the test, the session
time, the session busy time and, the average page time, the number of
transaction executed OK, cancelled, failed and the number of errors that
occurred.
The session time consists of the execution time of all transactions defined in the
dcluser section of a virtual user, without the initial and end transactions. The
session busy time is calculated as the session time minus think times.
For example:
dcluser
user
Vuser1
transactions
TInit : begin;
T1 : 1;
T2 : 3;
Tend : end;
180 SilkPerformer User Guide

7 ESTABLISHING BASELINE PERFORMANCE
Confirming Test Baselines
The session time is the average response time of T1 + 3 * average response time
of T2; the session busy time is the same without any think time.
If you are satisfied with these results, you can accept them as baseline results
and save them for further processing, such as the calculation of the number of
concurrent virtual users and the network bandwidth needed for the load test.
For each user type, the following detailed results are available:
• Summary Tables
This section contains summary measurements in a tabular form, that is,
the aggregate measurements of all virtual users. The first table provides
general information, such as the number of transactions that were
executed and the number of errors that occurred. All the following tables
provide summary information relevant to the type of application that was
tested.
• Transactions
This section contains summary measurements in a tabular form, that is,
aggregate measurements for all transactions of a specific user type. For
every transaction, the transaction response time and the transaction busy
time are displayed.
The transaction response time is measured from the beginning to the end
of the transaction, including all think times.
The transaction busy time is the transaction response time without any
think time.
• Page Timers
This section contains summary measurements in a tabular form for every
accessed Web page. For every page, the download time of the whole
page, the HTML-document download time, the server busy time, the
amount of data downloaded with this page, and the amount of data
downloaded for embedded objects are displayed.
• Web Forms
This section contains summary measurements in a tabular form for every
used Web form. For every form, the roundtrip time, the server busy time,
the number of HTTP hits, the amount of request data sent and the
response data received are displayed.
• Accept results button
If the results of the inspected load test are acceptable for use as a
baseline, you can store these results for further calculations and
processing by clicking on the Accept Results button. The results of tests
with all user types involved in the test are stored in the file
\BaselineResults\baselineReport_.brp. If
SilkPerformer User Guide 181

7 ESTABLISHING BASELINE PERFORMANCE
Setting Response Time Thresholds
you define different workloads in your load test project, every workload
must have a separately defined baseline. If you create a copy of a
workload, the accepted baseline results will also be copied.
Viewing Baseline
Reports
Viewing Accepted
Baseline Reports
Procedure To view a baseline report:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
2 In the ensuing dialog, click the Baseline Report link or button.
The Baseline Report appears, showing the results of the previous test
run.
3 Click the Accept Baseline button to store the displayed results of all user
types for further processing.
An Accepted Baseline Report contains the same information as a baseline report
for all accepted user types within the actual load test project.
The results stored in this report are used for the generation of thresholds for the
accessed Web pages and for thresholds for transaction response times and
custom timers. These thresholds are displayed in the overview report at the end
of a load test. With this information you can easily check if the response times of
a load test are acceptable.
The accepted baseline report will also be used to calculate the number of
concurrent virtual users and to show the network bandwidth needed for a test.
The accepted baseline reports are stored in the file:
\BaselineResults\baselineReport_.brp
If you define different workloads in your load test project, each workload must
have a separate baseline defined. If you create a copy of a workload, the
accepted baseline results will also be copied.
Viewing these reports can be useful for verifying the accepted results.
Procedure To view an accepted baseline report:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
2 On the Workflow - Confirm Baseline dialog, click the Accepted Baseline
Report link or button.
A report showing all accepted baseline results appears.
Setting Response Time Thresholds
With the accepted results from the previous baseline tests, thresholds for
response times can be set for subsequent load tests. These thresholds will be
used in the general overview report of a load test to separate good, acceptable
and unacceptable Web response times.
182 SilkPerformer User Guide

7 ESTABLISHING BASELINE PERFORMANCE
Setting Up Server Monitoring
Therefore, two boundaries must be specified:
- Good response times are below boundary 1.
- Unacceptable response times are above boundary 2.
- Acceptable response times are between these boundary values.
Thresholds can be set for:
- Transaction response times
- Custom timers
- HTML page timers
For the calculation of the boundaries from the baseline results, you must specify
a multiplier. The average response times of the timers will be multiplied with
these factors and the boundaries are set accordingly.
For example: A multiplier of 3 means that you set a boundary 3 times higher
than the average response time of the timer in the baseline test.
If the measured value in the baseline test was 0, you can specify a minimum
value for the boundaries.
The load testing scripts connected with the user types will be updated to set the
thresholds accordingly.
Setting response time
thresholds
Procedure To set response time threshold:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
2 On the Workflow - Confirm Baseline dialog, click the set response time
thresholds link or button.
The Automatic Threshold Generation dialog appears.
3 Select the timers you want to set thresholds for.
4 Specify the appropriate multipliers for your load tests (Lower bound and
Upper bound).
5 Specify minimum values, if the corresponding timers from the baseline
test are 0.
Setting Up Server Monitoring
One of the most important facilities provided by SilkPerformer during a load
test is server monitoring—where you can watch a live graphical display of the
performance of your server while a test is running.
Monitoring servers is also important in that it enables server-side results
information to be generated. This can then be viewed and correlated with other
test measurements at the results exploration stage.
SilkPerformer User Guide 183

7 ESTABLISHING BASELINE PERFORMANCE
Setting Up Server Monitoring
Monitoring a server helps you find out whether there are bottlenecks on the
server, and, if so, where exactly they are. In this way, you examine the
performance of both the operating system and the server application.
A template for server monitoring can be set up to monitor performance data or
you can use a default template that is already generally configured for the type
of application you are testing.
Procedure To set up a template for server monitoring:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
2 On the Workflow - Confirm Baseline dialog, click the Monitoring
template link or button.
The Profile Settings dialog opens, showing the Monitoring tab of the
Results category.
3 In the Monitoring options area, select the Automatically start monitoring
option to automatically launch the monitoring facility of Performance
Explorer when you start a load test. Performance Explorer will then
display the server performance data that is relevant to the type of server
you are testing.
- Select the Use default monitoring template option to activate the
default monitor template.
- Select the Use custom monitoring template option to use a custom
monitor template. Enter the name of the custom template file (.pew)
that you want to use to monitor your server. Click the browse (...)
button next to the field to locate the file.
4 Click the Create Custom Monitor Template button to create the template.
- To subsequently edit the template, click the Edit Custom Monitor
Template button; Performance Explorer will open and you will have
the choice of adding or removing any monitoring performance data.
Finally, save the workspace to apply your changes to the template
and close Performance Explorer.
5 Click OK.
6 Once you start a load test, server monitoring for the load test will start
and stop automatically. To save monitoring results for future exploration,
have the results of server monitoring written to a monitoring report file
(see “Monitoring server performance during tests” for details).
184 SilkPerformer User Guide

8
Defining 8Chapter
Workload
Introduction
What you will learn
This chapter explains how to define workload for SilkPerformer load tests.
This chapter contains the following sections:
Section
Page
Overview 185
Setting Up Workload 187
Adjusting Session Times 191
Maintaining Transaction Rates 191
Overview
The seventh step in conducting a SilkPerformer load test is to configure
workload. SilkPerformer offers different workload models to be used as a basis
for your load test. Before configuring workload you must select the model that
best fits your needs.
The following workload models are offered:
• Increasing workload model:
At the beginning of a load test, SilkPerformer simulates not the total
number of users defined, but only a specified part of them. Step by step,
the workload is increased until all the users specified in the user list are
actually running.
This workload model is especially useful when you want to find out at
which load level your system crashes or does not respond within
acceptable response times or error thresholds.
SilkPerformer User Guide 185

8 DEFINING WORKLOAD
Overview
• Steady state workload model:
In this model, the same number of virtual users is employed throughout
the test. Every virtual user executes the transactions defined in the load
testing script. When work is finished, the virtual user starts again with
executing the transactions. There is no delay between transactions. The
test finishes when the simulation time that was specified has been
reached.
This workload model is especially useful when you want to find out
about the behavior of your tested system at a specific load level.
• Dynamic workload model:
You can manually change the number of virtual users in the test while it
is being run. The maximum number of virtual users to be run is set;
within this limit, the number can be increased or decreased at any time
during the test. No simulation time is specified, and you must bring the
test to an end manually.
This workload model is especially useful when you want to experiment
with different load levels and to have the control over the load level
during a load test.
• All day workload model:
This workload model allows you to define the distribution of your load in
the most flexible way. You can assign different numbers of virtual users
to any interval of the load test. Each user type can use a different load
distribution. Therefore, you can design very complex workload scenarios
such as workday workloads or even weekly workloads. You can adjust
the load level during a load test for intervals that have not yet started
executing.
This workload model is especially useful when you want to model
complex, long lasting workload scenarios in the most realistic way
possible.
• Queuing workload model:
In this model, transactions are scheduled following a prescribed arrival
rate. This rate is a random value based on an average interval calculated
from the simulation time and the number of transactions you have
specified in your script (dcluser section: number of transactions per
user). The load test will finish when all of the virtual users have
completed their prescribed tasks.
Note that the test may take longer than the simulation time that was
specified, because of the randomized arrival rates. For example, if you
specify a simulation time of 3000 seconds and want to execute 100
transactions, then you will get an average transaction arrival rate of 30
seconds.
This workload model is especially useful when you want to simulate
186 SilkPerformer User Guide

8 DEFINING WORKLOAD
Setting Up Workload
workloads that use queuing mechanisms to handle multiple concurrent
requests. Typically, application servers like servlet engines or transaction
servers, which are receiving their requests from Web servers and not
from end users, can be accurately tested by using the queuing model.
• Verification workload model:
A verification test run is especially useful when combined with
SilkPerformer's extended verification functionality. This combination
can then be used for regression tests of Web-based applications. A
verification test run always runs a single user of a specific user type on a
specified agent computer.
This workload is especially useful when you want to automate the
verification of Web applications and when you want to start the
verification test from the command line interface.
Additionally the required network bandwidth per user type is displayed. This
helps you to check the network infrastructure used for the load test, if any
bottlenecks exist. The required bandwidth is calculated from the accepted
baseline results.
You can define more than one workload model in your load test project and save
them for further usage, but only one workload model can be active at a time. The
accepted baseline results are associated with a workload model, if you copy or
rename a workload model, the accepted baseline results are copied or renamed
accordingly.
Setting Up Workload
Procedure To set up the workload:
1 Click the Adjust Workload button on the SilkPerformer Workflow bar.
The Workflow - Select Workload Model dialog appears.
2 Select the workload model that you want for your load test. Choose from
either the Increasing, Steady State, Dynamic, All Day, Queuing, or
Verification option.
3 For the Increasing, Steady State and All Day workload models a
Workload Wizard is available to help you calculate the number of
concurrent virtual users and the required network bandwidth. If you have
selected on of these workload models, we recommend using the wizard
by clicking on the Workload Wizard button.
4 Enter the amount of time you want your load test to run and click Next.
SilkPerformer User Guide 187

8 DEFINING WORKLOAD
Setting Up Workload
Note Depending on the workload model you’ve selected, you can
adjust the duration of each phase of your load test on this dialog.
During the Increasing phase, workload increases in increments to
the specified maximum number of virtual users. In the Steady phase,
all specified virtual users are running. In the Decreasing phase
workload decreases in increments.
The Warm-Up phase specifies the time at the beginning of the load
test in which measurements are not factored into results calculations.
The Measurement phase, which begins at the completion of the
warm-up phase (or immediately in cases in which no warm-up time
is specified) restricts the time measurements that are taken for results
calculation.
5 The Calculate the Virtual Users dialog appears. If you know the number
of sessions per-hour per-virtual-user type, you can enter the value, and
the number of concurrent virtual users will be calculated automatically.
If you enter the number of virtual users, the number of expected sessions
per hour will be calculated. The required network bandwidth per user
type is also displayed.
Note This task is very important to achieve accurate results by
emulating real user behavior. If you know the number of expected
real user sessions per hour for your application under test, the
number of concurrent virtual users will be calculated based on the
results from the accepted baseline tests.
6 To adapt think times so as to modify session times or to guarantee
particular transaction rates, enable static or dynamic pacing of
transaction think times. When the Goal Session Time column is enabled
you can enter a desired session time for each user type. The wizard then
automatically calculates the think time factors and stores them in the
workload.
Note The think time adjustment mode can be set on a global basis
using this dialog. When think time adjustment is required on the
user-type level, the SetThinkTimeAdjustment() API function should
be applied.
7 Click Finish to close the wizard.
8 Back on the Workload Configuration screen, if you have chosen the All
Day workload model, some intervals proposed by SilkPerformer will be
displayed. You can change every value by editing the table, adding or
deleting rows, changing the number of virtual users or the duration of the
intervals.
188 SilkPerformer User Guide

8 DEFINING WORKLOAD
Setting Up Workload
Note Once a load test begins, you can change the number of virtual
users for intervals that have not yet started, but you cannot exceed
the maximum number of virtual users specified for the load test.
9 The diagram at the top of the Workload Configuration screen serves as a
graphical representation of the workload model you’ve selected. The
diagram uses the actual data of the user group selected in the workload
list at the bottom of the dialog.
10 Use the workload list to configure the user groups that are to be run in
your test scripts. In the User Type area, select the user types that you
want to run in your test. All of the user types selected prior to the
baseline test are listed here.
11 In the Max. Vusers column, specify the number of virtual users that are to
be run in each user type.
12 To specify that your load test for a specific user type will begin at a set
period after the time given in the Start Time field, enter a time value (in
seconds) in the Start Time Offset column.
13 In the Simulation Time column, enter the duration (in seconds) for which
each user type is to be run in your load test.
14 In the Warm-up Time column, enter the duration (in seconds) for the
period at the start of a load test during which no measurements are to be
performed. When you enter zero, there won't be a warm-up time and the
measuring will start immediately.
15 In the Measurement Time column, enter the duration (in seconds) for
which measurements will be performed during the load test. When a
value of zero is entered, then the measurement time is equal to the
simulation time.
16 In the Profile column, you can change the settings of the profile to be
used for each user type from the drop-down list(s).
17 In the Auto-assign column, select the assignment option for each user
type for which you want SilkPerformer to automatically compute and
assign virtual users to agent computers. To assign virtual users manually
to agent computers for a particular user type or types, leave this option
unselected.
18 Use the Agents column to check SilkPerformer's automatic assignment of
virtual users to agent computers, or to manually assign them. Click the
Browse (...) button to the left of the Agents column.
- For user types using the Auto-assign option, the Automatic Agent
Assignment for Script/Usertype dialog appears, where you can view
SilkPerformer's assignment of virtual users to agent computers.
SilkPerformer User Guide 189

8 DEFINING WORKLOAD
Setting Up Workload
- For user types not using the Auto-assign option, the Manual Agent
Assignment for Script/Usertype dialog appears.
19 In the Assigned Vusers column, enter the number of virtual users from
the relevant user type to run on each agent. All of the users from the user
type must be assigned to agents.
20 Click OK.
21 In the Start Time field in the Start time portion of the dialog, set the start
time for your load test in hours, minutes, and seconds.
- Select the Relative option to start your test after the time period you
have specified has elapsed.
- Select the Absolute option to have the test start at that specific time.
22 Check the Automatically start monitoring option, when you have
established a monitoring template and want Performance Explorer to
start and stop automatically with the load test.
23 Check the TrueLog On Error option to have SilkPerformer generate
TrueLog files for any transactions that generate errors.
24 Optionally, in the Load Test Description area, enter a description for the
load test. This feature is provided only for your project management
purposes.
25 Click the User distribution overview button to view the assignment of
virtual users to the agent computers that are currently available.
26 Click Close to close this dialog.
27 Back on the Workload Configuration dialog, click OK to save your
changes. Or, if enabled, click the Run button to execute the test. Or click
Connect to initialize the agent connection and then start the test manually
from the monitor view by clicking the Start all button.
Note To define additional workloads for your load test and save
them for future use, expand the Workloads section in the tree view,
right-click an existing workload and select New Workload or Copy
Workload from the context menu. If you create a new workload
model, the baseline will have to be accepted for the model. You can
copy previous accepted baseline results from an existing workload
model by copying and renaming an existing workload.
190 SilkPerformer User Guide

8 DEFINING WORKLOAD
Maintaining Transaction Rates
Adjusting Session Times
In some cases it’s desirable to adjust the session time of user types to a specific
duration. By adjusting a user type’s session time, throughput can be influenced
on a per-virtual-user basis to meet particular load test requirements.
Session time manipulation can be achieved by adjusting think times within a
session’s transactions. Think time adjustment is achieved on the user type level
by implicitly defining a think time factor, which is considered during think time
calculation.
SilkPerformer supports two different methods of adjusting think times: static
and dynamic.
Whereas static think time adaptation simply multiplies think times by a static
factor to adjust session time, dynamic think time adaptation factors in time slowdown
to keep throughput rates constant. See the following section for more
details on dynamic think time adaptation.
Maintaining Transaction Rates
A commonly used load test model is to keep transaction rates constant
regardless of the amount of load on the server. With dynamic think time
adaptation, you can specify workloads based on transaction rates that
SilkPerformer will meet even when server response times slow due to load
increases. This means that when server response time slows, SilkPerformer will
automatically reduce think times to avoid an increase in session times, which
would directly result in a decrease of throughput and ultimately result in server
busy times increasing as load on the server increases.
SilkPerformer offers dynamic think time adaptation for the following workload
models:
- Increasing
- Steady State
- All-Day
- Dynamic
SilkPerformer User Guide 191

8 DEFINING WORKLOAD
Maintaining Transaction Rates
192 SilkPerformer User Guide

9
Executing 9Chapter
Load Tests
Introduction
What you will learn
This chapter explains how to execute load tests.
This chapter contains the following sections:
Section
Page
Overview 193
Starting Load Tests 194
Monitoring Load Tests 194
Launching Load Tests Via the Command Line 205
Overview
The eighth step in conducting a SilkPerformer load test is to run the full load
test. Multiple virtual users are run, by means of the test script(s), to test the target
server. A large load test requires that an appropriate testing environment be set
up on the local area network, including a full complement of agent computers to
host the virtual users.
It is essential that you carefully set options for the appropriate type of test that is
to be used, accurately define the necessary workloads, and enable generation of
the kinds of test results you will need to assess the performance of the server.
The logging option, however, should be disabled to prevent interference with
load test results.
Comprehensive information is provided to testers while load tests are run. This
includes real-time information about agent computers, virtual users, and
transactions as they are conducted. In addition, real-time monitoring of the
target server is available in graphical form.
If a test is configured as a verification run, the following options will be
automatically set to the specified values:
SilkPerformer User Guide 193

9 EXECUTING LOAD TESTS
Starting Load Tests
- A Baseline report file will automatically be created.
- The Stop virtual users after simulation time (Queuing Workload)
option is off.
- The Virtual user log files (.log) option is set off.
- The Virtual user report files (.rpt) option is on.
- The Virtual user report on error files (.rpt) option is on.
- The Compute time series data (.tsd) option is off.
Starting Load Tests
Procedure To start a load test:
1 Click the Run Test button on the SilkPerformer Workflow bar.
The Workload Configuration dialog appears.
You can also access this dialog by right-clicking the project name in the
tree-view area of the main SilkPerformer window, and selecting Run
Test; or by selecting Run/Run Test from the SilkPerformer menu bar.
2 Configure the workload that you will use in your load test.
3 Click Run to start the load test, or click Connect to initialize the agent
connection and then start the test manually from the monitor view by
clicking the Start all button.
Monitoring Load Tests
Detailed real-time information is available to testers while SilkPerformer load
tests are run. Graphic displays and full textual reporting of activity on both the
client side and the server side offer intuitive monitoring of the progress of tests
as they are happening.
Directly from the workbench where the test is conducted, the tester can see
comprehensive overview information about the agent computers and virtual
users in the test. There is full control of the level of detail of the information
presented, from a global view of the progress of all the agent computers in the
test down to exhaustive detail of the transactions conducted by each virtual user.
Progress information for each agent and each user is available in multiple
categories. Run-time details for each user include customizable, color-coded
readouts on transactions, timers, functions, and errors as they happen.
In addition, real-time monitoring of the performance of the target server is
available in graphical form. Charts display the most relevant performance
information from a comprehensive collection of the Web servers, application
194 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Monitoring all agent
computers
Monitoring a specific
agent computer
Monitoring a specific
virtual user
Monitoring server
performance during
tests
servers, and database servers running on all of the operating systems most
widely used today. Multiple charts can be open at the same time, and these can
be juxtaposed to provide the most relevant comparisons and contrasts for the
tester. A tree-view editor allows elements from any data source to be combined
in the charts. Performance information from the client application—for
example, response times—can easily be placed in the same chart as performance
data from the server. This enables a direct visual comparison to be made, so that
you can see how shortcomings on the server influence client behavior.
Procedure To monitor all agent computers:
1 While a load test is running, view progress in the Monitor window.
2 View information about the progress of agent computers and user groups
in the top part of the window. Among the comprehensive number of
information options presented, you can see the status of a particular
agent, what percentage of the test is complete on the agent, and the
number of transactions executed.
Procedure To monitor a specific agent computer:
1 In the top part of the Monitor window, select the specific agent to
monitor.
Information on all of the virtual users running on that agent appears in
the bottom of the Monitor window.
2 Among the many information options presented for each user, you can
check its status, the name of the current transaction, what percentage of
its work is complete, and the number of transactions executed.
Procedure To monitor a specific virtual user:
1 In the bottom part of the Monitor window, right-click the virtual user you
want to monitor.
2 Select Show Output of Vuser.
SilkPerformer displays detailed run-time information about the selected
user in the Virtual User window, for example, the transactions and
functions the user executes, and the data the user sends to and receives
from the server.
To view the columns that are to be shown, right click the virtual user
area, and click Select Columns. Then select all the columns you want to
view.
Procedure To monitor server performance in a graph:
1 Click the Confirm Baseline button on the SilkPerformer Workflow bar.
The Workflow - Confirm Baseline dialog appears.
SilkPerformer User Guide 195

9 EXECUTING LOAD TESTS
Monitoring Load Tests
2 Click the Monitoring Template button or link to specify the settings for
receiving online performance data. In the Profile Results dialog choose
to automatically start monitoring while running a load test and then
choose either the default-monitoring template or to create a customized
monitoring template.
3 Click Create/Edit Custom Monitor Template.
Performance Explorer appears.
4 Close all monitor windows that you are not currently using.
5 Click the Monitor Server button on the Performance Explorer Workflow
bar.
The Data Source Wizard appears at the System Selection dialog.
You can also access the wizard by selecting Results/Monitor Server from
the SilkPerformer menu bar.
6 Select from predefined data sources if you are sure which data sources
the server provides, otherwise let Performance Explorer scan the server
for available data sources.
7 In the tree view, expand the folder that corresponds to the operating
system on which the server and the application you are testing is running.
8 Select the server application you want to monitor from the list that
appears. To monitor the operating system, select System.
9 Click Next.
The Connection Parameters dialog appears.
10 In the Hostname field, enter connection parameters such as the host
name or IP address of the computer that hosts the application, the
connection port, the user name, and the password. The data you enter
here depends on the operating system running on the computer you are
monitoring.
11 Click Next.
The Select Displayed Measures dialog appears.
12 Expand the tree view and select the factors you want to monitor.
13 Click Finish.
A Monitor Graph appears, with the elements you specified being shown
in a live, color-coded display of the server performance. Beneath the
graph is a list of the elements included, with the key to the color-coding,
and performance information about each element.
Vendors of hardware and software systems provide information about the nature
of their systems via MIBs (Management Information Base). This information
can be retrieved with the help of the SNMP protocol by specifying the
196 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Monitoring server
performance in a
graph via a SNMP
appropriate OID (Object Identifier). MIB browsers assist you in selecting the
right OID stored in the MIB.
Performance Explorer offers you the possibility of selecting OIDs of MIBs from
several sources for use as data sources for server monitoring. For example,
database systems, such as Oracle, offer extensive performance data to be
collected via a SNMP protocol.
Procedure To monitor server performance in a graph via a SNMP:
1 Click the Monitor Server button on the Performance Explorer Workflow
bar.
The Data Source Wizard appears in the System Selection dialog.
2 Select from the predefined data sources.
3 In the tree view, expand the Custom Data folder.
4 Select the SNMP Data option.
5 Click Next.
The Connection Parameters dialog appears.
6 In the Host name field, enter connection parameters, such as the host
name or IP address of the server system you want to monitor, the port
number and other data required to connect to the data source. You must
have access to the specified server, access rights, and the SNMP service
must be configured properly on the server system.
7 Click Next.
The Add SNMP Measures dialog appears.
To use the contents of an MIB with Performance Explorer, the MIB must
be compiled first. A number of common MIBs are delivered in a
precompiled format.
8 Select one of the precompiled MIBs delivered with Performance
Explorer, or select Compile New MIB for an MIB that has been shipped
with the server system. See the procedure below for compiling new
MIBs.
Once you have selected a compiled MIB, its contents will be shown in a
tree view.
9 Select the OID you are interested in and click Get Value(s).
Connection to the server will then be established and the actual data item
shown.
10 Select the corresponding data item and click Add.
11 Repeat step 8 and 9 until you have selected all the OIDs you want.
12 Click Close.
The Select Displayed Measures dialog appears.
SilkPerformer User Guide 197

9 EXECUTING LOAD TESTS
Monitoring Load Tests
13 Select the factors to be monitored and click Finish.
A monitor graph appears with the elements you specified shown in a
live, color-coded server performance display.
To use the contents of an MIB with Performance Explorer, the MIB must
first be compiled. A number of common MIBs are delivered in a
precompiled format.
Compiling a new MIB
Monitoring server
performance in a
report
Monitoring multiple
servers
Procedure To compile a new MIB:
1 In the Add SNMP Measure dialog, select Compile New MIB for an MIB
that has been shipped with the server system.
The Compile New MIB dialog appears.
2 Browse for the MIB file you want to compile and click Open.
3 Click Compile.
If the compilation succeeds, a new name will be added to the list of
compiled MIBs.
If the error message "Could not IMPORT from -MIB" is
returned, compile the MIB referred to in the error message.
4 Click Close.
Now the new MIB can be used as outlined in the above procedure.
Procedure To monitor server performance in a report:
1 Follow the steps for monitoring server performance in a graph, as
explained above.
2 Click Finish.
3 Select Monitor/Clone Monitor Report from the Performance Explorer
menu bar.
A monitor report appears containing the same performance factors as
selected in the graph.
Monitoring information now appears in tabular form in the report.
Note To save the monitoring report to be used for results
exploration together with the load test results, select Monitor/Write
Monitor Data from the Performance Explorer menu bar.
Procedure To monitor multiple servers:
1 For each server you want to monitor, follow the procedure given at the
top of this topic for monitoring server performance in a graph.
2 Separate graphs will appear for each server.
3 Reposition and resize the graphs to facilitate viewing and comparison.
4 To monitor factors from more than one server in a single graph, use the
tree-view area to locate the factors you want to view.
198 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Monitoring Load Tests
5 Drag and drop them onto your selected graph.
Storing server
monitoring data
Procedure To store server-monitoring data:
1 Create a monitor report that monitors server performance, as outlined in
the above procedures. You can store server-monitoring data only from a
monitor report, not from a monitor graph.
2 Once you have populated the monitor report by dragging and dropping
the relevant performance factors, select Monitor/Write Monitor Data
from the Performance Explorer menu bar. The monitor data will be
stored in a time series data (.tsd) file.
3 The file name appears in the File section of the Monitor information area
in the report dialog.
4 To view the data in a graph, select File/New/Monitor Graph from the
Performance Explorer menu bar.
A new monitor graph appears.
5 Drag and drop the performance factors of your choice from the report
onto the graph.
The data appears in graphical form.
Requirements for server monitoring
This section gives information about the installation and configuration
requirements for the servers for which real-time monitoring is available in
SilkPerformer.
Server
Active Server Pages
Anny Way Wap Server
1.0
ColdFusion 4.0, 4.5, 5.0
Installation and configuration requirements
SilkPerformer can monitor Active Server Pages without custom
configuration. Varying counters may be displayed depending on which
Internet Information Server version is installed.
Anny Way Wap Servers are monitored by the Performance Monitor.
No special configuration is required.
To monitor ColdFusion, select the Enable Monitoring option in the
Debug dialog box of the ColdFusion administration site. This is the default
value for Windows NT. For UNIX, you need to perform this same action
after you have installed SilkPerformer.
To monitor ColdFusion on UNIX systems, SNMP is used.
SilkPerformer User Guide 199

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Server
BEA WebLogic 8.x and
9.x
Borland Application
Server 6.6
IBM WebSphere
Application Servers 3.5 /
4.0
IBM WebSphere 6.0, 6.1
Internet Information
Server 3.0 / 4.0 / 5.0/5.1/
6.0
Installation and configuration requirements
SilkPerformer can monitor WebLogic 8.x and 9.x servers without custom
configuration of the application server. The JNDI protocol is used to
connect via the Management EJB (MEJB) to the MBean Server. Default
connection parameters are preconfigured in SilkPerformer.
SilkPerformer needs to know the appserver installation directory or a copy
of it to build the correct classpath.
A JMX browser allows the selection of the MBean attributes to monitor.
An Easy JMX profile can be used to pre-select interesting MBeans.
SilkPerformer can monitor Borland Application Server 6.6 via JSR-160
standard using the Corbaloc protocol over IIOP. The application server
must be configured to allow access to the MBean Server via IIOP.
SilkPerformer needs to know the appserver installation directory or a copy
of it to build the correct classpath.
A JMX browser allows the selection of the MBean attributes to monitor.
An Easy JMX profile can be used to pre-select interesting MBeans.
Certain MBeans will not work properly without applying patch 7697 to
Borland Application Server.
A browser is provided for selecting WebSphere performance counters.
Requirements: WebSphere AdminService must be running on the
WebSphere host. It runs by default unless IBM WebSphere Application
Server Standard Edition is used.
SilkPerformer can monitor WebSphere 6.0 and 6.1 Server without custom
configuration of the application server.
JNDI protocol is used to connect via the Management EJB (MEJB) to the
MBean Server. Default connection parameters are preconfigured in
SilkPerformer.
SilkPerformer needs to know the appserver installation directory or a copy
of it to build the correct classpath. SilkPerformer needs to use the same
IBM JVM version for monitoring.
A JMX browser allows the selection of the MBean attributes to monitor.
SilkPerformer can monitor IIS Web servers without custom configuration.
SNMP is supported only with a custom monitoring configuration.
Windows NT SNMP Service must be installed before the IIS Web server is
installed.
200 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Server
JBoss 4.0
Netdynamics 5
Netscape Enterprise
Server 3.0
Installation and configuration requirements
SilkPerformer can monitor JBoss 4.0 Server without custom configuration
of the application server.
JNDI protocol is used to connect via the Management EJB (MEJB) to the
MBean Server. Default connection parameters are preconfigured in
SilkPerformer.
SilkPerformer needs to know the appserver installation directory or a copy
of it to build the correct classpath.
A JMX browser allows the selection of the MBean attributes to monitor.
Two Easy JMX profiles may be used to pre-select interesting MBeans.
Netdynamics, is supported using SNMP.
For standard SNMP, Netdynamics uses a custom port (default 14600)
instead of port 161.
If the SNMP agent still needs to be configured, you may run the ndConfig
utility.
Note Access to MIB data is restricted to a select list of
communities and computers. You must add the machine on
which SilkPerformer is running to this list. The file containing this
list is called jaw.acl. It is located in the directory
\[Netdynamics?Home]\program \properties\[partitionname]\.
Without this file, any user from any machine would be able to
access the data.
See the NetDynamics Administration Guide, Chapter 7, for the MIB
Scheme of the NetDynamics MIB.
Netscape Enterprise 3.0 server is monitored using the Performance
Monitor. No custom configuration is required.
SNMP is supported only with custom monitoring configuration. Use the
Custom Data Source dialog to create custom measurements.
To support SNMP, a subagent must be deployed for each instance of
Netscape's server manager.
On Solaris, a master agent must be deployed to collect and coordinate
subagent data.
SilkPerformer User Guide 201

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Server
Netscape Enterprise
Server 4.x
Netscape Enterprise
Server 5.x
Oracle Application
Server 10.1.3
Installation and configuration requirements
Netscape Enterprise Server 4.x can only be monitored with a custom
configuration. You can use either NT Performance Monitor Data or SNMP
Data. Run the Custom Data Source dialog to create the custom measures.
Netscape Enterprise 4.0 Setup automatically installs and configures
SilkPerformer counters.
Netscape Server 4.1 has been known to have problems automatically
registering SilkPerformer counters. In such instances, an error message is
generated in the event log. You may have to manually rectify this problem.
To support SNMP, a subagent must be deployed for each instance of
Netscape's server manager.
On Solaris, a master agent must be deployed to collect and coordinate all
subagent data.
Netscape Enterprise Server 5.x can only be monitored using a custom
configuration. You can use either NT Performance Monitor Data or SNMP
Data. Run the Custom Data Source dialog to create the custom measures.
Netscape Enterprise 5.0 Setup automatically installs and configures
SilkPerformer counters.
To support SNMP, a subagent must be deployed for each instance of
Netscape's server manager.
On Solaris, a master agent must be deployed to collect and coordinate all
subagent data.
SilkPerformer can monitor Oracle Application Server 10.1.3 without
custom configuration of the application server.
JNDI protocol is used to connect via the Management EJB (MEJB) to the
MBean Server. Default connection parameters are preconfigured in
SilkPerformer.
SilkPerformer needs to know the appserver installation directory or a copy
of it to build the correct classpath.
A JMX browser allows the selection of the MBean attributes to monitor.
An Easy JMX profile can be used to pre-select interesting MBeans.
202 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Server
Installation and configuration requirements
Oracle 7.3
SilkPerformer can monitor Oracle 7.3 servers without custom
configuration. All counters specified in the opctrnm.ini file can be
monitored. However, only the most important ones are automatically
displayed in SilkPerformer.
If registration of performance counters fails during installation of Oracle
server, you can manually register the opctrnm.ini file by running the lodctr
command.
For SNMP support (custom configuration monitoring only), the Oracle
Intelligent Agent must be installed and running.
Oracle 8
Same as Oracle 7.3, however the .ini file for the counters is called
opctrn8.ini.
No custom configuration is required.
For SNMP support (custom configuration monitoring only), the Oracle
Intelligent Agent must be installed and running.
Oracle 9 Same as Oracle 8.
No special configuration is required.
Silverstream 3.0, 3.7 This application server can be monitored on Windows NT using SNMP.
SUN JVM 1.5 and 1.6
Procedure To enable SNMP support:
1 Run SilverStream designer.
2 Select SilverMaster3 database/Objects.
3 Open AgSNMPGetStats in design mode.
4 Double-click the trigger symbol to open the Properties page.
5 Set the Enabled option to true.
6 Close and save the changes.
SilkPerformer can monitor SUN JVM 1.5 and 1.6 via JSR-160 standard
using the RMI protocol. Any application or application server using the
platform MBean Server may be monitored this way. No custom
configuration is necessary.
A JMX browser allows the selection of the MBean attributes to monitor.
An Easy JMX profile can be used to pre-select interesting MBeans.
SilkPerformer User Guide 203

9 EXECUTING LOAD TESTS
Monitoring Load Tests
Server
Apache
Installation and configuration requirements
The monitoring of Apache Web Servers is supported for the following
operating systems:
HP-UX, Linux, Solaris, OSF1
Requirements: With Apache Web servers, you have the option of using the
built-in status report functionality. This requires that the "mod_status"
module be built in, which is done by default. To get more detailed
information it’s useful to select the compile option "ExtendedStatus On".
To enable status reports only for browsers from the foo.com domain, add
code to your "access.conf" configuration file. The Apache configuration is
maintained in a set of files, usually located in /usr/local/apache/etc or /usr/
local/apache/conf (Solaris) or /etc/httpd/conf (Linux). By default the code
below should be “commented out” in one of the configuration files. If it’s
not, add it to the main Apache config file, httpd.conf. Ensure that the
"Allow" statements are correct, otherwise monitoring won’t work.
SetHandler server-status
order deny, allow deny from all allow from .foo.com
Restart your server. To verify the server's performance module, type the
following url into your browser:
http://hostname/server-status?auto
Monitoring will work if you get a response such as that below:
Total Accesses: 210
Total kBytes: 94
CPULoad: .000278279
Uptime: 366539
ReqPerSec: .000572927
BytesPerSec: .262608
BytesPerReq: 458.362
BusyServers: 1
IdleServers: 6
Scoreboard: ______W.................................................................................
Note: The default Apache monitor only works if Telnet is enabled. If you want to
monitor an Apache server with Telnet disabled, you need to use the WGet
command instead of the default Rexec commands, for example:
wget -q -O - http://hostname/server-status | grep "requests/sec" |
cut -c5- | cut -d\ -f1
204 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Launching Load Tests Via the Command Line
Server
Stronghold
iPlanet Directory Server
5.0
Netscape Directory
Server 4.0/4.1
Operating System Data
Microsoft Proxy Server
2.0
SQL Server 6.5, 7, 2000
Installation and configuration requirements
Stronghold Web Servers running on Linux can be monitored using the
SilkPerformer. Setting up monitoring for Stronghold Web Servers is
analogous to setting up monitoring for Apache Web Servers.
To verify whether Stronghold's performance module is active, type the
following url into your browser:
http://hostname/stronghold-status?auto
iPlanet Directory Servers are monitored by the Performance Monitor.
No special configuration is required.
Netscape Directory Servers are monitored by the Performance Monitor.
No special configuration is required.
Operating System Data can be monitored for the following platforms:
• NT
• Windows 2000
• Windows XP
• Linux
• Solaris
• HP-UX
• AIX
• OSF1
If available on the host, information such as processor utilization, disc
usage, swapfile utilization, and network packets can be monitored.
On Microsoft systems, operating system data is monitored by the
Performance Monitor. For UNIX systems, a ‘Remote Execution Daemon’
must be running. Such daemons run on UNIX systems by default.
Microsoft Proxy Servers are monitored by the Performance Monitor.
No special configuration is required.
The most significant versions of SQL Server can be monitored with the
Performance Monitor.
No special configuration is required.
Launching Load Tests Via the Command Line
SilkPerformer enables load tests to be started directly from a command prompt
window. This is useful when, for example, you want to perform a number of
load tests consecutively by creating and executing a batch file.
SilkPerformer User Guide 205

9 EXECUTING LOAD TESTS
Launching Load Tests Via the Command Line
To start SilkPerformer from a command line, observe the following syntax:
performer [project file [/Automation [refresh rate] [/
WL:workloadname] [/StartMonitor]] [/PatrolReport] [/
Dev]]
Parameter
project file
/Automation
refresh rate
/WL:workloadname
/StartMonitor
/PatrolReport
/Dev
/SOA
Description
Name of the project file; that defines
the load test to be executed.
Automatically start the load test that
is defined by the specified project
file. If you omit this parameter,
SilkPerformer opens the specified
project.
Refresh rate, in seconds, at which
the Monitor window is to be
refreshed. If you omit this
parameter, then no refresh will
occur.
Name of the workload in the
specified project file. If you specify
a workload, it will be set as the
Active Workload in the project file.
Start Performance Explorer’s
monitoring facility when running
the test.
Create a result file, named
LoadTestReport.xml, at the end of
the load test where all agent
summary measurements are stored.
Note: You must be licensed for Silk
Performer KM to use this switch.
Specifies that SilkPerformer
Developer Workbench be launched.
Specifies that SilkPerformer SOA
Workbench be launched.
/? or -? Displays help about the
SilkPerformer command line
interface.
Table 8: Command line parameters
206 SilkPerformer User Guide

9 EXECUTING LOAD TESTS
Launching Load Tests Via the Command Line
Example
performer
"\Working\Projects\AllDayWorkload\AllDayWorkl
oad.ltp"
/Automation 5 /WL:AllDayWorkload10 /PatrolReport
Note The parentheses (") around the project path are needed when
files or folders in the project path contain spaces.
Note SilkPerformer will close immediately if an error (e.g., a
misspelled project path, script error, workload not in the specified
project file) occurs during the execution of the command line
instruction. Errors can be found in the Event Viewer of the operating
system.
SilkPerformer User Guide 207

9 EXECUTING LOAD TESTS
Launching Load Tests Via the Command Line
208 SilkPerformer User Guide

10
Exploring 10 Chapter
Test Results
Introduction
What you will learn
This chapter explains how to analyze the test results that are generated by
SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 209
Performance Explorer 212
Reports 229
Measurements 230
Overview
The ninth and final step in conducting a SilkPerformer load test is to display and
analyze the results of the test to assess the performance of the application and
the server under test.
Results overview
SilkPerformer’s Results tab in the top left area of the GUI offers full control over
your test results. Selecting the Results tab displays a tree view containing all test
results of the currently active project, including TryScript runs, load test results,
and Test Manager results (if your SilkPerformer project is linked to a Test
Manager test definition).
Depending on the logging options that you have set for test runs (see “Result
options”), the following items are displayed in the results tree:
• Baseline Report
• Time Series
• TrueLog Files (contains TrueLog files for each virtual user)
SilkPerformer User Guide 209

10 EXPLORING TEST RESULTS
Overview
• User Results (contains user profiles, which list all result files per virtual
user)
• Test Manager (displays results from Test Manager, if the project is
linked to a Test Manager test definition)
To view results, double-click a file in the results tree, or use the context menu by
right-clicking a result file and selecting Open or Explore. Opening a results file
invokes the respective viewing facility (see “Results viewing facilities”).
To delete locally stored results (this does not affect results that are stored in Test
Manager), right-click a top-level results node and select Delete Results.
To refresh displayed results (the results that are displayed in the tree view do not
update if they are changed outside of SilkPerformer), right-click a top-level
results node and select Update Results.
If the icon of a parent node in the results tree is marked with a red X symbol, an
error has occurred during the test execution. If an icon (parent node or result file)
is marked with a blue exclamation mark (!), an error log file contains an
informational message, a warning, or an error.
Test Manager
integration
Results viewing
facilities
SilkPerformer is tightly integrated with SilkCentral Test Manager. If your
SilkPerformer project is linked to Test Manager, you can store your test
results centrally in Test Manager where they can be accessed from both Test
Manager and SilkPerformer.
To view Test Manager results, your project must be linked to a Test Manager
test definition (see “SilkCentral Test Manager Integration” for detailed
information on linking SilkPerformer projects to Test Manager). Once your
project is linked to Test Manager, you can view the centrally stored results by
clicking the Click here to add Test Manager results node in the results tree
(see “Downloading Test Results from Test Manager” for information on
downloading Test Manager results).
Additionally, you can upload results to Test Manager by right-clicking a toplevel
results node and selecting Upload Results to Test Manager (see
“Uploading Test Results to Test Manager” for detailed information).
Comprehensive facilities are available in SilkPerformer for displaying,
reporting, and analyzing test results. Defined measurements take place during
tests and can be displayed in a variety of graphical and tabular forms.
• “Performance Explorer”: This is the primary area for viewing test
results. A fully comprehensive array of graphic features displays the
results—primarily in user-defined graphs—with as many elements as are
required. The results of different tests can be compared, and there are
extensive features for server monitoring. A comprehensive HTML based
overview report that combines usergroup data from the baseline report
file (baselineReport.brp) with time series test result information, is
available.
210 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Overview
Verification under load
• “TrueLog On Error”: SilkPerformer provides full visual verification
under load capabilities for Web applications. It allows you to combine
extensive content verification checks with full error drill-down analysis
during load tests.
• “Virtual User Report files”: These files contain the simulation results for
each user, if their generation is enabled. Details of the measurements for
each individual user are presented in tabular form.
• “Virtual User Output files”: These files contain the output of write
statements used in test scripts, if their generation is enabled.
• “Viewing Baseline Reports”: A detailed XML/XSL-based report that
provides you with a summary table, transaction response-time details,
timers for all accessed HTML pages, Web forms, and errors that
occurred. This information is available for all user types involved in
baseline tests.
• SilkCentral Test Manager Reports: SilkPerformer projects can be
integrated into SilkCentral Test Manager (Test Manager) test plans and
directly executed via Test Manager. This allows for powerful test-result
analysis and reporting. For detailed information on Test Manager
reporting, please refer to the SilkCentral Test Manager User Guide.
SilkPerformer provides full visual verification under load capabilities for Web
applications. It allows you to combine extensive content verification checks
with full error drill down analysis during a load test. SilkPerformer verification
checks can be done from all virtual users without major performance loss.
Thanks to SilkPerformer's innovative TrueLog technology, it is possible to
collect details regarding the errors of all virtual users that take part in a load test.
Traditional probing clients, separated from the load testing tool, are only able to
detect content verification errors that occur for users that actually perform
verifications. Since a functional testing tool is only able to drive a very limited
number of concurrent users, the functional test covers only a very small subset
of all users that visit your Web application. Such a probing client cannot detect
verification errors that occur to users that are not contained within this small
subset. Hence, SilkPerformer redefines the type of testing that can be done with
a load testing tool.
With the TrueLog technology, you are able to find a new class of errors and
their root causes—including errors that usually occur to only a subset of users
when your application is under a heavy load. For most applications, this is the
type of load that will most likely be experienced once the application is
deployed in the real world. Typical errors include incorrect text on a Web page,
incorrectly computed and displayed values, or application-related messages,
such as "Servlet Error" or "Server Too Busy" errors. These are not system-level
errors and are displayed on Web pages with HTTP 200 status codes.
SilkPerformer User Guide 211

10 EXPLORING TEST RESULTS
Performance Explorer
Additionally, SilkPerformer provides a visual view of its new verification under
load capability to simplify its use:
- Visual content verification allows you to visually define the content
that needs to be verified.
- TrueLog On Error generation and TrueLog On Error analysis allow
you to visually analyze errors to find their root cause.
TrueLog On Error
As a foundation for full visual root cause analysis, SilkPerformer can save the
complete history of errors within a TrueLog. TrueLog On Error can even be
generated in large-scale load tests. TrueLog On Error helps you uncover the true
source of your application's error output under all possible load conditions.
Generated log files can be displayed by means of TrueLog Explorer to verify the
cause of errors raised by application servers.
1 Before you begin a load test, select the TrueLog On Error checkbox on
the Workload Configuration dialog.
2 If errors occur during load tests, you can view the visual content of the
TrueLog by clicking the Explore Results button on SilkPerformer's
workflow bar.
3 TrueLog Explorer opens and displays the content of the TrueLogs.
4 If TrueLog On Error files are not displayed automatically, click Analyze
TrueLog On Error and navigate to the directory of the actual load test.
Select the TrueLog files to be displayed and click OK.
5 You can navigate from one occurrence of an error to the next.
6 To display the history of an error, click the API nodes in the tree view on
the left.
Performance Explorer
Performance Explorer is the primary tool used for viewing load test results. A
vast array of graphic facilities allows both real-time monitoring of the target
server while tests are being run and exhaustive analysis of the results after tests
have completed. Exploring test results is made easy by a workflow bar with a
click-through user interface, enhanced by drag-and-drop functionality.
In real-time monitoring, live charts provide a customizable display of the most
relevant performance information from the target server. Monitoring is available
for a comprehensive collection of the Web servers, application servers, and
database servers running on all of the operating systems most widely used
today. Multiple charts can be open at the same time, so that, for example, the
tester can watch a graphic display of Web server performance and operating
212 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
Performance Data
Collection Engine
(PDCE)
Viewing Overview
Summary Reports
system performance simultaneously. A tree-view editor with drag-and-drop
functionality allows elements from any data source to be combined in the charts.
After a test, the performance of the target server can be charted from both the
client side and the server side. Response time measurements display the client
perspective, while throughput data gives the perspective from the server side.
Charts and graphs are fully customizable, and they can contain as many or as
few of the measurements taken during the test as the tester needs. Multiple
charts, using information from one or from different tests, can be opened at once
to facilitate contrast/compare operations. Templates for the most typical test
scenarios (Web, Database, IIOP) are provided, and these default charts can be
populated easily and quickly with the data the tester requires. Here also, dragand-drop
functionality enables chart elements to be combined from any data
source. Information on client response times and server performance can be
placed in a single chart, so that testers can see directly how server performance
affects client behavior.
A comprehensive XML/XSL based overview report, which combines usergroup
data from the baseline report file (baseline.brp) with time series test result
information, is available. You can customize the overview report to your needs
by changing the provided texts and by inserting charts of your interest. Saving
your changes as a template allows the reuse of your individual settings.
The PDCE (Performance Data Collection Engine) is a Performance Explorer
engine that is used for online monitoring. Its interface can alternately be
accessed via Benchmark Description Language (BDL), which is used by
SilkPerformer to monitor data sources, such as PerfMon, SNMP and
WebSphere 4.0. The BDL interface for the PDCE is described in the pdce.bdh
include file.
Each of the measurements that Performance Explorer monitors can
alternately be monitored via SilkPerformer using BDL scripts. After
registering measurements of interest, a monitoring data set can be retrieved
from the PDCE via the provided polling functions (PDCE API). These
functions provide data, status information, and descriptions of the
measurements (if available). Further information regarding the PDCE API
can be found in the BDL Function Reference.
Procedure To view an overview summary report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button or link.
Performance Explorer opens and displays an overview summary report
for the most recent load test, if you have selected the Generate overview
report automatically option in the Settings/Options/Reporting dialog.
Additionally, you can choose to use a previously stored template for the
SilkPerformer User Guide 213

10 EXPLORING TEST RESULTS
Performance Explorer
Customizing Overview
Summary Reports
Adding a default chart
Adding a custom chart
generation of an overview report in this dialog. This setting is a global
setting and will be used with Performance Explorer, regardless of the
workload project you are using.
3 If the overview report does not appear automatically, click the Overview
Report button on the Workflow bar.
4 Navigate to the directory of the load test you are working with and select
the corresponding .tsd file.
5 In the next step you are asked to select the template you want to use for
the generation of the overview report.
6 Click Finish and the overview report will appear.
Performance Explorer enables to you add customized charts and reports to
overview reports. You can remove charts and reports from the overview report
at any time and you can save your changes as a template to be used for the
generation of overview reports in the future.
Procedure To add a default chart to the overview report:
1 Select the overview report to which you want to add the chart.
2 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom charts section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog then appears.
3 Select all sections that you want displayed in the overview report. By
default, all available sections are displayed. Click Next.
4 Select Add new view and click Next.
5 Expand the Graphs folder and select the type of graph you want to insert
into the overview report.
6 Click Finish.
Performance Explorer opens a new view with the type of chart you have
selected. In addition, Performance Explorer inserts the chart into the
custom charts section of the overview report.
Note When you add new measurements to the view that contains the
graph, or when you remove measurements from the view, the
corresponding graph in the overview report is automatically
updated.
Procedure To add a custom chart to the overview report:
1 Generate the graph you want to add to the overview report.
2 Select the overview report to which you want to add the chart.
214 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
Adding a default
report
3 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom charts section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog opens.
4 Select all sections that you want to have displayed in the overview
report. By default, all available sections are displayed. Click Next.
5 Expand the Charts folder and select the chart you have just created. If
you want to add a new chart to the overview report, select the Add new
view option and click Next.
6 Click Finish.
Performance Explorer inserts the chart you have selected into the custom
charts section of the overview report.
Note To change the name and the description of a chart, right click
the bottom area of the graph in Performance Explorer and select
properties. There you can specify a name and a description, which
also appears in the overview report.
Note When you add new measurements to the view that contains the
graph, or when you remove measurements from the view, the
corresponding graph in the overview report is automatically
updated.
Procedure To add a default report to an overview report:
1 Select the overview report to which you want to add the report.
2 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom tables section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog opens.
3 Expand the Reports folder and select the type of graph you want to insert
into the overview report.
4 Click Finish.
Performance Explorer opens a new view with the type of report you have
selected. In addition, Performance Explorer inserts this report into the
custom tables section of the overview report.
Note When you add new measurements to the view that contains the
report, or when you remove measurements from the view, the
corresponding report in the overview report is automatically
updated.
SilkPerformer User Guide 215

10 EXPLORING TEST RESULTS
Performance Explorer
Adding a custom
report
Removing a chart
Removing a report
Procedure To add a custom report to the overview report:
1 Generate the report you want to add to the overview report.
2 Select the overview report to which you want to add the report.
3 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom reports section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog opens.
4 Expand the Reports folder and select the report you have just created.
5 Click Finish.
6 Performance Explorer inserts the report you have selected into the
custom tables section of the overview report.
Note To change the name and the description of a report, right click
the report in Performance Explorer and select properties. There you
can specify a name and a description, which will also appear in the
overview report.
Note When you add new measurements to the view that contains the
report, or when you remove measurements from the view, the
corresponding report in the overview report is automatically
updated.
Procedure To remove a chart from the overview report:
1 Select the overview report from which you want to remove a chart.
2 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom charts section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog opens.
3 Expand the Charts folder and disable the chart you want to remove from
the overview report. If you want to add a new chart to the overview
report, select the Add new view option and click Next.
4 Click Finish.
Performance Explorer removes the chart you have selected from the
custom charts section of the overview report.
Procedure To remove a report from the overview report:
1 Select the overview report from which you want to remove a report.
2 In the Workflow toolbar, click the Customize Report button.
Alternately, navigate to the custom tables section of the overview report
and click the Click here to customize this section link.
The Custom Overview Report Wizard dialog opens.
216 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
Changing text areas
Saving report layouts
Specifying a default
template
Viewing results
3 Expand the Reports folder and disable the report you want to remove
from the overview report.
4 Click Finish.
Performance Explorer removes the report you have selected from the
custom tables section of the overview report.
Procedure To change text areas in the overview report:
1 In an overview report, many text areas are predefined, but you can
change them according to your needs. When you see click here to edit
you can change the corresponding text and save the report as a template
to be used for further reports.
Procedure To save the layout of a report as a template:
1 Customize the overview report based on your needs.
2 Select File - Save as template from the menu bar and provide a name for
the new template.
The template will be stored in the \working\include folder, to be used for
all projects. Have a look at Settings / Options / Reporting, in which you
can specify this template to be used when creating a new overview
report.
Procedure To set a template as the default template:
1 From the menu bar, select Settings / Options and then the Reporting tab.
2 Check Use template when generating a new overview report and select a
previously stored template.
The selected template will be used to create a new overview report for all
projects.
Performance Explorer offers numerous options for viewing results.
Default results
Viewing default
results in a graph
Procedure To view the default set of results in a graph:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Click the View Graph button on the Performance Explorer Workflow
bar.
The default graph for the current project appears.
SilkPerformer User Guide 217

10 EXPLORING TEST RESULTS
Performance Explorer
Viewing default
results in a report
Procedure To view the default set of results in a report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Click the View Report button on the Performance Explorer Workflow
bar.
The default report for the current project appears.
Specific results
Viewing specific
results in a graph
Viewing specific
results in a report
Procedure To view a specific set of results in a graph:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Click the Select Graph button on the Performance Explorer Workflow
bar.
The Template dialog appears.
4 In the Templates area, expand the folder for the measurement type you
want to view, either Response Time or Throughput. To view the default
graph, select General.
5 Select the measure you want to view.
6 Click Finish.
A graph displaying the measure you selected appears.
Procedure To view a specific set of results in a report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Click the Select Report button on the Performance Explorer Workflow
bar.
The Template dialog appears.
4 In the Templates area, expand the folder for the measurement type you
want to view, either Response Time or Throughput. To view the default
report, select General.
5 Select the measure you want to view.
218 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
6 Click Finish.
A report giving information on the measure you selected appears.
Custom results
Viewing custom
results in a graph
Viewing custom
results in a report
Procedure To view a custom set of results in a graph:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Graph from the Performance Explorer menu bar.
The Select Result File dialog appears.
4 Locate the time series data (.tsd) file for which you want to see a graph
and click Open.
Performance Explorer displays a tree view with all the measurements
that the selected file contains.
5 In the tree-view area, expand the folder containing the measure types you
want to add to the graph.
6 Drag and drop the measure you want onto the graph.
7 Repeat the drag-and-drop procedure as often as is necessary to add
elements to the graph.
Procedure To view a custom set of results in a report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Report from the Performance Explorer menu bar.
The Select Result File dialog appears.
4 Locate the time series data (.tsd) file for which you want to see a report
and click Open.
5 In the tree-view area, expand the folder that contains the measure types
you want to add to the report.
6 Drag and drop the measure you want onto the report.
7 Repeat the drag-and-drop procedure as often as is necessary to add
elements to the report.
SilkPerformer User Guide 219

10 EXPLORING TEST RESULTS
Performance Explorer
Multiple tests
Viewing multiple test
results in a graph
Viewing multiple test
results in a report
Procedure To view the results of multiple load tests in a graph:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Graph from the Performance Explorer menu bar.
A new, empty graph appears.
4 From the Performance Explorer menu bar, select Explore/Add Results;
or right-click the project name in the tree-view area and choose Add
Results.
The Select Result File dialog appears.
5 Use this dialog to specify results information from any load test to be
made available for the graph.
The data is added to the resources available in the tree-view area.
6 Continue adding results information from as many load tests as is
required.
7 In the tree-view area, expand the folders for the load tests and measure
types you want to see in your graph.
8 Drag and drop the measures you require onto the graph.
Procedure To view the results of multiple load tests in a report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Report in the Performance Explorer menu bar.
A new, empty report appears.
4 From the Performance Explorer menu bar, select Explore/Add Results;
or right-click the project name in the tree-view area and choose Add
Results.
The Select Result File dialog appears.
5 Use this dialog to specify results information from any load test to be
made available for the report.
The data is added to the resources available in the tree-view area.
6 Continue adding results information from as many load tests as is
required.
220 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
7 In the tree-view area, expand the folders for the load tests and measure
types you want to see in your report.
8 Drag and drop the measures you require onto the report.
Monitoring data
Viewing results and
monitoring data in a
graph
Viewing results and
monitoring data in a
report
Procedure To view test results and monitoring data in a graph:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Graph from the Performance Explorer menu bar.
A new, empty graph appears.
4 From the Performance Explorer menu bar, select Explore/Add Results;
or right-click the project name in the tree-view area and choose Add
Results.
The Select Result File dialog appears.
5 Use this dialog to specify results information from the load test you want
to be made available for the graph.
These tsd files contain the following data: m@... contains merged times
series data of the load test, i@... contains information from the agent,
u@... contains user specific data, r@... contains server monitor data
The data is added to the resources available in the tree-view area.
6 Use the Select Result File dialog again, this time to locate monitoring
information generated during server monitoring.
7 In the tree-view area, expand the folders for the load test and monitoring
data you want to see in your graph.
8 Drag and drop the measures you require onto the graph.
To view test results and monitoring data in a report:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
3 Select File/New/Report from the Performance Explorer menu bar.
A new, empty report appears.
SilkPerformer User Guide 221

10 EXPLORING TEST RESULTS
Performance Explorer
Customizing graphs
4 From the Performance Explorer menu bar, select Explore/Add Results;
or right-click the project name in the tree-view area and choose Add
Results.
The Select Result File dialog appears.
5 Use this dialog to specify results information from the load test you
require to be made available for the report.
The data is added to the resources available in the tree-view window.
6 Use the Select Result File dialog again, this time to locate monitoring
information generated during server monitoring. Files beginning with
r@... contain server monitoring data.
7 In the tree-view area, expand the folders for the load test and monitoring
data you want to see in your report.
8 Drag and drop the measures you require onto the report.
Procedure To customize the graph you are viewing:
1 To view and edit properties for your graph, click the View Properties
button on the Performance Explorer toolbar.
The Properties dialog appears at the General tab.
2 Use the View Description area to see information about the current
graph.
3 Click the Chart tab to edit display options for the current graph.
4 Use the View Options area to specify chart settings. You can specify
whether to display the currently selected series as a thicker line and
whether to assume that all time series start at the same time.
5 Select the Highlight Selected Series option to display the currently
selected series as a thicker line. While this option may be useful as long
as you are working with your charts, you should disable the option
before you export or print your charts.
6 Select the Overlay series option to assume that all time series that are
displayed in a chart start at the same time. You should use this option
when you are viewing a chart that contains time series from load tests
with different starting times.
7 From the Background Color drop-down lists, select two colors that you
want to use to enhance your charts. Performance Explorer displays a
vertical color transition from the first to the second color as the
background image. To use custom colors, click the buttons to the right of
the drop-down lists.
222 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
8 From the Text and grid color drop-down list, select the color that you
want to use for the axis information and the grid in your charts. To use a
custom color, click the button to the right of the drop-down list.
9 Use the Scaling Options area to determine whether you want to display
real time series data values or to scale values between 0 and 100.
10 Select the Scale displayed values between 0 and 100 option to display
scaled values rather than real time series data values.
Select the Display real values option to display real time series data
values rather than scaled ones.
11 Click OK exit the dialog and implement your changes in the graph.
12 To change the appearance of a selected time series element in your
graph, use the Series Color, Line Width, and Line Style drop-down lists
in the Performance Explorer toolbar.
13 To expand an area in your graph, click the left edge of the area you want
to expand and drag to the right edge of the area.
The area is highlighted as you drag, and is expanded when you release
the mouse button.
14 To undo the area expansion, drag in the opposite direction on the graph,
or click the Zoom out button on the toolbar.
15 Use the Source Type drop-down list to choose the way in which
measurement information for the time period in question is presented.
Choose to have Performance Explorer calculate and use in the graph a
sum of the values, their average, a count, the minimum or maximum
value, or the standard deviation from the average.
16 Use the Series scaling button to alter the multiple by which selected
elements in a graph are scaled.
17 Use the Series Type button to display graph elements as a simple line (the
default), as a candle graph, as a candle graph with sticks or as an area
series type. In a candle graph, each measurement is displayed with a
vertical marker with a central button. The central button represents the
average value for the time period involved, and the top and bottom of the
marker represent the highest and lowest values of the standard deviation
from the average. In a candle graph with sticks, two further values are
displayed as lines on top of and below the marker; these are the
maximum and minimum values for the period in question. In an area
series type graph the data area is completely colored in.
18 Use the Show Series Markers button to alternate between displaying
actual measurement values on the graph itself or not.
SilkPerformer User Guide 223

10 EXPLORING TEST RESULTS
Performance Explorer
Comparing graphs
Viewing a results
graph in a Web
browser
19 Use the Show Null Values button to show values where no measurements
were made during the interval.
20 Use the Toggle Highlight button to alternate between showing a selected
graph element (or elements) highlighted with a bold line or not.
21 To choose between seeing the lower part of the graph window in which
the measure type legends appear or not, use the Toggle legends button.
22 To choose between having a frame with title bar or not on your graph,
use the Toggle Frames button.
23 Use the Toggle Scaling button to alternate between showing real values
and showing values scaled between 0 and 100 in the vertical axis in your
graph. Note that Performance Explorer initially chooses a scale for the
graph that will provide a clear display.
24 To use logarithmic scaling for the vertical axis of your graph, use the
Logarithmic Y-axis button.
25 Use the Toggle Overlay button to have the start points of all measures
displayed in your graph coordinated at the zero point on the horizontal
axis. This is useful where comparisons need to be made between
measurements from load tests that started at different times.
Procedure To use similar graphs for comparing different sets of results:
1 With the initial graph that you want to use for comparison purposes
open, click the Compare Graph button on the Performance Explorer
Workflow bar.
The Compare TSD File dialog opens.
2 Locate and open the time series data (.tsd) file from the load test whose
results you want to see in a graph similar to your initial one.
A new graph appears, populated with the same elements as the initial
graph, if available.
3 Resize and relocate the graphs as necessary to facilitate comparisons.
Procedure To view a results graph in a Web browser:
1 In Performance Explorer, create or open the graph that you want to view
in a Web browser.
2 Click the View as HTML button on the Performance Explorer Workflow
bar.
The View as HTML dialog appears.
3 Click OK.
Your Web browser opens and displays your graph, now ready for
publication on the Web, if you want.
224 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
Viewing a results
report in a Web
browser
Setting chart options
Procedure To view a results report in a Web browser:
1 In Performance Explorer, create or open the report that you want to view
in a Web browser.
2 Click the View as HTML button on the Performance Explorer Workflow
bar.
The View as HTML dialog appears.
3 Click OK.
Your Web browser opens and displays your report.
Procedure To set default charting options:
1 Click the Explore Results button on the SilkPerformer Workflow bar.
The Workflow - Explore Results dialog appears.
2 Click the Performance Explorer button.
Performance Explorer opens.
You can also access Performance Explorer by selecting Results/Explore
Time Series from the SilkPerformer menu bar.
3 Select Settings/Options from the Performance Explorer menu bar.
The Performance Explorer Options dialog opens at the Chart tab.
4 Use the View Options area to specify chart settings. You can specify
whether to display the currently selected series with an emphasized,
thicker line and whether to assume that all time series start at the same
time.
5 Select the Highlight Selected Series option to display the currently
selected series with an emphasized, thicker line. While this option may
be useful as long as you are working with your charts, you should disable
the option before you export or print your charts.
6 Select the Overlay series option to assume that all time series that are
displayed in a chart start at the same time. You should use this option
when you are viewing a chart that contains time series from load tests
with different starting times.
7 From the Background Color drop-down lists, select two colors that you
want to use to enhance your charts. Performance Explorer displays a
vertical color transition from the first to the second color as the
background image. To use custom colors, click the buttons to the right of
the drop-down lists.
8 From the Text and grid color drop-down list, select the color that you
want to use for the axis information and the grid in your charts. To use a
custom color, click the button to the right of the drop-down list.
SilkPerformer User Guide 225

10 EXPLORING TEST RESULTS
Performance Explorer
Setting time series
data options
Setting options for
exporting result
graphs
9 Use the Scaling Options area to determine whether you want to display
real time series data values or to scale values between 0 and 100.
10 Select the Scale displayed values between 0 and 100 option to display
scaled values rather than real time series data values.
Select Display real values option to display real time series data values
rather than scaled ones.
11 Click OK.
Procedure To set default options that are specific to time series data:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
2 Click the Series tab.
3 In the Default settings area, select the Ignore NULL values option if you
want to hide null values by default. You can use the Show Null Values
button on the Performance Explorer toolbar to specify whether to display
null values separately for each series.
4 From the Show summary and counter values per drop-down list, select
the time unit for which you want time series data to be displayed. Select
the Interval at which you want to have data displayed for the
computation interval contained in the tsd, or select Second to have data
displayed scaled to one-second intervals. To change the default
computation interval, use the Time Series tab of the Profile Settings –
Results dialog.
5 From the Default line width drop-down list, select a line width that will
be used whenever you add a new time series to a chart.
6 Use the Markers area to specify whether to display chart markers
indicating exact measurement values for each series.
7 Select the Show chart markers option to display a chart marker for each
measure value of each series, indicating the exact measure value.
8 In the Decimal digits field, enter the number of digits to be displayed in
the chart markers. Chart markers can be used to display exact
measurement values for each series within a chart.
9 Click OK.
Procedure To set default options for exporting results graphs:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
2 Click the HTML View tab.
226 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Performance Explorer
Setting options for
candle graphs
Setting monitoring
options
3 Select an option in the Chart picture resolution area to determine the
resolution of the image that is generated when you export a chart as an
HTML document. Keep in mind that the higher the resolution, the larger
the generated image file will be and the longer it will take to transfer the
HTML document over the Web.
Select the User-defined option to use a custom resolution, and add
appropriate values in the fields.
4 Use the Picture type area to select the picture format of your chart in an
HTML document.
Select the JPEG option to include the chart in an HTML document as a
JPEG file (Joint Photographic Experts Group).
Select the BMP (only with Internet Explorer) option to include the chart
in an HTML document as a BMP file (Microsoft Bitmap).
5 Click OK.
Procedure To set default options for candle graphs:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
2 Click the Candles tab.
3 Use the Candle style area to define the width and the border color for the
candles in your current graph.
4 From the Border color drop-down list, select the color you want to use
for the candle borders in your current graph. Click the button to the right
of the drop-down list to define a custom color.
5 In the Candle width field, enter the width for the candles in your current
graph. Click the buttons to the right of the field to raise and lower the
width.
6 Use the Marker style area to define the height and color for the markers
that indicate the mean values of the series in your current chart.
7 From the Marker color drop-down list, select the color you want to use
for the markers. Click the Browse (...) button to the right of the dropdown
list to define a custom color.
8 In the Marker height field, enter the height for the markers. Click the
buttons to the right of the field to raise and lower the height.
9 Click OK.
Procedure To set default monitoring options:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
SilkPerformer User Guide 227

10 EXPLORING TEST RESULTS
Performance Explorer
Setting time series
data writer options
Setting report options
2 Click the Monitor tab.
3 In the Interval fields in the Monitor options area, set the length of the
intervals at which the graph is updated. Click the buttons to the right of
the field to raise and lower the interval and to set the interval time unit.
4 In the Page size field, set the number of measure points currently
displayed per graph on the horizontal axis.
5 In the History size field, set the number of measure points stored for the
entire graph on the horizontal axis.
6 Use the View options area to set display options for your real-time charts.
You can change the color settings in your charts, in the axis information,
and in your chart grid.
7 From the Background colors drop-down lists, select two colors that you
want to use to enhance your charts. Performance Explorer displays a
vertical color transition from the first to the second color as the
background image. Click the buttons to the right of the drop-down lists
to select custom colors.
8 From the Text and grid color drop-down list, select the color that you
want to use for the axis information and the grid in your charts. Click the
Browse (...) button to the right of the drop-down lists to select a custom
color.
9 Click OK.
Procedure To set default options for the time series data writer:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
2 Click the TSD-Writer tab.
3 Use the TSD-Writer area to set options for the facility that creates time
series date (.tsd) files from monitoring data.
4 In the Interval fields in the TSD-Writer area, select a time interval value
and a time unit for which time series data will be written.
5 Click OK.
Procedure To set default options for reporting:
1 From the Performance Explorer menu bar, select Settings/Options.
The Performance Explorer Options dialog opens at the Chart tab.
2 Click the Reporting tab.
3 Select the Generate overview reports automatically option for
Performance Explorer to automatically display a summary overview
report when you start exploring the results of a load test. An overview
228 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Reports
report contains the most important results of a load test, for example, the
number of virtual users that were active, the number of transactions that
were executed, and the number of errors that occurred.
4 Select the Show NULL values in charts option if you want null values to
be displayed in the charts. If you disable this option, Performance
Explorer replaces each null value with a direct line from the preceding to
the following value.
5 Select Use template when creating a new overview report if you want to
use a previously stored template for the creation of overview reports.
Select an appropriate template with the browse button.
6 Click OK.
Reports
Virtual User Report
files
Virtual User Output
files
Overview Report
Report files contain the simulation results for each virtual user. Details of the
measurements for each individual user are presented in tabular form. The many
sections in these files include reports on transactions, timers, Web page timers,
and counters, and also sections on IIOP, Web form, TUXEDO, and SQL results.
Output files contain the output of write statements used in the test script. An
output file is generated for a particular user only if write statements are executed
by that user.
However, generating output files alters the time measurements of a load test.
Therefore, these files should be used for debugging purposes only, and should
never be generated for a full load test.
After a load test has completed, Performance Explorer provides an overview
report for the load test. This report includes the most important test results in
tabular and graphical form. It comprises the following sections:
• General information
• Summary tables
• User types
• Custom charts
• Custom tables
• Detailed charts
General information
The general information section includes administrative information in tabular
form as well as important load test results in a graphical form.
Administrative information includes the project name, a description of the
project, the load test number, a description of the load test, the date of the load
SilkPerformer User Guide 229

10 EXPLORING TEST RESULTS
Measurements
test, the duration of the load test, the number of used agent computers, and the
number of virtual users that were running.
The charts display the number of active virtual users, response time
measurements for transactions, and the number of errors that occur over time.
Transaction response times are provided for successfully executed transactions,
for failed transactions, and for cancelled transactions.
Additional charts display summary measurements related to the type of load
testing project. For example, in the case of Web application testing, response
time measurements for Web pages are presented in a graph.
Summary tables
This section contains summary measurements in tabular form, that is, aggregate
measurements for all virtual users. The first table provides general information,
such as the number of transactions that were executed and the number of errors
that occurred. All the following tables provide summary information relevant to
the type of application that was tested.
User types
For each user group, this section provides detailed measurements in tabular
form. The measurements include transaction response times, individual timers,
counters, and response time and throughput measurements related to the type of
application that was tested (Web, database, CORBA, COM, or TUXEDO). In
addition, errors and warnings for all user groups are listed.
Custom charts
This section contains graphs that you have added manually. You can add charts
to and remove charts from this section at any time. You can save your changes
as a template to be displayed for every summary report.
Custom tables
This section contains tables that you have added manually. You can add tables
to and remove tables from this section at any time. You can save your changes
as a template to be displayed for every summary report.
Detailed charts
This section provides enlarged versions of the charts included in the report.
Click a reduced version of a chart to jump to the enlarged version, and vice
versa.
Measurements
A measure (also named a “measure type”) is a generic term for a measurement
of performance taken during a load test. Measures can be either timers or
counters. Timers collect data related to response times, and counters collect data
on throughput and load test events.
SilkPerformer provides the following types of measurements:
230 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Summary
measurements
• “Summary measurements”
• “Individual measurements”
• “Web page measurements”
• “Web form measurements”
• “WAP-specific measurements”
• “CORBA-specific measurements”
• “Database-specific measurements”
• “Middleware-specific measurements”
• “Citrix-specific measurements function”
• “Custom measurements”
The Summary measurement group contains summarized measurements on a
global level. It contains measurement types that aggregate individual
measurements from other measurement groups as well as measurement types
that represent information on a global level that is not included in other
measurement groups. The summary measurement group contains only counters;
it does not contain timers. It represents a fixed set of global measurements that
are collected only during the measurement period, thus excluding the warm-up
and shutdown period of the load test. The Summary group contains the
following measurement types, which are available as real-time data and time
series data (tsd):
General measurement types
Measurement
Description
Active users The number of active virtual users. A virtual user is
regarded as active if it has been started and is
currently in one of the following states: executing,
warm up, wait database, loading documents, wait
resource, rendezvous, thinktime, and shutdown.
Transactions The number of SilkPerformer transactions that
have completed.
Errors
The number of API errors, including Internet,
WAP, database, and middle ware APIs. An API
error is counted if the severity is "error" or higher
("transaction exit" or "process exit"). An error is
ignored if the severity is "informational" or
"warning."
Table 9: General measurement types
SilkPerformer User Guide 231

10 EXPLORING TEST RESULTS
Measurements
Agent health control measurements
Measurement
Internet-related measurement types
Description
CPU utilization The percentage of time that a processor is busy
executing tasks. If the processor utilization is
equal to or near 100%, the computer has run
out of available processing capacity.
Memory
The percentage amount of memory allocated
by the driver machine in relation to the
available amount of RAM. High memory
utilization indicates that the driver machine is
low on available memory, and this can cause
disk swapping, which degrades the overall
performance of the computer.
Responsiveness The ratio of the defined execution time for a
task to its measured execution time, expressed
as a percentage. It indicates how quickly the
computer becomes available for a specific task.
In the case of a lightly loaded driver machine,
the ratio will be equal to or very close to 100%.
Table 10: Agent health control measurements
Measurement
Description
Concurrent connections The number of currently open TCP/IP
connections.
Requests sent
The number of HTTP requests sent by
browser-level or WAP functions, the
number of data packets sent by TCP/IPlevel
functions, or the number of
commands sent by POP3, SMTP, FTP, or
LDAP function calls. SilkPerformer also
includes LDAP traffic in this measurement;
it reports the number of LDAP requests
sent. Note that IIOP uses other counters.
Requests failed
The number of the requests listed above
that are not sent or executed successfully.
Note that IIOP uses other counters.
Table 11: Internet-related measurement types
232 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Measurement
Description
Request data sent[KB] The amount of data (in KB) sent to the
server, including header and body content
information as well as all TCP/IP-related
traffic (HTTP, native TCP/IP, IIOP, POP3,
SMTP, FTP, LDAP), secure traffic over
SSL/TLS, and WAP. It does not include
data overhead caused by encryption using
SSL/TLS or other API-related traffic like
TUXEDO ATMI, ODBC, and Oracle OCI.
Responses received The number of HTTP responses received
by browser-level and WAP functions, the
number of data packets received by TCP/
IP-level functions, and the number of
responses received by POP3, SMTP, FTP,
and LDAP functions. Note that IIOP uses
other counters.
Responses failed
The number of the expected responses that
are not received successfully. Note that
IIOP uses other counters.
Response data received[KB] The amount of data (in KB) received from
the server, including header and body
content information as well as TCP/IPrelated
traffic (HTTP, native TCP/IP, IIOP,
POP3, SMTP, FTP, and LDAP), secure
traffic over SSL/TLS, and WAP. It does
not include API-related traffic like
TUXEDO ATMI, ODBC, and Oracle OCI.
Connects successful The number of successful TCP/IP connects
to the remote hosts established by browserlevel,
TCP/IP-level, IIOP, SMTP, POP3,
FTP, or LDAP functions; this include also
the number of successful connects by WAP
functions.
Table 11: Internet-related measurement types
SilkPerformer User Guide 233

10 EXPLORING TEST RESULTS
Measurements
Measurement
Connects failed
The number of TCP/IP connection failures
caused by invalid host names, invalid IP
addresses, invalid ports, unavailable remote
hosts, network-related problems, or
authentication failures. This counter is used
by browser-level, TCP/IP-level, IIOP,
SMTP, POP3, FTP, and LDAP functions.
Note that by default browser-level
functions perform up to three connect
attempts (up to two retries).
Note: By default browser-level functions
perform up to three connection attempts
(up to two retries). Only if all attempts fail,
a failure is recorded.
This measurement includes also the
number of connections that could not be
established by WAP functions.
Connects retries
The number of attempts to connect to the
server if the initial connection fails.
Table 11: Internet-related measurement types
Web-specific measurements
Measurement
Description
Description
HTTP request retries The number of times a client is forced to retry a
request. This may happen when the server
closes a persistent connection (available with
Keep-Alive or HTTP/1.1) because of a timeout
occurring while the client tries to send a
request. This counter is used only by browserlevel
functions.
HTTP redirections The number of HTTP redirections performed
due to a Web server response with status code
3xx. This measurement is incremented only by
Web and WAP functions.
Table 12: Web-specific measurements
234 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Measurement
HTTP re-authentications
HTTP hits
Hits failed
HTTP 1xx responses
HTTP 2xx responses
HTTP 3xx responses
HTTP 4xx responses
HTTP 5xx responses
HTTP cache conditional
reloads
HTTP cache hits
Description
Table 12: Web-specific measurements
The number of re-authentications necessary
because of a status code 401 or 407 returned by
the Web server. This counter is used only by
Web and WAP functions that perform basic
authentication.
The number of HTTP requests that arrive at the
Web server. This measurement is incremented
only by Web and WAP functions.
The number of failed HTTP requests. This
measurement is incremented only by Web and
WAP functions.
The number of HTTP responses containing
status codes indicating that their content has the
status "informational". This measurement is
incremented by Web and WAP functions.
The number of HTTP responses containing
status codes indicating successful requests.
This measurement is incremented by Web and
WAP functions.
The number of HTTP responses containing
status codes indicating redirected requests. This
measurement is incremented by Web and WAP
functions.
The number of HTTP responses containing
status codes indicating client errors. This
measurement is incremented by Web and WAP
functions.
The number of HTTP responses containing
status codes indicating server errors. This
measurement is incremented by Web and WAP
functions.
The number of requests sent to the server to
download an item if its content has changed
since it was last placed in the cache.
The number of times requests are fulfilled from
the cache.
SilkPerformer User Guide 235

10 EXPLORING TEST RESULTS
Measurements
Measurement
HTTP cookies sent The number of cookies sent by browser-level
functions. Note that the header line of an HTTP
cookie request may contain multiple cookies;
each cookie in the header is counted.
HTTP cookies received The number of cookies received by browserlevel
functions. Note that an HTTP response
may include multiple cookie header lines; each
cookie of these header lines is counted.
HTTP pages
The number of HTML pages requested by the
page-level API or by the browser-level API
when Java script simulation is enabled.
Table 12: Web-specific measurements
IIOP-related measurement types
Measurement
Description
Description
IIOP messages
The number of IIOP messages sent by
SilkPerformer.
IIOP requests
The number of IIOP request messages sent by
SilkPerformer.
IIOP loc requests The number of IIOP locate request messages
sent by SilkPerformer.
IIOP fragments sent The number of fragmented IIOP messages
sent by SilkPerformer.
IIOP replies OK
The number of IIOP reply messages with
status OK received by SilkPerformer.
IIOP replies exception The number of IIOP reply messages with the
status "exception raised" received by
SilkPerformer. This count includes system
exceptions as well as user exceptions.
IIOP replies obj moved The number of IIOP reply messages with the
status "object has moved." This status
message causes SilkPerformer to reissue the
request to the new location.
Table 13: IIOP-related measurement types
236 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Measurement
IIOP request timed out
IIOP loc replies obj here
IIOP loc replies obj
unknown
IIOP loc replies obj moved
IIOP message errors
Database-specific measurements
The number of times the time limit for
receiving data (specified in the load testing
script) was exceeded while waiting for a
pending reply. Whenever the time limit is
exceeded, no reply is received and
SilkPerformer sends an IIOP "Cancel
Request" message to inform the server about
the aborted operation.
The number of IIOP "Locate Reply" messages
with status "object here" received by
SilkPerformer.
The number of IIOP "Locate Reply" messages
with status "object unknown" received by
SilkPerformer.
The number of IIOP "Locate Reply" messages
with status "object moved" received by
SilkPerformer.
The number of IIOP error messages received
by SilkPerformer.
IIOP fragments received The number of fragmented IIOP messages
received by SilkPerformer.
IIOP loc request timed out The number of times that the time limit for
receiving data (specified in the load testing
script) was exceeded while waiting for a
pending locate reply. Whenever the time limit
is exceeded, no locate reply is received and
SilkPerformer sends an IIOP "Cancel
Request" message to inform the server about
the aborted operation.
Table 13: IIOP-related measurement types
Measurement
Description
Description
SQL commands The number of SQL commands sent to the
DBMS for execution.
Table 14: Database-specific measurements
SilkPerformer User Guide 237

10 EXPLORING TEST RESULTS
Measurements
Measurement
Open database logins
Open database cursors
Description
Table 14: Database-specific measurements
Counts the currently open database sessions.
This is relevant for the OCI, ODBC, and
ODBC high level interfaces (This measurement
is not available in report files).
Counts the currently opened database cursors
(handles). This is relevant for the OCI, ODBC,
and ODBC high level interfaces. For ODBC,
only OdbcClose calls with the SQL_DROP
option will decrease this measure value. (This
measurement is not available in report files).
Individual
measurements
Transactions
measurement group
The SilkPerformer controller computer collects a subset of the summary
measurements at regular time intervals for all of the virtual users and agents
participating in a load test. It generates a time series data file (.tsd). After a load
test is complete, the data from the controller computer can be analyzed using
Performance Explorer.
The transactions measurement group contains response time data on the
transaction level. For each transaction defined in your load test script, a
measurement of this group type is automatically created, using the name of the
transaction as the key for the measurements. The measurement group contains
the following measurement types:
Measurement
Description
Trans. ok[s]
The response time for successful transactions,
in seconds. A transaction time is reported in
this measure type if no API function call
performed within the transaction returns an
error message.
Trans.(busy) ok[s] The busy time for successful transactions, in
seconds. A transaction busy time is the
transaction response time without any think
time. A transaction time is reported in this
measure type if no API function call performed
within the transaction returns an error message.
Trans. failed[s] The response time for failed transactions. A
transaction time is reported in this measure type
if at least one error of severity "error" or higher
occurs within the transaction.
Table 15: Transaction measurement types
238 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Measurement
Description
Trans.(busy) failed[s] The busy time for failed transactions. A
transaction busy time is the transaction
response time without any think time. A
transaction time is reported in this measure type
if at least one error of severity “error” or higher
occurs within the transaction.
Trans. canceled[s] The response time for canceled transactions. A
transaction time is reported in this measure type
if a return statement in the script is issued with
a return value of 1.
Table 15: Transaction measurement types
Web page
measurements
The Web page measurement group contains the measurement types related to
Web pages. For each Web page that is downloaded during a load test, a
measurement group will be created automatically with the name of the Web
page as the key. This measurement group contains the following types of
measurements.
Measurement
Page time[s]
Document download
time[s]
Server busy time[s]
Page data[kB]
Embedded objects
data[kB]
Description
Table 16: Web page measurement types
The time it takes to retrieve an entire Web page
from the server, that is, for retrieving all HTML
documents and all embedded documents, such
as images, videos, and audio files.
The time it takes to retrieve all HTML
documents of a Web page.
The time the server takes to process a Web page
request. This measurement starts after a virtual
user has sent the last byte of the request and
ends when the user retrieves the first byte of the
response.
The sum of data for all requests and responses
related to the request and retrieval of an entire
Web page including all embedded objects.
The sum of data for all requests and responses
related to the request and retrieval of objects
embedded on a Web page.
Web form
measurements
The Web form measurement group contains the measure types related to forms
declared in the dclform section of a load test script. The Web form measurement
SilkPerformer User Guide 239

10 EXPLORING TEST RESULTS
Measurements
group provides response time and throughput rates for form submissions with
the POST, GET, and HEAD command sent by scripting function calls. For each
form declared in the dclform section, a measurement group will be created
automatically with the name of the form as the key. The Web form
measurement group contains the following measurement types:
Measurement
Server busy time[s]
Round trip time[s]
Hits
Request data sent[KB]
Response data
received[KB]
Description
Table 17: Web form measurement types
Measurement starts after the user has sent the
complete request, and ends when the user
receives the first part of the response.
The time it takes to send a request and receive
the appropriate response. It also includes the
connect and shutdown time when the
connection is not persistent (the connection may
persist if the Keep-Alive option is enabled or
HTTP/1.1 is used).
The number of successful form submissions.
The amount of data (in KB) sent to the Web
server during submission of a form. This also
includes HTTP header and body content
information. The count specifies the number of
Web server requests sent.
The amount of data (in KB) received from the
Web server in response to the submission of a
form. This also includes HTTP response header
and body content information. The count
specifies the number of Web server responses
received.
WAP-specific
measurements
The WAP measurement group contains the measure types related to WAP
requests and WAP pushes. It contains the following measurement types:
Measurement
Description
WAP request
Measures the time from sending a URL request
to receiving the reply.
WAP send retries Counts the number of protocol-specific retries
to send a WAP package; when a request packet
is sent successfully, the virtual user will receive
a reply package from the gateway.
Table 18: WAP-specific measurement types
240 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
Measurement
Description
WAP pushes CL Counts the number of incoming connectionless
pushes.
WAP pushes CO Counts the number of incoming connectionoriented
pushes.
WAP pushes
Counts the number of incoming pushes
(connection-less and connection-oriented).
Table 18: WAP-specific measurement types
CORBA-specific
measurements
The IIOP measurement group contains the measurement types related to IIOP
operations that are called within a load testing script. It provides response time
and throughput ratings for the IIOP operation calls issued by the API function
IiopRequest. With the IiopObjectSetMeasure and the IiopObjectBind function
you can specify a prefix to the name of the measurement groups associated with
the operations of a certain CORBA object. If you decide not to do so, the name
of the operation is used as the name of the corresponding measurement group.
The IIOP group contains the following measurement types:
Measurement
Description
Round trip time[s] The time measured from when the processing
of the IiopRequest function call begins to when
the processing of the reply of the CORBA
object ends.
Server busy time[s] Measurement starts after a user has sent the last
byte of the CORBA operation call, and ends
when the user retrieves the first byte of the
response of the operation call from the CORBA
object.
Table 19: CORBA-specific measurement types
For non-blocking IIOP requests, these measurements are performed in the same
way. Note that this can lead to a situation where the sum of all your measured
times exceeds the total time of your load test.
If a CORBA object decides to redirect a request, the IiopRequest function has to
reissue this request to a different location. The result is two request/reply
message pairs, where both are measured separately. As in the case of the
measurement times for the IiopRequest function, the sum of both (separately
measured) request/reply pairs is supplied.
SilkPerformer User Guide 241

10 EXPLORING TEST RESULTS
Measurements
Database-specific
measurements
Middleware-specific
measurements
The SQL measurement group contains the following measurement types:
Measurement
Description
Parse[s]
The time it takes to prepare (Oracle
terminology: parse) a SQL command. Preparing
a SQL command usually includes a server round
trip during which the SQL command is sent to
the server. There the SQL command is parsed
and transformed into an internal format that can
be reused on subsequent calls of that SQL
command.
Note: If you use the OraParse function with the
ORA_DEFERRED option, the parsing of the
SQL command is deferred to the execute step,
and therefore the server round-trip time is
produced for the parse step. In this case, this
measurement type is meaningless.
Exec[s]
The time in which a SQL command is executed.
Executing a SQL command includes at least one
server round trip for sending the command and
returning the result to enable the SQL command
to be executed. Because of the considerable
amount of data processing required on the server
for a SQL command, executing SQL commands
usually generates a major part of the load on the
database server. There is a separate
measurement type in the SQL measurement
group for fetching data. Fetching data can either
be included as part of the execute step or in a
separate step. In the latter case, results are not
included in the SQL measurement group.
ExecDirect[s]
The time in which a SQL command is prepared
and executed.
Table 20: Database-specific measurement types
The TUXEDO measurement group contains the measurement types related to
TUXEDO service calls performed in a load testing script. It provides response
time and throughput ratings for the TUXEDO service calls issued by the Tux_
tpcall, Tux_tpacall, Tux_tpsend, Tux_tprecv, Tux_tpgetrply, Tux_tpenqueue,
and Tux_tpdequeue functions. For each TUXEDO service used in these
TUXEDO functions, a measurement group is created automatically with the
242 SilkPerformer User Guide

10 EXPLORING TEST RESULTS
Measurements
name of the service as the key. The TUXEDO group contains the following
measurement types:
Measurement
Response time[s]
Request data sent[KB]
Response data
received[KB]
Description
The response time for a TUXEDO service call.
For synchronous service calls, the response
time includes the time the server needs to
process the call. For asynchronous service
calls, SilkPerformer reports the time that
elapses between sending the request and
receiving the reply, plus the time spent waiting
for the request. Asynchronous service calls do
not measure the time a server needs to execute
the service. For send/request calls,
SilkPerformer measures the time for sending
data across an open connection as well as the
time for receiving data from an open
connection. For queuing communication, both
the time required for sending the service call to
the queue and the time for retrieving the results
of a service call are measured.
The amount of data (in KB) sent to the
TUXEDO server for submission of a given
service. SilkPerformer measures the
application data buffer size used by the service
calls rather than the transferred bytes on the
network.
The amount of data (in KB) received from the
TUXEDO server by a given service.
SilkPerformer measures the application data
buffer size used by the service calls rather than
the transferred bytes on the network.
Table 21: Middleware-specific measurement types
Citrix-specific
measurements
function
There is currently only one Citrix-specific measurement available. The Citrix
synchronization points measurement states how many synchronization
functions already have been executed. The main purpose of this measurement is
to visualize the progress of a transaction.
Measurement
Citrix Synchronization
points
Description
Table 22: Citrix-specific measurement types
The number of executed Citrix synchronization
functions (CitrixWaitForXXX).
SilkPerformer User Guide 243

10 EXPLORING TEST RESULTS
Measurements
Custom
measurements
Custom timer measurement group
The Custom Timer measurement group contains the timers that are defined in
the load testing script with the MeasureStart and the MeasureStop functions:
Measurement
Description
Response time[s] The time measured between MeasureStart and
MeasureStop function calls.
Table 23: Custom timer measurement types
Note: You can pause and resume timers with the MeasurePause and the
MeasureResume function.
Custom counter measurement group
The Custom Counter measurement group contains the counters that are defined
in the load testing script with the MeasureInc function:
Measurement
Description
Custom counter Counts values that are set with the MeasureInc
function.
Table 24: Custom counter measurement types
244 SilkPerformer User Guide

11
Additional 1Chapter
Workbench
Features
Introduction
What you will learn
This chapter explains some unique SilkPerformer Workbench features.
This chapter contains the following sections:
Section
Page
Overview 245
Teamwork Features 246
Code Completion 249
Parameter Wizard 252
Java Integration 256
Converting CSV Data for Use With Silk Performance Explorer 259
Overview
This section details the following Workbench features:
• “Teamwork Features”
• “Code Completion”
• “Parameter Wizard”
• “Java Integration”
• “Converting CSV Data for Use With Silk Performance Explorer”
SilkPerformer User Guide 245

11 ADDITIONAL WORKBENCH FEATURES
Teamwork Features
Teamwork Features
Importing/exporting
projects
Exporting projects
Importing projects
SilkPerformer Workbench enables you to export, import, email and upload
SilkPerformer projects.
SilkPerformer offers the option of exporting entire test projects to a separate
directory or an archive file. Such exported project files can subsequently be
imported into SilkPerformer.
Procedure To export a project:
1 From the SilkPerformer menu bar, select File/Export Project.
The Export Project dialog appears.
2 In the Export location area, specify the location to which you want to
export your current load testing project. If you want to export all the
project files to a directory, specify the directory name. If you want to
export all the files to a ZIP archive, specify the archive name.
Click the Browse (...) button to the right of the field to locate the
directory or file to which you want to export your load testing project.
3 Select the ZIP to single archive file option to export all files belonging to
your current load testing project to a ZIP archive. This is useful when
you want to backup your project or when you want to send a copy of
your project to someone.
4 In the Options area, select the Include results files of all tests option to
export the results of all tests belonging to this project (i.e., the results of
all try-script tests, find-baseline tests, and actual load tests). If you
disable this option, only the results of the recent try-script test will be
exported.
5 Select the Protect ZIP file with password option to protect the export file
(.ltz) with a password. This is useful when your project contains
confidential data (e.g., user names and passwords for the application you
are testing).
If you select this option, enter a password and confirm it.
6 Click OK.
SilkPerformer then exports the project and displays the progress of
packing all relevant files into a ZIP archive in a status bar.
Procedure To import a project:
1 From the SilkPerformer menu bar, select File/Import Project.
A standard Open dialog appears, enabling you to select the archive file
that you want to import.
2 After selecting the archive file, click OK.
246 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Teamwork Features
Uploading projects
Emailing projects
3 The Import Project dialog appears.
4 In the Import target directory area, specify the location to which you
want to extract your load testing project.
If you configured password protection you will be prompted to enter
your password. Enter your password and confirm it.
5 Click OK.
SilkPerformer then extracts the archive file to the specified target
directory and opens the load testing project.
Procedure To upload a project to SilkCentral Test Manager:
1 From the SilkPerformer menu bar, select File/Upload Project.
The Upload Project dialog appears.
2 In the Upload project name area, enter the name of the file in which the
load testing project should be uploaded.
3 In the Upload URL field, enter the URL to which you want to upload the
load testing project.
4 In the Upload area, select the Use secure connection option to use a
secure connection to upload your current load testing project. This is
useful when your project contains confidential data (e.g., user names and
passwords for the application under test.
5 Select the Remember login information option to have SilkPerformer
remember the user name and password that you enter when you upload a
load testing project for the first time. This is useful when you plan to
upload multiple projects.
6 Click the Internet Options button to define the Internet-related settings
for your computer and, most importantly, to determine how to connect
your computer to the Internet. In the ensuing Internet Properties dialog,
click the Connections tab and select the options that correspond to your
environment.
When you are done specifying Internet options, close the Internet
Options dialog.
7 In the Project Upload dialog, click OK.
SilkPerformer then uploads the project to the SilkCentral installation
URL specified in the Upload URL field, and displays the progress of
packing all relevant files into a ZIP archive in a status bar.
Procedure To email a project:
1 From the SilkPerformer menu bar, select File/Email Project.
The Email Project dialog appears.
SilkPerformer User Guide 247

11 ADDITIONAL WORKBENCH FEATURES
Teamwork Features
Enabling file locking
2 In the Email address area, enter the email address to which you want to
send your current load testing project.
3 In the Options area, select the Include results files of all tests option to
send the results of all tests belonging to this project along with the email,
that is, the results of all try-script tests, all find-baseline tests, and all
actual load tests. If you disable this option, only the results of the recent
try-script test will be sent.
4 Select the Protect ZIP file with password option to protect the contents of
the email with a password. This is useful when your project contains
confidential data (e.g., user names and passwords for the application you
are testing).
If you select this option, enter a password and confirm it. Also, do not
forget to tell the recipient of the email the password in a separate email or
phone call.
5 Select the Include a system diagnostic report option to send, along with
the email, a system diagnostic report. This information will list the
processors in the computer, the name and version of the operating
system, and give information regarding the drives, memory, services,
and drivers. In addition, relevant information from the Windows registry
will be included.
6 Click OK.
SilkPerformer then emails the project and displays the progress of
packing all relevant files into a ZIP archive in a status bar.
SilkPerformer’s document file locking mechanism allows multiple users (on
SilkPerformer Controller machines) to work simultaneously on the same
document files (bdf script files, bdh include files, data files) without overwriting
each other’s work. Documents that are locked (i.e., “checked out”) by another
user can be viewed in read-only format, but cannot be edited.
Procedure To the enable the document file locking mechanism:
1 From the SilkPerformer menu bar, select Settings/System.
2 The System Settings configuration dialog appears.
3 Click the Workspace tab.
4 In the Document file options area select the Enable document file locking
option.
5 Click OK.
248 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Code Completion
Code Completion
Code completion makes it easier to work with BDL. It significantly reduces
scripting errors and decreases the need to type text into your BDL files by
automatically completing functions, variables and constants. There are four
Code Completion options:
• Code List: Shows a list box with all currently matching function names.
• Code completion: Completes the current function, variable, or constant
name.
• Parameter info: Shows a tool tip with types and additional information
for function parameters.
• Quick info: Shows information about functions in a tool tip.
• Smart indention: Positions the cursor based on current logical context.
Code Completion works with both SilkPerformer-defined and user-defined
BDL files.
To enable Code Completion you must name and save new BDL files as either
.bdf or .bdh files. Once a BDL file has been saved, Code Completion recognizes
changes you make to it using the BDL Editor and includes all files you reference
via BDL use statements.
Note At all times, with all features of Code Completion, you are
only able to receive information regarding functions, variables and
constants that are defined in your local .bdf files, or in the included
.bdh files of those files.
Code list
One of the key features of Code Completion is its Code List mechanism, which
is used to offer you a list of functions, variables, and constants. The Code List
includes either:
- The entire list of available functions, variables, and constants
- Or, a list of functions, variables, and constants that match the
characters of the word currently selected in the script
The Code List can be manually invoked using Ctrl+Space. The Code List is
invoked automatically when you type one of the following API prefixes:
Attribute Com DB_ Dotnet Get Iiop
Java Jolt Measure Odbc Ora Pdce
Set Str Tux Wap Web Xml
Table 25: API prefixes for code list
SilkPerformer User Guide 249

11 ADDITIONAL WORKBENCH FEATURES
Code Completion
When the Code List is invoked, a list control with the following behavior
displays:
- If the cursor isn’t within a word or at the border of a word, the list
control is scrolled to the top, and the first item receives focus, but it
is not selected.
- If the cursor is within a word or at the border of a word, the list
control is scrolled to the first (selected) item that matches the word,
and the item is selected.
- By pressing the Tab, Space, or Enter key, the currently selected
string is inserted into the view. If a string is not selected, the string
with the focus is selected.
- Double-clicking an item in the list inserts the name into the view.
Code completion
Parameter info
Quick info
If a word that should be used with the Code List feature is unique, the Code List
won’t open when you invoke it via Ctrl+Space, however the full word will be
inserted.
Uniqueness in this context means that only one word in the list control matches
the partial word in question.
Parameter info is displayed when you place an opening bracket after a function
name in a script. Parameter info can also be invoked manually by pressing
Ctrl+I when the cursor is placed between the opening and closing brackets of a
function call. Parameter info is displayed as a tool tip that remains visible until
(1) the Esc key is hit, (2) a closing bracket is inserted, (3) the arrow keys are
used to move outside the brackets, or (4) input focus is set to another control.
Parameter info includes the following for functions:
- Function name
- Function parameters with names, types and modifiers (e.g., in, out,
and allownull)
- The current parameter is displayed with bold letters and, if available,
an additional description
- Return value type and, if available, a description
Quick info is similar to parameter info with the difference that it shows
information not only for functions, but also for variables and constants.
Quick info is presented in a form similar to a tool tip when you hold your cursor
over a function, variable or constant name. Quick info tool tips disappear when
you move your cursor away.
For functions, available information is the same as with parameter info, with the
exception that all parameters are shown with their descriptions rather than only
the current parameter.
250 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Code Completion
Smart indention
Enabling Code
Completion
User-defined
functions, variables,
and constants
For variables and constants, only name, data type and optional descriptions are
displayed.
Smart indention is enabled when the Enter key is pressed within a BDL related
view and a new line is to be inserted. Upon invocation, it places the cursor at a
position that is calculated with respect to the current context (e.g., indenting
after inserting an initial keyword).
Procedure To enable Code Completion:
1 From the SilkPerformer menu bar, select Settings/System.
The System Settings configuration dialog appears.
2 Click the Layout tab.
3 In the Editor area, select the Enable Code Completion option.
4 Click OK.
To receive Code Completion features for user-defined functions, variables, and
constants, you must create a new file with the same name—in the same
directory as your .bdf or .bdh file that defines the function, variable or constant
you want to enhance—though with the extension .bdd.
A .bdd file has the following structure:
FuncDesc
ParamName
ParamDesc
RetValTyp
RetValDesc
GlobalDesc
SilkPerformer User Guide 251

11 ADDITIONAL WORKBENCH FEATURES
Parameter Wizard
This means that for each function you want to instrument with Code Completion
information, you must insert a function node. For all others (variables and
constants) you must insert a global node.
FuncID
Name of the function in your .bdf or
.bdh file, for which Code
Completion information should be
provided.
FuncDesc
Textual description of the function.
ParamName
Parameter name
ParamDesc
Textual description of the
parameter.
RetValTyp
Data type of the return value.
RetValDesc
Textual description of the return
value.
GlobalId
Name of the global variable or
constant.
GlobalDesc
Textual description of the variable
or constant.
Table 26: Code completion nodes
Parameter Wizard
The Parameter Wizard enables realistic functionality to be created for virtual
users during load tests by replacing recorded or scripted values with constant
values, random values or values pulled from multi-column data files. Each of
these value types can be generated using the wizard. This means that you can
furnish virtual users with varied personal data for all the information commonly
required by Web forms (name, address, zip code, credit card number, etc.).
The Parameter Wizard guides you through the process of creating variables and
then places the variables in test scripts. A wide range of variable types are
provided. Each variable type can generate different value types (strings
containing text and/or numbers, serial numbers, whole numbers, and floatingpoint
numbers).
Values can be randomized using different distribution patterns. Comprehensive
sets of preset data (names, addresses, passwords, etc.) are also provided.
252 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Parameter Wizard
Constant values
Random variables
Procedure To insert a constant value into your script:
1 Click the Customize Test button in the SilkPerformer Workflow Bar.
The Workflow - Customize Test dialog appears.
2 Click the Create New Parameter Wizard button or link.
The Parameter Wizard launches.
OPTION 1: You can also access the wizard by selecting Script/Create
New Parameter in the SilkPerformer menu bar.
OPTION 2: The Parameter Wizard can also be accessed by moving your
cursor to a customizable value (e.g., a string), right-clicking, and
selecting Customize Value from the context menu
3 Select Constant value and click Next.
4 In the Name field, type a name for the new constant value or accept the
default name.
5 In the Value field, type the value of the new constant value.
6 Click Finish.
The new constant is added to the parameter section in you script.
If you opened the wizard via the context menu in your script, the value
you selected is automatically replaced with the newly created variable.
7 Launch the wizard as often as is required to create more constant values.
If you launched the wizard via another method, replace the constant
values in your script with the variables you created using
SilkPerformer’s BDL scripting language.
For help using Benchmark Description Language, refer to the BDL
Function Reference.
Procedure To insert a random value into a script:
1 Click the Customize Test button on the SilkPerformer Workflow bar.
The Workflow - Customize Test dialog appears.
2 Click the Create New Parameter Wizard button or link.
The Parameter Wizard appears.
OPTION 1: You can also access the wizard by selecting Script/Create
New Parameter from the SilkPerformer menu bar.
OPTION 2: The Parameter Wizard can also be accessed by moving your
cursor to a customizable value (e.g., a string), right-clicking, and
selecting Customize Value from the context menu.
3 Select Parameter from Random Variable and click Next.
SilkPerformer User Guide 253

11 ADDITIONAL WORKBENCH FEATURES
Parameter Wizard
Multi-column data files
4 In the wizard’s Choose the type of random variable window, choose a
random variable type from the Random type drop-down list
Use the type description and variable declaration preview provided for
each random variable type to choose the required variable and specify its
attributes.
5 Click Next.
The Name the variable and specify its attributes window appears
6 In the Name field, type a name for the new variable or accept the default
name
7 Specify the attributes of the variable. These will vary depending on the
type of variable selected. Use the Back button to review the type
description of the variable if necessary.
8 Click Finish.
The new random variable is then added to the random variables section
of your script.
If you opened the wizard via the context menu in your script, the value
you selected is automatically replaced with the newly created variable.
9 Launch the wizard as often as is required to create more constant values.
If you launched the wizard via another method, replace the constant
values in your script with the variables you created using
SilkPerformer’s BDL scripting language.
For help using Benchmark Description Language, refer to the BDL
Function Reference.
Procedure To insert a parameter with values from a multi-column data file:
1 Click the Customize Test button on the SilkPerformer Workflow bar.
The Workflow - Customize Test dialog appears.
2 Click the Create New Parameter Wizard button or link.
The Parameter Wizard appears.
OPTION 1: You can also access the wizard by selecting Script/Create
New Parameter from the SilkPerformer menu bar.
OPTION 2: The Parameter Wizard can also be accessed by moving your
cursor to a customizable value (e.g., a string), right-clicking, and
selecting Customize Value from the context menu.
3 Select Parameter from multi-column data file and click Next.
4 In the wizard’s Parameter from multi-column data file window, choose a
multi-column data file from the File Name drop-down list. Then select
an appropriate separator from the Separator drop-down list, or type a
custom separator.
In the Handle name field, type a name for the file handle of the multi-
254 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Parameter Wizard
column data file variable or accept the default name.
In the Parameter name field, type a name for the new parameter or
accept the default name.
5 Click Next.
The Name the variable and specify its attributes window appears.
6 In the Name field, type a name for the new variable or accept the default
name.
7 Specify the attributes of the variable. These will vary depending on the
type of variable selected. If necessary, use the Back button to review the
type description of the variable.
8 Click Finish.
The new parameter from a multi-column data file is then added to the
parameter section of your script.
The multi-column data file you selected is also added to the data files of
your SilkPerformer project.
If you opened the wizard via the context menu in your script, the value
you selected is automatically replaced with the newly created variable.
9 Launch the wizard as often as is required to create more constant values.
If you launched the wizard via another method, replace the constant
values in your script with the variables you created using
SilkPerformer’s BDL scripting language.
For help with Benchmark Description Language, refer to the BDL
Function Reference.
Editing Multi-Column Data Files
Editing multi-column
data files
Procedure To edit a multi-column data file:
1 Open the Data Files sub-tree of SilkPerformer’s Project tab.
2 Double-click the multi-column data file you want to edit.
The Edit CSV File window appears.
3 Choose an appropriate separator for the multi-column data file from the
Separator drop down list or type a custom separator.
4 Modify, add, or delete values from the multi-column data files.
5 Click the Save As button to save the multi-column data file with a
different name.
6 Click OK to save the changes of the multi-column data file and close the
window.
7 Click Cancel to ignore all changes and close the window.
SilkPerformer User Guide 255

11 ADDITIONAL WORKBENCH FEATURES
Java Integration
Existing parameters
Procedure To use an existing parameter for your script:
1 Move your cursor to a customizable value (e.g., a string) in your script.
2 Right-click and select Customize Value from the context menu.
The Parameter Wizard appears.
Note: The Customize Value menu item will be disabled if the value in
your script is not customizable.
3 In the wizard’s Customize Value window, select Use existing parameter.
4 Click Next.
The wizard’s Use Existing Parameter window appears.
5 Select a variable from the list and click Finish.
You can choose the data type of the shown parameters from the Show all
Parameters of type drop-down list.
Note: SilkPerformer automatically shows the window with a selected
data type that is compatible with the data type of the customized value.
6 The value you selected is then automatically replaced with the existing
parameter.
Java Integration
As a powerful extension to SilkPerformer’s Benchmark Description Language,
the Java Framework enables you to implement user behavior in Java. When
testing an existing Java application you don’t need to spend much time creating
test scripts—the only effort required is embedding existing Java source code
into the framework.
To generate a benchmark executable and run a load test with the help of the Java
framework, two source files are typically required:
- A Java class that implements the behavior of a virtual user
- A SilkPerformer test script that invokes Java method calls
Note Both source files need to have the same name, with extensions
indicating the file types (.bdf for the SilkPerformer test script and
.java for the Java source file).
With SilkPerformer it’s possible to call Java functions on any Java object from a
BDL context with any number and type of parameter and return parameter.
When a Java function returns a complex object, a handle for the Java object is
returned in its place. The handle can be passed to other function calls as an input
parameter, or public methods may be invoked on the object. There is also
support for static methods, cast operations, and exception handling.
256 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Java Integration
Creating Java projects
Adding Java files to
projects
Procedure To create a Java project:
1 Click the Start here button on the SilkPerformer Workflow bar or select
File/New Project from the main menu bar.
The Workflow - Outline Project dialog appears.
2 Enter a name for your project in the Project name field.
3 Enter a description for your project in the Project description field.
Note that this description is for your own project management purposes
only.
4 From the Application type tree list, choose Java/Java Framework.
5 Click OK.
An icon for the new project appears in the tree-view area of the main
SilkPerformer window.
6 If you need to add additional resources to the project, right-click the
project icon in the tree list. It is particularly important that all the user
data files (.csv), random data files (.rnd), and .idl files needed by
SilkPerformer are set up for the project.
Procedure To generate Java Framework files:
1 Click the Model Script button on the SilkPerformer Workflow bar.
The Workflow – Model Script dialog appears.
2 Select New script and click OK.
The New Java Framework Script dialog appears.
(Alternately you can select File/New.../Java Framework Scripts (.bdf/
.java) from the main menu bar to reach this step)
Enter a name for your Java class.
3 Click OK.
A Java file containing a class with the given name is generated and
opened along with a BDL file, which is used to load test the Java class.
4 If you save your files, you will be asked if you want to add them to your
project.
Procedure To generate Java files without a BDL file:
1 Select File/New.../Java Scripts (.java) from the main menu bar.
The New Java Framework Script dialog appears.
2 Enter a name for your Java class.
3 Click OK.
A Java file containing a class with the given name is generated and
opened.
SilkPerformer User Guide 257

11 ADDITIONAL WORKBENCH FEATURES
Java Integration
4 If you save your files, you will be asked if you want to add them to your
project.
You have the following options for adding a Java file to a project:
• When saving a new Java file for the first time, you are asked if you want
to add the file to the current project.
• Right-click the Scripts subnode of the project and select Add Script... to
open the Select Script(s) dialog. Change the File of type drop down list to
Java files (*.java). Select your file, which will be added to the project
when you click Open.
• Right-click in the view of a Java file to open the context menu. Click Add
to Project to add the file to the project.
See “JUnit Integration” for details regarding JUnit testing with SilkPerformer.
Compiling Java files
Before a Java file can be compiled it’s necessary to ensure that a valid path to a
Java SDK and a valid classpath are specified.
Procedure To specify Java settings:
1 Select Settings/Active Profile.
2 Scroll down to and select the Java icon in the left-hand scroll window.
3 The General tab appears. Select the Java version you are using from the
Version drop-down list.
4 Use the Browse (...) button to select the directory where the Java SDK is
installed.
5 Add any required entries to the classpath
Note Alternately, you can use the JavaSetOption API function with
the appropriate option number to set the Java version, Java home
directory, and classpath.
Manual compilation can be achieved in one of the following ways:
• Open a Java file and press F7.
• Open a Java file and right-click in the view to open a context menu.
Select the Compile command.
• Right-click a Java file in the Project tab to open a context menu. Select
the Compile command.
Initiating Try Scripts, finding baselines, and running tests will also
automatically launch compilation for all Java and BDL files.
258 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Converting CSV Data for Use With Silk Performance Explorer
During compilation, all output from the Java Compiler is redirected to
SilkPerformer’s Compiler tab. Resulting class files are generated in the
directory of the current project.
Note See the BDL Function Reference for details regarding Javaspecific
functions.
Converting CSV Data for Use With Silk Performance Explorer
Generating a CSV file
on a remote server
SilkPerformer offers a command line tool (csv2tsd.exe) that enables you to
convert CSV (Comma Separated Values) files into TSD (Time Series Data)
files.
This is useful when you are monitoring external Windows data that, for security
reasons, can not be monitored by SilkPerformer. By converting CSV data to
TSD format, external data can be analyzed alongside internal SilkPerformer
monitoring data via Silk Performance Explorer.
For example, if a performance counter is used to measure the behavior of a
remote server and the results of that monitor are saved in CSV format, those
results can be imported into Performance Explorer for analysis alongside the
results of a SilkPerformer load test involving the same server. With
Performance Explorer you then have the opportunity to look for correlations
between the two data sets.
Any CSV log file you specify in the command line tool must follow the format
generated by the Windows Performance Monitor tool (Adminstrative Tools/
Performance). So, before using csv2tsd.exe you must create a performance
counter with settings that will allow you to import its data.
Procedure To generate an eligible performance counter log prior to
conversion:
1 Open the Windows Performance Monitor tool (Adminstrative Tools/
Performance).
2 Expand the Performance Logs and Alerts tree node.
3 Right-click the Counter Logs node and select New Log Settings from the
context menu.
4 Enter a name for the new log settings file on the New Log Settings dialog
and click OK.
5 The properties dialog for the new log settings file is displayed. Click the
Add Counters button to open the Add Counters dialog.
6 Specify the computer, performance object, counter and counter instances
for which you want to track results.
SilkPerformer User Guide 259

11 ADDITIONAL WORKBENCH FEATURES
Converting CSV Data for Use With Silk Performance Explorer
7 Close the Add Counters dialog.
8 Select the Log Files tab on the Properties dialog.
9 From the Log file type list box, select Text File (Comma delimited).
10 Click OK to close the properties dialog.
11 Close the Performance Monitor.
Once you have configured a performance counter so that its log results can be
converted to TSD format using csv2tsd.exe, you can start capturing behavioral
data using the performance counter.
Procedure To begin capturing performance counter log data:
1 Open the Windows Performance Monitor tool (Adminstrative Tools/
Performance).
2 Expand the Performance Logs and Alerts tree node.
3 Double-click the Counter Logs node.
4 Right-click a counter log in the right-hand window and select Start from
the context menu.
Active counters are shown in green. Inactive counters are shown in red.
Generated CSV counter log files have the following format:
"(PDH-CSV 4.0) (W. Europe Standard Time)(-
60)","\\JONES\Processor(_Total)\% Interrupt
Time","\\JONES\Processor(_Total)\% Privileged
Time","\\JONES\Processor(_Total)\% Processor Time"
"02/25/2008 23:24:23.826"," "," "," "
"02/25/2008
23:24:38.841","0","0.10406093659265107","0.1015008710549
7141"
"02/25/2008
23:24:53.841","0.052084666700800877","1.8229633345280307
","2.4454192693999621"
"02/25/2008
23:25:08.840","0.052084666700800877","3.0729953353472514
","5.4663299380464103"
"02/25/2008
23:25:23.840","0.052084666700800877","0.5729313337088095
5","1.0391332684783428"
Procedure To convert a properly formatted CSV file to TSD format using
csv2tsd.exe:
1 Launch csv2tsd.exe (All Programs/Borland//Analysis
Tools/TSD File Export Tool)
260 SilkPerformer User Guide

11 ADDITIONAL WORKBENCH FEATURES
Converting CSV Data for Use With Silk Performance Explorer
2 Enter file import and export settings using the following format:
Csv2Tsd [[+/-]]
is the full path of the CSV File to import
is the full path of the TSD File to be generated
is the time difference in seconds. (default: 0)
3 Hit Enter to execute the conversion.
The example CSV counter log file above (which includes 20 rows) can be
converted to TSD format using the following command at the prompt in
csv2tsd.exe:
Csv2Tsd.exe C:\perflogs\Test1.csv c:\perflogs\Test1.tsd
SilkPerformer User Guide 261

11 ADDITIONAL WORKBENCH FEATURES
Converting CSV Data for Use With Silk Performance Explorer
262 SilkPerformer User Guide

12
SilkPerformer 12 Chapter
Plug-Ins
Introduction
What you will learn
This chapter explains SilkPerformer’s plug-in interface.
This chapter contains the following sections:
Section
Page
Overview 263
Checking and Initializing Plug-Ins 264
Message Output 264
Borland Optimizeit ServerTrace Plug-In 265
dynaTrace Diagnostics Plug-In 267
Overview
SilkPerformer offers a plug-in interface for SilkPerformer Workbench. The
plug-in interface extends SilkPerformer’s compatibility to third-party
technologies, enabling coordination and control of third-party technologies from
within the SilkPerformer environment.
Generic plug-in
interface controls
SilkPerformer's plug-in interface extends SilkPerformer’s compatibility with
third-party technologies, enabling coordination and control of third-party
technologies from within SilkPerformer.
The controls that are available for plug-ins vary based on the functionality
offered by each plug-in. By default, each plug-in offers the following attributes
via the SilkPerformer System Settings dialog:
On the General tab:
- Plug-in name, version, and description details
- Enabling/disabling of the plug-in
SilkPerformer User Guide 263

12 SILKPERFORMER PLUG-INS
Checking and Initializing Plug-Ins
- A list of features supported by the plug-in
On the Attributes tab:
- Name, description, datatype, and value for each available attribute
Depending on available plug-in functionality, other tabs and configurable
attributes may also be available.
Available plug-ins • “Borland Optimizeit ServerTrace Plug-In”
• “dynaTrace Diagnostics Plug-In”
Checking and Initializing Plug-Ins
All available plug-ins are installed to the following directory during
SilkPerformer setup: \Working\Plugins\...>.
When SilkPerformer Workbench is launched, it checks for installed plug-ins.
All installed plug-ins are then loaded, verified, and initialized.
Plug-ins can be enabled, disabled, or configured by users via SilkPerformer’s
System Settings dialog. The System Settings dialog adds a new icon in the lefthand
pane for each available plug-in. On the General tab, basic information
(including description, version, and supported features) and the ability to enable/
disable the plug-in is provided for each installed plug-in. Each plug-in is also
provided an Attributes tab that lists all attributes associated with the plug-in.
Some plug-ins may include additional tabs related to available functionality.
Plug-ins are disabled by default and must be enabled before use.
Message Output
A new tab is inserted into SilkPerformer’s Output window for each installed and
enabled plug-in.
SilkPerformer outputs all messages from plug-ins into these tabs. Output tabs
are removed when plug-ins are disabled.
Output tabs display the following information:
- Timestamp (HH:MM:SS)
- Severity (“Error,” “Warning,” or “Info”).
- Numerical message Code
- The text of the Message. Messages are displayed in the color that
corresponds to their severity (Error = red; Warning = orange; Info =
magenta).
264 SilkPerformer User Guide

12 SILKPERFORMER PLUG-INS
Borland Optimizeit ServerTrace Plug-In
Borland Optimizeit ServerTrace Plug-In
ServerTrace controls
System settings
The Borland ServerTrace Plug-In for Optimizeit ServerTrace makes use of
SilkPerformer’s plug-in interface. The Borland ServerTrace plug-in enables
SilkPerformer to initiate Borland ServerTrace snapshots within SilkPerformer
load test environments. Future versions of the plug-in may offer capturing and
analysis of ServerTrace results.
To install the Optimizeit ServerTrace Plug-In, download and execute the setup
package, which is available from http://sso.borland.com/download/. After
logging in, download the installer that corresponds to the latest version of
SilkPerformer you have purchased.
The Borland Optimizeit ServerTrace plug-in enables SilkPerformer to initiate
ServerTrace snapshots from within a SilkPerformer load test environment.
A ServerTrace snapshot offers a glimpse of the status of a server application. It
provides key performance metrics and statistics for standard J2EE components,
such as JDBC, JNDI, Servlet and JSP, EJB, JMS, and RMI.
During SilkPerformer load tests, the Optimizeit ServerTrace plug-in is in
contact with each configured ServerTrace agent and instructs it to start or stop
profiling or generate a snapshot.
For details regarding use of Borland Optimizeit ServerTrace, please see Borland
Optimizeit documentation.
Procedure To view and configure available system settings:
1 Open the System Settings dialog (Settings / System).
2 Select the ServerTrace icon in the left-hand pane.
3 The General tab appears. This tab displays read-only information about
the plug-in (name, version, and description). A read-only listbox displays
the features that are supported by the plug-in.
The General tab also enables you to enable/disable the plug-in (via the
Enable plug-in checkbox).
Note: By enabling/disabling the plug-in, you also enable/disable the
ServerTrace toolbar button and the ability to configure plug-in attributes
on the System Settings / Attributes tab.
4 Click OK to accept your settings, or select another tab to configure
additional plug-in settings.
Attributes Tab
Available plug-in attributes can be configured on the System Settings/Attributes
tab. Available attributes are listed in a grid control that includes Name, Type and
attribute Value. A read-only description of each attribute is also available in the
text field at the bottom of the tab.
SilkPerformer User Guide 265

12 SILKPERFORMER PLUG-INS
Borland Optimizeit ServerTrace Plug-In
Attribute values can be configured based on attribute datatype and available
values. These values are configurable only when the plug-in is enabled.
Click OK to accept your settings, or select another tab to configure additional
plug-in settings.
Agents Tab
On the Agents tab you can configure the available Borland ServerTrace agents
so that SilkPerformer can execute snapshots during load tests. Available agents
are listed in a grid control with Host name and Port text fields.
Agent Host and Port information must be entered manually.
Click OK to accept ServerTrace agent settings.
Using SilkPerformer to
control ServerTrace
ServerTrace Toolbar
The ServerTrace toolbar gives you the option of manually:
- Starting ServerTrace
- Hibernating ServerTrace
- Resetting ServerTrace
- Generating ServerTrace snapshots
Each action is executed on all configured ServerTrace agents.
ServerTrace snapshots
SilkPerformer serves as an interface for ServerTrace so that you can schedule
and execute ServerTrace snapshots from within SilkPerformer.
When a SilkPerformer load test is launched, stopped, or killed, the
SilkPerformer controller communicates its status to the ServerTrace Plug-In.
Depending on the plug-in settings, SilkPerformer triggers ServerTrace agents to
gather J2EE performance metrics and statistics.
For details regarding use of Borland Optimizeit ServerTrace, please see Borland
ServerTrace documentation.
Message output
A new tab is inserted into SilkPerformer’s Output window when the Borland
ServerTrace Plug-In is enabled.
The ServerTrace Output Window tab indicates such status information as when
contact is established with ServerTrace agents and when snapshots are initiated,
started, or failed.
266 SilkPerformer User Guide

12 SILKPERFORMER PLUG-INS
dynaTrace Diagnostics Plug-In
The ServerTrace Output Window tab displays the following ServerTrace
snapshot details:
• Timestamp (HH:MM:SS)
• Severity (“Error,” “Warning,” or “Info”).
• Numerical message Code
• The text of the Message. Messages are displayed in the color that
corresponds to their severity (Error = red; Warning = orange; Info =
magenta).
dynaTrace Diagnostics Plug-In
SilkPerformer offers a plug-in framework for dynaTrace Diagnostics and other
third-party diagnostics tools. The framework enables third-party tool vendors to
integrate into one or more of the following three Borland tools:
• SilkPerformer Workbench
For details on integration, see “dynaTrace Plug-In”.
• Silk TrueLog Explorer
The SilkPerformer Workbench plug-in can be used to modify profile
settings so that the runtime sends an additional HTTP header tag with
each page request. The same functionality is available via the
WebSetHttpTag BDL function.
When a user selects an API node, TrueLog Explorer passes the plug-in
header and body, including the additional HTTP tag. TrueLog Explorer’s
lower-right window includes an additional tab for each enabled plug-in.
The plug-in tabs are Internet Explorer controls, so plug-ins can present
anything that can be expressed via HTML.
Plug-ins can use the HTTP tag to identify traces and present complete
component break-downs for each Web page call.
• Silk Performance Explorer
Plug-ins that implement the PerfExp plug-in interface can be called from
the (OVR) Overview Report/Ranking section. Overview reports show plugin
related icons for each ranking measurement for which the plug-in can
provide in-depth information.
dynaTrace Plug-In
The dynaTrace Plug-In is installed along with the SilkPerformer setup. It is
disabled by default.
Note dynaTrace information is only produced when running a load
test. When running a TryScript test run, no dynaTrace information is
produced.
SilkPerformer User Guide 267

12 SILKPERFORMER PLUG-INS
dynaTrace Diagnostics Plug-In
Procedure To configure dynaTrace plug-in settings:
1 From SilkPerformer, select Settings/System Settings.
2 Select the the dynaTrace Diagnostics Plug-In group icon.
3 On the General tab, select the Enable plug-in check box.
4 Configure Name/Value pairs on the Attributes tab as required.
5 Click OK to save your settings.
268 SilkPerformer User Guide

13
Component 13 Chapter
Testing (Java &
.NET)
Introduction
What you will learn
This chapter explains now to manage component testing with SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 269
.NET Support 269
Java Support 271
Overview
With SilkPerformer .NET Explorer and Java Explorer, you can test Web
Services, .NET Remoting objects, Enterprise JavaBeans (EJB), RMI objects,
and other GUI-less .NET/Java objects.
.NET Support
SilkPerformer .NET Explorer, which was developed using .NET, and the
SilkPerformer .NET Framework allow you to test Web Services, .NET
Remoting objects, and other GUI-less .NET objects. .NET Explorer allows you
to define and execute complete test scenarios with different test cases without
requiring manual programming—everything is done visually via point and click
operations. Test scripts are visual and easy to understand—even for staff
members who aren't familiar with .NET programming languages.
SilkPerformer User Guide 269

13 COMPONENT TESTING (JAVA & .NET)
.NET Support
Launching .NET
Explorer
Test scenarios created with SilkPerformer .NET Explorer can be exported to
SilkPerformer Workbench for immediate reuse in concurrency and load testing;
and to Microsoft Visual Studio .NET for further customization.
Procedure To launch SilkPerformer .NET Explorer:
1 Go to .../Development Tools/SilkPerformer .NET
Explorer.
Alternately you can launch SilkPerformer Workbench. Go to
.../SilkPerformer Workbench.
Then create a new project with the application type .NET/.NET Explorer
or Web Services/.NET Explorer.
SilkEssential for .NET
Explorer
Add-In for Microsoft
Visual Studio .NET
SilkPerformer ships with a SilkEssential that allows you to hook .NET Explorer
to get a BDL script with tested WebServices.
Procedure To make use of the SilkEssential:
1 Create a new project based on the SilkEssential (Project Type:
WebServices/.NET Explorer).
2 Model the script.
3 While the recorder is running, start .NET Explorer from the Start menu.
4 Load a test scenario that you have defined in a previous .NET Explorer
session.
5 Execute an animated run of the scenario.
6 Close .NET Explorer and the recorder.
7 Perform a Try Script run of the recorded BDL script.
8 Perform XML customization/verification in the TrueLog as required.
See the BDL Function Reference for a description of DotNET framework.
The SilkPerformer Add-In for Microsoft Visual Studio .NET allows you to
implement test drivers in Visual Studio .NET that are compatible with
SilkPerformer Workbench. Such test drivers can be augmented with
SilkPerformer features that facilitate test organization, verification, performance
measurement, test data generation, and reporting.
Tests created with the Add-In can either be run within Visual Studio (with full
access to SilkPerformer's functionality) or within SilkPerformer Workbench (for
concurrency and load testing scenarios).
270 SilkPerformer User Guide

13 COMPONENT TESTING (JAVA & .NET)
Java Support
Procedure To launch SilkPerformer Add-In for Microsoft Visual Studio
.NET:
1 Go to: Start/Programs/Visual Studio .NET/Visual Studio .NET and create
a new SilkPerformer Visual Studio project.
Alternately you can launch SilkPerformer Workbench. Go to
.../SilkPerformer Workbench.
.NET Resources
Documentation
Create a new project using the .NET/.NET Framework using the Visual
Studio .Net Add-In application type.
Note The Add-In requires an installed version of either Visual
Studio .NET or Visual Studio .NET 2003/2005.
Visit the following links for information about .NET:
http://msdn.microsoft.com/net
http://www.gotdotnet.com
For full details regarding testing of .NET with SilkPerformer, please see the
.NET Explorer User Guide and the .NET Framework Developer Guide.
Java Support
Launching Java
Explorer
SilkPerformer Java Explorer, which was developed using Java, and the
SilkPerformer Java Framework allow you to test Web Services, Enterprise
JavaBeans (EJB), RMI objects, and other GUI-less Java objects. Java Explorer
allows you to define and execute complete test scenarios with multiple test cases
without requiring manual programming—everything can be done visually via
point and click operations. Test scripts are visual and easy to understand—even
for personnel who are not familiar with the Java programming language.
Test scenarios created with SilkPerformer Java Explorer can be exported to
SilkPerformer Workbench for immediate reuse in concurrency and load testing.
Note Java Explorer is compatible only with JDK versions 1.2 and
higher (v1.4 or higher recommended).
Procedure To launch SilkPerformer Java Explorer:
1 Go to: .../Development Tools/SilkPerformer Java
Explorer.
Alternately you can launch SilkPerformer Workbench. Go to:
.../SilkPerformer Workbench.
SilkPerformer User Guide 271

13 COMPONENT TESTING (JAVA & .NET)
Java Support
Create a new project using either the Java/Java Explorer or Web
Services/Java Explorer application type.
Java resources
Documentation
Visit the following links for information about Java:
http://java.sun.com
http://www.javaworld.com
http://www-106.ibm.com/developerworks/java/
For full details regarding testing of Java with SilkPerformer, please see the Java
Explorer User Guide and the Java Framework Developer Guide.
272 SilkPerformer User Guide

14
Unit 14 Chapter
Test Integration (JUnit
& NUnit)
Introduction
What you will learn
This chapter explains how to manage unit testing with SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 273
JUnit Integration 273
NUnit Integration 281
Overview
SilkPerformer enables you to integrate existing JUnit and NUnit test scripts into
SilkPerformer for the support of remote-component testing.
JUnit Integration
SilkPerformer facilitates smooth integration of existing JUnit test scripts
(version 3.8.x and 4.x) into SilkPerformer for the support of remote-component
testing under realistic concurrent-access server conditions.
SilkPerformer provides an import tool that enables you to import existing JUnit
test classes and other Java classes. Tested Java methods may have parameters
and return values; code for setting the in-parameters of these functions is
generated automatically. By definition, JUnit test methods can’t have
parameters.
SilkPerformer User Guide 273

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
Prerequisites
Setting up new JUnit
projects
Importing JUnit test
classes
Importing Java test
classes
Modeling Scripts
SilkPerformer’s Unit Test Import tool offers you the option of selecting specific
test methods. It automatically generates BDL stub code (a benchmark
description file) that calls those selected test methods. Existing JUnit test classes
can be called from SilkPerformer without requiring modification of the test
classes.
JUnit 4.x testing requires Java version 1.5 or higher.
The respective JUnit librarires must be available in the classpath of the Java
settings of the active profile.
Procedure To set up a JUnit SilkPerformer project:
1 Click the Outline Project button.
2 Enter a Project name and Project description for the project.
3 Select Unit Testing/JUnit from the Application type tree control on the
Outline Project dialog.
4 Click OK to create the project.
Procedure To Import a JUnit test class:
1 Select Project/Import Unit Test/JUnit Test Class to open the Import
JUnit Test Class dialog.
2 Follow the instructions outlined below in “Modeling Scripts”.
Note For JUnit4 module tests, JDK 1.5 or higher is required.
Modify your Java profile settings accordingly before importing
JUnit4 module tests.
Procedure To import a Java test class:
1 From the menu bar, select Project/Import Unit Test/Java Test Class to
open the Import Java Test Class dialog.
2 Follow the instructions outlined in the following section, Modeling
Scripts.
On the Import JUnit Test Class dialog, select an archive that is to be tested using
the File field; click the Browse (...) button to browse for the file. The archive is
automatically added to the profile classpath. The available classes are then
retrieved and displayed, sorted alphabetically in the Class field. Select one of the
available classes from the drop-down box for testing.
When you don’t specify a specific archive for testing, the Model Script dialog
enables you to specify a class that is available in the profile classpath. Simply
type the fully qualified class name into the Class field (e.g., “java.lang.String”).
Select the appropriate Constructor for instantiating the imported class. If a
constructor is not selected (entry “”), only static methods will be displayed in the method list.
274 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
Java profile settings
Select the Methods that are to be called. Note that when you enter a class name
in the Class field, the available constructors and methods are automatically
retrieved and displayed.
The JUnit or Java test class you want to import and all third-party libraries
needed by the test class must be available in the profile classpath. SilkPerformer
automatically adds all archives entered in the File field to the profile classpath.
Also, all test classes compiled in SilkPerformer Workbench are compiled to the
project directory, which is also automatically included in the profile classpath.
Click the Active Profile button (the checklist icon above the Methods field) to
display the Profile Settings dialog, open to the Java/General page for Java
projects (JUnit project type).
Note Changes made to these settings (e.g., Java Classpath) may lead
to different results. Selections made in the Class, Constructor, and
Methods fields will be updated with the new results.
Note For JUnit4 module tests, JDK 1.5 or higher is required.
Modify your Java profile settings accordingly before importing
JUnit4 module tests.
If you change the Java version, Java home directory, or JVM DLL, you must
restart SilkPerformer for the changes to take effect.
Click the Advanced Settings button (the funnel icon above the Methods field) to
filter the methods that are shown in the Methods field.
Parameterless Functions Only
Unit Test Functions Only
(selected by default for JUnit
import)
Member Functions Only
Public Functions (selected by
default)
Ignore all functions that expect
parameters
Ignore all functions that are not JUnit
conform
Ignore all non-member functions
Include public functions
Protected Functions
Include protected functions
Private Functions
Include functions with private access
Package Functions
Include functions with package access
Declared Functions Only Ignore functions from the base classes
Table 27: Filter options for JUnit methods
SilkPerformer User Guide 275

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
Complex Functions (selected by
default)
Autodetect Unit Test Functions
(selected by default)
Table 27: Filter options for JUnit methods
Show functions that take complex
parameters. In Java, complex
parameters are scripted by
JavaSetObject with NULL as the
default value.
Automatically detect and script JUnit
functions according to the following
rules:
• JUnit 3.8.x scripts are derived from
junit.framework.TestCase class.
• JUnit 3.8.x functions must not have
parameters or return values.
• setUp() and tearDown() are
invoked prior to and following each
JUnit 3.8.x method.
• JUnit 4.x scripts no longer require
to be derived from
junit.framework.TestCase class.
• JUnit 4.x functions are recognized
by the @Test annotation and added
to the script, unless the @Ignore
annotation is used.
• Methods having @BeforeClass and
@AfterClass annotation are
invoked once per test class.
• Methods having @Before and
@After annotation are invoked
prior to and following each JUnit
4.x method.
Generating a BDL
script for JUnit 3.8.x
Once you have customized filter settings, click OK to update the Methods field.
The following script is generated by importing a sample JUnit 3.8.x test case
and selecting the three methods testRound, testSqrt, and testMax:
transaction TInit
var
hPerf : number;
begin
JavaCreateJavaVM();
JavaSetString(JAVA_STATIC_METHOD, "TestCaseName");
ghTestObj := JavaLoadObject("JUTestClass",
"JUTestClass.");
end TInit;
276 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
transaction TMain
begin
JUnitCallMethod(ghTestObj, "testRound",
"testRound");
JUnitCallMethod(ghTestObj, "testSqrt", "testSqrt");
JUnitCallMethod(ghTestObj, "testMax", "testMax");
end TMain;
Generating a BDL
script for JUnit 4.x
transaction TEnd
begin
JavaFreeObject(ghTestObj);
end TEnd;
JavaLoadObject in the TInit transaction instantiates the JUnit test class
“JUTestClass”.
JUnitCallMethod in the TMain transaction calls one of the three test methods in
the same way that a JUnit test runner would. First the setUp() method is
invoked, then the test method itself (e.g., “testSqrt()”), finally the tearDown()
method is invoked.
Timers:
When an optional timer parameter is specified for a Java method call, the
execution time of the constructor, test method, setup method, and tear-down
method are measured. For the example above you would receive the following
measures:
For the constructor: JUTestClass.
For the methods: testRound, testSqrt, testMax
For the set up methods: testRound_setup, testSqrt_setup, testMax_setup
For the tear down methods: testRound_tearDown, testSqrt_tearDown, testMax_
tearDown
The following script is generated by importing a sample JUnit 4.x test case and
selecting the methods doFoo and doFoo2. The sample JUnit test script contains
several annotations for test setup and tear-down.
var
ghTestObj :number;
dcluser
user
JavaUser
transactions
TInit :begin;
TMain :1;
TEnd :end;
SilkPerformer User Guide 277

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
dclfunc
function JUnit4Before
begin
// @Before
JavaCallMethod(ghTestObj, "myBefore", "myBefore");
end JUnit4Before;
function JUnit4After
begin
// @After
JavaCallMethod(ghTestObj, "myAfter", "myAfter");
end JUnit4After;
function JUnit4CallFunc(hJavaObj :number; sName
:string; sTimerName :string optional; sExceptionBuffer
:string optional) : boolean
begin
JUnit4Before();
// @Test
JUnit4CallFunc := JavaCallMethod(hJavaObj, sName,
sTimerName);
if not StrIsNull(sExceptionBuffer) then
JavaGetLastException(hJavaObj, sExceptionBuffer);
end;
JUnit4After();
end JUnit4CallFunc;
dcltrans
transaction TInit
var
hPerf : number;
begin
JavaCreateJavaVM();
ghTestObj := JavaLoadObject("JU4ImporterTest",
"JU4ImporterTest.");
// @BeforeClass
JavaCallMethod(JAVA_STATIC_METHOD,
"JU4ImporterTest.myBeforeClass", "myBeforeClass");
end TInit;
transaction TMain
var
sBuffer :string;
begin
JUnit4CallFunc(ghTestObj, "doFoo", "doFoo");
JUnit4CallFunc(ghTestObj, "doFoo2", "doFoo2");
278 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
end TMain;
transaction TEnd
var
sBuffer :string;
begin
// @AfterClass
JavaCallMethod(JAVA_STATIC_METHOD,
"JU4ImporterTest.myAfterClass", "myAfterClass");
JUnit 4.x exception
handling
JavaFreeObject(ghTestObj);
end TEnd;
In the dclfunc section, the helper functions for test setup, tear-down, and
exception handling are defined. These functions will be used in the transactions.
JavaLoadObject in the TInit transaction instantiates the JUnit test class
“JU4ImporterTest”. All JUnit methods using the @BeforeClass annotation are
called in the TInit transaction.
JUnit4CallFunc in the TMain transaction calls all the test methods that were
selected for the JUnit test import. First the methods using the @Before
annotation are invoked, then the test method itself (e.g., “doFoo()”), finally the
methods using the @After annotation are invoked.
Timers:
When an optional timer parameter is specified for a Java method call, the
execution time of the constructor, test method, test setup, and test tear-down are
measured. For the example above you would receive the following measures:
For the constructor: JU4ImporterTest.
For the methods: doFoo, doFoo2
For the test set up: myBeforeClass, myBefore
For the test tear-down: myAfterClass, myAfter
JUnit 4.x allows the definition of methods that throw exceptions. Throwing such
exceptions is validated by JUnit. SilkPerformer imports JUnit methods that
throw exceptions including the validation of the thrown message.
//@Test(expected=java.lang.IllegalArgumentException)
JavaRegisterException(500,
"java.lang.IllegalArgumentException", JAVA_OPTION_MATCH_
SUBSTRING);
ErrorAdd(FACILITY_NATIVE_JAVA, 500, SEVERITY_
INFORMATIONAL);
JUnit4CallFunc(ghTestObj, "doFoo", "doFoo",
sBuffer);
ErrorRemove(FACILITY_NATIVE_JAVA, 500);
SilkPerformer User Guide 279

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
JUnit Integration
Importing JUnit test
suites
Executing JUnit test
suites
if (StrSearch(sBuffer,
"java.lang.IllegalArgumentException", STR_SEARCH_FIRST)
= 0) then
RepMessage("method did not throw:
java.lang.IllegalArgumentException", SEVERITY_ERROR);
end;
The sample above tests a JUnit method that is expected to throw a
java.lang.IllegalArgumentException. The function JavaRegisterException
assigns the BDL error code 500 to the exception. The function ErrorAdd sets the
error severity to informational, because the error is expected. After that, the
JUnit method doFoo is invoked, where the parameter sBuffer retrieves the error
message. The ErrorRemove function sets the error handling back to its original
state. It also verifies that the expected exception occurred, and throws a BDL
error if this is not the case.
Note When modifying the BDL script manually, make sure that you
do not map a Java exception to different BDL error numbers, or vice
versa.
Procedure To import a JUnit test suite:
Note For JUnit4 module tests, JDK 1.5 or higher is required.
Modify your Java profile settings accordingly before importing
JUnit4 module tests.
1 Proceed as described in “Importing JUnit test classes”
2 On the Import Unit Test - JUnit Test Class dialog, click the Advanced
Settings button (the funnel icon above the Methods field).
3 On the Advanced Settings dialog, make sure to apply the following
selections:
- "Unit Test Functions Only": Not selected
- "Declared Functions Only": Selected
- "Member Functions Only": Not selected
4 Load the JUnit test class as usual (via browsing for an archive or by
typing the class name).
5 Select the suite() method, then click OK.
If the JUnit Import Tool (Project -> Import Unit Test -> JUnit Test Class) scripts
code for invoking a method that has a return value, the code needed for
retrieving the return value is scripted but commented out.
If a Java method returns a junit.framework.Test method, additional commentedout
code is scripted, which allows to use the FRunTestSuite function from
junit.bdh to execute the test suite. junit.bdh is automatically added to the include
280 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
section of the BDL script. To execute the test suite, you need to un-comment the
respective calls.
transaction TMain
begin
// junit.framework.Test suite()
JavaCallMethod(JAVA_STATIC_METHOD,
"JU4ImporterTestSimple.suite", "suite");
// hObj := JavaGetObject(JAVA_STATIC_METHOD);
// use the following function to execute JUnit
testsuites
// FRunTestSuite(hObj);
// JavaFreeObject(hObj);
end TMain;
NUnit Integration
Setting up new
projects
SilkPerformer facilitates smooth integration of existing NUnit and Microsoft
Unit test scripts into SilkPerformer for the support of remote-component testing
under realistic concurrent-access server conditions.
SilkPerformer provides an import tool that enables you to import existing NUnit
assemblies and other .NET assemblies. Methods may have parameters and
return values; code for setting the in-parameters of these functions is generated
automatically.
SilkPerformer’s Unit Import Tool offers you the option of selecting specific test
case methods. It automatically generates BDL stub code (a benchmark
description file) that calls those selected test case methods. Existing NUnit test
classes can be called from SilkPerformer without requiring modification of the
test classes.
Note SilkPerformer’s NUnit import utility supports NUnit versions
2.1.4 and higher.
Procedure To set up an NUnit or .NET Testing SilkPerformer project:
1 Click the Outline Project button.
2 Enter a Project name and Project description for the project.
3 Select Unit Testing/NUnit or .NET Testing from the Application type tree
control on the Outline Project dialog.
4 Click OK to create the project.
SilkPerformer User Guide 281

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
5 Click the Model Script workflow button to customize the script (See the
following section, Importing NUnit/.Net Assemblies for details on
modeling NUnit test scripts).
Importing NUnit/.NET
assemblies
Modeling scripts
Procedure To Import an NUnit assembly:
1 Select Project/Import Unit Test/NUnit Test Class to open the Import
NUnit Test Class dialog.
2 Follow the instructions outlined below.
On the Import NUnit Test Class dialog, select a file that is to be tested using the
File field; click the Browse (...) button to browse for the file. The available
classes are then retrieved and displayed in the Class field.
When you don’t specify a specific file for testing, the Model Script dialog
enables you to specify a Class name from the available system classes (or
another file), select the Methods that are to be called, and define the appropriate
Constructor for instantiating the imported class. Note that when you enter a class
name in the Class field, the available constructors and methods are
automatically retrieved and displayed.
Click the Active Profile button (the checklist icon above the Methods field) to
display the Profile Settings dialog, open to the .NET/General page for .NET
projects (NUnit & .NET Testing project types). Changes made to these settings
may lead to different results. Selections made in the Class, Constructor, and
Methods fields are updated with the new results.
Click the Advanced Settings button (the checklist icon above the Methods field)
to filter the methods that are shown in the Methods field.
Parameterless Functions Only
Unit Test Functions Only (selected
by default)
Member Functions Only
Public Functions (selected by
default)
Ignore all functions that expect
parameters
Ignore all functions that are not
NUnit
Ignore all non-member functions
Include public functions
Protected Functions
Include protected functions
Private Functions
Include functions with private
access
Package Functions
Include functions with package
access
Table 28: Filter options for NUnit methods
282 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
Declared Functions Only
Complex Functions (selected by
default)
Autodetect Unit Test Functions
(selected by default)
Table 28: Filter options for NUnit methods
Ignore functions from the base
classes
Show functions that take complex
parameters. In Java, complex
parameters are scripted by
DotNetSetObject with NULL as the
default value.
Automatically detect and script
functions. Functions must not have
parameters or return values.
Methods marked with [SetUp] and
[TearDown] (for NUnit)
respectively [TestInitialize] and
[TestCleanup] (for Microsoft Unit
Tests) are invoked prior to and
following each method.
Generating NUnit BDL
scripts
Once you have customized filter settings, click OK to update the Methods field.
The following script is generated by importing NUnit Money and selecting the
three methods BagMultiply, BagNegate, and BagSimpleAhdd.
transaction TInit
var
sFileName : string;
begin
DotNetSetOption(DOTNET_OPT_REDIRECT_CONSOLE, 1);
// ==============================================
// Unit Test TestClass Information:
// Used Framework: NUnit Test Framework
// Initialize method: SetUp
// Class contains 21 test methods!
// ==============================================
GetDataFilePath("nunitmoneysample.dll", sFileName);
ghTestObj := DotNetUnitTestLoadObject(sFileName,
"NUnit.Samples.Money.MoneyTest",
"NUnit.Samples.Money.MoneyTest");
end TInit;
transaction TMain
begin
// BagMultiply
DotNetUnitTestCallMethod(ghTestObj, "BagMultiply",
"BagMultiply");
// BagNegate
SilkPerformer User Guide 283

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
DotNetUnitTestCallMethod(ghTestObj, "BagNegate",
"BagNegate");
// BagSimpleAdd
DotNetUnitTestCallMethod(ghTestObj, "BagSimpleAdd",
"BagSimpleAdd");
end TMain;
Timers
transaction TEnd
begin
DotNetUnitTestFreeObject(ghTestObj,
"NUnit.Samples.Money.MoneyTest");
end TEnd;
DotNetUnitTestLoadObject loads the NUnit assembly and creates an instance
of the test class. If the test class has a global initialize function implemented
(marked with TestFixtureSetup - ClassInitialize), this function will be called
right after the object is created.
DotNetUnitTestFreeObject in the TEnd transaction frees the reference to the test
object. If the test class has a global uninitialize function implemented (marked
with TestFixtureTearDown - ClassCleanup), this function will be called prior to
release of the object.
DotNetUnitTestCallMethod calls the TestMethod in the same way that the
NUnit Microsoft Unit Test Engine does. If there is a SetUp - TestInitialize
method implemented, it will be called before the test method is called.
If there is a TearDown - TestCleanup method implemented, it will be called
after the test method is called.
Test methods that expect an exception to be thrown (those marked with
ExpectedException) are only considered successful when an exception is
actually thrown.
When a test method writes information to the error or out console, the
information can be viewed in the TrueLog. A separate element is logged as child
node of the DotNetUnitTestCallMethod.
The same applies for unexpected exceptions. Stack trace and exception
messages are logged to the TrueLog.
A new optional timer parameter has been introduced for the original DotNet
API calls and the new DotNetUnitTest API calls. When this parameter is
specified, the execution times of the constructor, test method, set up method,
and tear down method are measured. For the example above you would receive
the following measures:
For the constructor: NUnit.Sample.Money.MoneyTest
For the methods: BagMultiply, BagNegate, BagSimpleAdd
284 SilkPerformer User Guide

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
For the setup methods: BagMultiply_Setup, BagNegate_Setup, BagSimpleAdd_
Setup
SilkPerformer User Guide 285

14 UNIT TEST INTEGRATION (JUNIT & NUNIT)
NUnit Integration
286 SilkPerformer User Guide

15
SilkCentral 15 Chapter
Test Manager
Integration
Introduction
What you will learn
This chapter explains SilkPerformer’s integration with SilkCentral Test
Manager.
This chapter contains the following sections:
Section
Page
Overview 287
Configuring Test Manager Integration 288
Opening SilkPerformer Projects from Test Manager 289
Importing SilkPerformer Projects from Test Manager 291
Uploading SilkPerformer Projects to Test Manager 292
Uploading Test Results to Test Manager 293
Downloading Test Results from Test Manager 294
Overview
SilkPerformer is fully integrated with SilkCentral Test Manager’s test planning
and test execution functionality. SilkPerformer projects can be integrated into
Test Manager test plans and directly executed via Test Manager. This allows for
powerful test-result analysis and reporting. It also enables “unattended” testing
(tests that are run automatically by Test Manager based on pre-configured
schedules).
SilkPerformer User Guide 287

15 SILKCENTRAL TEST MANAGER INTEGRATION
Configuring Test Manager Integration
Test Manager projects can be downloaded to SilkPerformer where scripts and
settings can be edited. Edited projects can subsequently be uploaded to Test
Manager to make them available for future test executions.
Note The term ‘Project’ is used differently in SilkPerformer than it
is in Test Manager. A SilkPerformer ‘project’, when uploaded to
Test Manager, becomes the core element of a Test Manager ‘test
definition’. Test Manager ‘projects’ are higher-level entities that
may include multiple test definitions, executions definitions, and
requirements.
Silk Performance Explorer can be used for in-depth analysis of test runs.
Performance Explorer can be launched directly from Test Manager’s
Executions unit via execution runs on the Runs tab (see the SilkCentral Test
Manager User Guide for details), from SilkPerformer, or from Performance
Explorer itself (see the Performance Explorer User Guide for details regarding
Performance Explorer’s integration with Test Manager).
The results of SilkPerformer load test runs can be uploaded to Test Manager and
associated with test definitions. To achieve this, SilkPerformer searches its
results directories; the files in the results directories are then uploaded to Test
Manager. See “Uploading Test Results to Test Manager” for details.
For additional information about Test Manager’s integration with
SilkPerformer, please consult the SilkCentral Test Manager User Guide and the
Silk Performance Explorer User Guide.
Configuring Test Manager Integration
Procedure To configure Test Manager access:
1 From the SilkPerformer menu bar, select Settings/System. The System
Settings - Workbench dialog appears open at the Agent Pool tab of the
Workbench group.
2 Select the Test Manager tab to access Test Manager configuration.
3 In the Hostname field, specify the host name or IP address of the host
where SilkCentral Test Manager’s front-end server is running.
4 In the Port field, specify the port number on which SilkCentral Test
Manager is listening. This is usually port 80 (if you are using an ISAPI
Web server) or port 19120 (if you are using Test Manager’s standalone
Web server). If you leave this field empty, SilkPerformer automatically
uses the default port number (80, 443 for secure connection).
5 Select the Secure checkbox to connect to Test Manager via SSL (Secure
Sockets Layer). The default SSL port is 443.
288 SilkPerformer User Guide

15 SILKCENTRAL TEST MANAGER INTEGRATION
Opening SilkPerformer Projects from Test Manager
6 Enter valid user credentials in the Username and Password fields.
7 Click the Test Connection button to test the connection to SilkCentral
Test Manager.
8 Click the Internet Options button to define the Internet-related settings
for your computer and, most importantly, to determine how to connect
your computer to the Internet. In the ensuing Internet Properties dialog,
click the Connections tab and select the options that correspond to your
environment.
Opening SilkPerformer Projects from Test Manager
Procedure To open a SilkPerformer project from Test Manager:
1 From the File menu, select the Open Project from Test Manager
command. You will then be presented with the connection settings you
entered on the System Settings - Workbench - Test Manager tab.
2 (Optional) If you don't want to use the default connection settings
specify:
- Hostname of your Test Manager installation and the Port where Test
Manager is listening (values of ‘0’ or ‘omitted’ indicate the default
port).
- Secure connection option to connect to Test Manager via SSL
(Secure Sockets Layer, HTTPS).
- Valid Username and Password to connect to Test Manager.
3 (Optional) Click the Set as Default button to define the connection
settings you have entered as the default settings for future connections.
4 (Optional) Click the Internet Options button to define Internet
connection properties (e.g., proxy settings, security settings, browser
connection settings). Click OK on the Internet Properties dialog to save
the settings.
5 Click Next on the Open Project from Test Manager dialog to see a list of
the Test Manager projects to which the active user has read access (this is
the same list that is available in Test Manager’s Projects unit).
6 Select the Test Manager project you want to open and click Next (or
simply double-click the project you want to open).
7 In the project tree, open test containers and/or test folders as required and
select the SilkPerformer test definition that you want to open in
SilkPerformer.
SilkPerformer User Guide 289

15 SILKCENTRAL TEST MANAGER INTEGRATION
Opening SilkPerformer Projects from Test Manager
Editing SilkPerformer
projects from Test
Manager
8 If the SilkPerformer test definition is located on an UNC path, click
Finish. Otherwise click Next. This brings up the Target Directory
Selection dialog where you can specify the location where the project is
to be copied. If you are using source-control integration for Test
Manager (e.g., Visual SourceSafe) you will be presented with a login
screen for your source-control client. Enter valid user connection settings
and click OK to continue.
Note SilkPerformer projects can also be opened directly from Test
Manager via the Open Project icon on the Test Manager Properties
tab in the Test Plan unit. See the SilkCentral Test Manager User
Guide for details.
Note See the Settings chapter of the SilkCentral Test Manager User
Guide for information regarding the configuration of source-control
integration with Test Manager.
Procedure To edit a SilkPerformer project from Test Manager:
1 Once you have downloaded a Test Manager project, the project appears
in SilkPerformer, where you can edit the project before saving it back to
Test Manager.
Note If you have configured source-control integration for your
Test Manager installation, SilkPerformer will have automatically
checked the SilkPerformer project file out of the source-control
database (SilkPerformer projects that have been checked out via Test
Manager source-control integration are indicated with red
checkmark symbols in the SilkPerformer tree view). This integration
is facilitated by SilkPerformer using Web Services to connect to Test
Manager and request the source control link information (connection
string, username, and .ltp project file name).
2 Once you have completed and saved all required changes to the
SilkPerformer project file, select the Close Project command from the
File menu and, on the subsequent dialog, confirm that you want to check
the edited file back into the source-control database by clicking Yes
(alternately, you can select the Check In (and Close Project) command
from the project tree context menu).
All subsequent executions of this project via Test Manager will use this updated
project file.
290 SilkPerformer User Guide

15 SILKCENTRAL TEST MANAGER INTEGRATION
Importing SilkPerformer Projects from Test Manager
Importing SilkPerformer Projects from Test Manager
Whereas opening a Test Manager project (as described in “Opening
SilkPerformer Projects from Test Manager”) involves “checking out” a Test
Manager project from a source-control tool (or retrieving a project from a UNC
path), editing the project, and checking the project back in, importing a project
involves only downloading a copy of the project and working with it
independent of Test Manager. Any changes you make to an imported project
will have no effect on Test Manager. All you need to import a Test Manager
project is a normal Web connection (no source-control tool connection or
credentials are required).
Importing a file into SilkPerformer enables you to perform “attended” test runs
independently of Test Manager. This allows you to execute test runs manually in
SilkPerformer. This differs from tests that are run automatically based on
predefined schedules in Test Manager. Test results can subsequently be
uploaded back to Test Manager and associated with a test definition.
Even if you have configured source code control integration with Test Manager,
imported project files will not be checked out of your source control tool
because ‘imported’ project files are saved locally to your hard disk as
compressed (WinZip) archives. Attended tests can also be initiated from the Test
Manager side (see the SilkCentral Test Manager User Guide for details).
Procedure To import a SilkPerformer project from Test Manager:
1 From the File menu, select the Import Project from Test Manager
command. You will then be presented with the connection settings you
entered on the System Settings - Workbench - Test Manager tab.
2 (Optional) If you don't want to use the default connection settings
specify:
- Host name or IP address of your Test Manager installation and the
Port where Test Manager is listening (values of ‘0’ or ‘omitted’
indicate the default port).
- Secure connection option to connect to Test Manager via SSL
(Secure Sockets Layer, HTTPS).
- Username and Password of your connection.
3 (Optional) Click the Set as Default button to define the connection
settings you have entered as the default settings for future connections.
4 (Optional) Click the Internet Options button to define Internet
connection properties (e.g., proxy settings, security settings, browser
connection settings). Click OK on the Internet Properties dialog to save
the settings.
SilkPerformer User Guide 291

15 SILKCENTRAL TEST MANAGER INTEGRATION
Uploading SilkPerformer Projects to Test Manager
5 Click Next on the Import Project from Test Manager dialog to see a list
of the Test Manager projects to which the active user has read access
(this is the same list that is available in Test Manager’s Projects unit).
6 Select the Test Manager project you want to import and click Next (or
simply double-click the project you want to import).
7 In the project tree, open test containers and/or test folders as required and
select the SilkPerformer test definition that you want to import to
SilkPerformer.
8 If you don't want to use the default target directory, enter a directory of
your choice into the Target Directory field.
9 Click Finish.
Uploading SilkPerformer Projects to Test Manager
Uploading
SilkPerformer projects
to Test Manager
Before uploading a project to Test Manager, ensure that you have properly
configured the project’s workload via SilkPerformer. Workload settings are
subsequently specified along with uploaded projects in the SilkPerformer Test
Properties portion of the Properties tab in Test Manager’s Test Plan unit.
Procedure To upload a SilkPerformer project to Test Manager:
1 With the project that you want to upload opened in SilkPerformer, select
Upload Project to Test Manager from the File menu to open the Upload
Project to Test Manager wizard.
2 If the project has never been uploaded to Test Manager before, the Test
Manager connection settings that you configured will be displayed (see
“Configuring Test Manager Integration” for details on configuring these
settings). Click Next. Please proceed with step 3.
If the project has been uploaded to Test Manager before, the association
with Test Manager’s test definition is already known and pre-selected.
Please proceed with step 4.
3 From the project list, select the SilkCentral Test Manager project to
which you want to upload the SilkPerformer project. Click Next.
4 You can select an existing test definition that is to be overwritten or you
can create a new test definition by right-clicking the folder or container
where the test definition is to reside and selecting New Child Test
Definition from the context menu.
You can also create a new folder within an existing test container by
right-clicking a container and selecting New Child Test Folder from the
context menu.
292 SilkPerformer User Guide

15 SILKCENTRAL TEST MANAGER INTEGRATION
Uploading Test Results to Test Manager
Note Newly created test definitions and folders are displayed with
bold text to indicate that they have not yet been written to the Test
Manager database. If you click Cancel the items you have created
will not be saved.
5 If you do not want to upload test run results now, do not enable the
Enable Results Upload checkbox and proceed with step 8.
Enable the Enable Results Upload checkbox to select specific test run
results that you want to upload with the project. Click Next.
6 Check the appropriate checkboxes in the Results list to select the results
that you want to upload along with the project. Click Next.
7 Specify Version and Build number for the assigned product to which the
uploaded results belong.
8 Click Finish to upload the project to Test Manager. The newly uploaded
items now appear in Test Manager’s test plan/project tree.
Uploading Test Results to Test Manager
Procedure To upload test results to Test Manager:
1 Select the Upload Results to Test Manager command from the Results
menu.
Alternative Right-click a results node in the Results list of
SilkPerformer and select Upload Results to Test Manager from the
context menu.
2 If the project has never been uploaded to Test Manager before, the Test
Manager connection settings that you configured will be displayed (see
“Configuring Test Manager Integration” for details on configuring these
settings). Click Next. Please proceed with step 3.
If the project has been uploaded to Test Manager before, the association
with Test Manager’s test definition is already known and pre-selected.
Please proceed with step 5.
3 From the project list, select the SilkCentral Test Manager project to
which you want to upload the SilkPerformer results. Click Next.
4 From the tree list, select the test definition to which you want to upload
the results. Click Next.
SilkPerformer User Guide 293

15 SILKCENTRAL TEST MANAGER INTEGRATION
Downloading Test Results from Test Manager
Note Alternately you can right-click in the tree and use the
commands on the context menu to create a new test definition, child
test definition, test folder, and/or child test folder to which the results
are to be saved.
5 On the following dialog you can specify Version and Build number for
the assigned product to which the uploaded results belong. Also specify
the SilkPerformer Test result status (Passed, Failed).
6 Click Finish to upload the results.
Downloading Test Results from Test Manager
Procedure To download test results from Test Manager:
1 Select SilkPerformer’s Results tab in the top left area of the GUI.
2 Click the Click here to add Test Manager results node in the results tree.
3 The Open Results from Test Manager wizard opens. The corresponding
Test Manager’s test definition is pre-selected. You can select multiple
executions and download the results of multiple test executions.
Note To download results from another test defintion, another
project, or another Test Manager server, click the Back button to
navigate back though Test Manager’s tree view.
4 Click Finish to download the results.
294 SilkPerformer User Guide

16
Source 16 Chapter
Control Integration
Introduction
What you will learn
This chapter explains SilkPerformer’s integration with source control systems.
This chapter contains the following sections:
Section
Page
Overview 295
Configuring Source Control Integration 296
Placing Projects and Files Under Source Control 298
Checking Files In 300
Checking Files Out 301
Getting the Latest File Versions 302
Checking In/Out Project Files 302
Removing Source Control Linkage 303
Synchronizing Source Control Status 303
Known Issues and Workarounds 304
Overview
SilkPerformer's source control integration (SCC integration, also referred to as
source code control integration) enables you and other members of your
organization to manage and share your SilkPerformer project files, test scripts,
custom include files, and data files in a common repository (i.e., source control
system)—enabling file check-in, check-out, get-latest-version, and other
functions typical of source control systems.
SilkPerformer’s source control integration is facilitated by commands on the
SilkPerformer File menu (File/Source Control) and context menus in
SilkPerformer User Guide 295

16 SOURCE CONTROL INTEGRATION
Configuring Source Control Integration
SilkPerformer’s project tree. All project nodes, script nodes, custom include file
nodes, and data file nodes in the project tree offer SCC commands.
In addition to offering commands that select all the files of complete projects
(i.e., project file, custom include files, data files, and scripts), a sub-set of SCC
commands that select only the project file are also available (i.e., you can easily
check-in/check-out project files without overwriting associated script and data
files). See “Checking In/Out Project Files” for details.
Status icons in SilkPerformer’s project tree indicate the status of files under
source control:
- Files that have been placed under source control and are currently
checked out (i.e., the source control version of the file is writable)
are indicated with red checkmark icons.
- Files that have been placed under source control, but are currently
NOT checked out (i.e., the source control version of the file is readonly)
are indicated with blue padlock icons.
Note SilkPerformer supports StarTeam, Microsoft Visual
SourceSafe, and PVCS. Source control integration must be enabled
in SilkPerformer Workbench system settings (Settings/System/
Source Control). You must have StarTeam, PVCS, or Visual
SourceSafe installed on your system before source control
integration can be enabled.
Configuring Source Control Integration
For configuring MS VSS or PVCS SCC integration with SilkPerformer,
proceed with “General SCC integration configuration”.
For configuring StarTeam SCC integration with SilkPerformer, proceed with
“Integrating StarTeam SCC with SilkPerformer”.
Integrating StarTeam
SCC with
SilkPerformer
The integration consists of downloading the integration module and running the
installation wizard. The following subtasks detail this workflow:
• “To download the StarTeam SCC Integration Module:”
• “To run the StarTeam Microsoft SCC Integration Setup wizard:”
Procedure To download the StarTeam SCC Integration Module:
1 Navigate to the StarTeam download area at http://www.borland.com/
downloads/download_starteam.html and select StarTeam 2005 & 2005
Rel 2. Integrations. You may need to login or verify your email address.
296 SilkPerformer User Guide

16 SOURCE CONTROL INTEGRATION
Configuring Source Control Integration
2 Scroll to the Others table at the bottom of the page. In the SCC row,
choose HTTP or FTP to download the integration module.
3 Click Save to save the integration module to the default location.
4 The Borland Download Manager downloads the file to your local
computer.
General SCC
integration
configuration
Procedure To run the StarTeam Microsoft SCC Integration Setup wizard:
1 Click Launch on the Download Manager to save files and launch the
installation wizard.
2 Enter the name of the folder to save the temporary installation file to and
click Next.
3 The StarTeam Microsoft SCC Integration Setup wizard is displayed.
4 Click Next and then click Yes to accept the license agreement. Enter your
user name and company and click Next.
5 The StarTeam SDK runtime is installed if needed.
6 On the Choose Destination Location page of the wizard, choose the
folder to install the StarTeam SCC Integration Module to and click Next.
7 On the Select Components page of the wizard, choose SCC API
Integration and click Next.
8 Click Yes to make StarTeam the default SCC Provider.
9 Click Next to accept the default program icon setup.
10 Click Finish on the Setup Complete page of the wizard.
11 Click Exit to close the Download Manager.
Important: It is critical that you set up and manage your projects using the
StarTeam client. This includes creating projects, adding files to a project, and
removing files from a project.
Procedure To configure source code control integration in SilkPerformer:
1 With StarTeam, MS VSS, or PVCS installed on your local system, select
System from the Settings menu.
2 Select the Workbench group icon.
3 Select the Source Control tab (this tab is farthest to the right of the tab
list; use the small right-pointing arrow to access this tab).
4 Select the Enable Source Control Integration checkbox. All SCC
providers that are installed on your local machine and supported by
SilkPerformer (i.e., StarTeam, VSS, or PVCS) are listed in the SCC
Provider list box.
SilkPerformer User Guide 297

16 SOURCE CONTROL INTEGRATION
Placing Projects and Files Under Source Control
5 Enter a username in the Username field. The username you enter here
will be suggested automatically as the default username value in future
operations, such as login dialogs.
Note Custom include and user data files are configured
independently from project and system files for the SCC because
SilkPerformer may require that these files be stored separately in the
SCC (e.g., an organization might have a common repository for
include files that are shared amongst a group of developers. These
developers might also be required to check in files from their
individual projects).
6 In the Custom user data files SCC path field, click the Browse (…) button
and select the location in the SCC database where the custom user data
files are to be saved.
Note At this point you will be required to log into your SCC system
(see StarTeam, MS VSS, and PVCS documentation for full details
regarding the use of your source control system). The username that
you entered above will be entered into the login dialog
automatically. Your SCC system will also likely require that you
enter the password that corresponds to the submitted username. If
not already selected, you may need to select the target SCC database.
Following that, browse the file structure of your SCC system and
select the directory to which your custom user data files are to be
saved (you may also create a new directory for this purpose).
7 In the Custom include files SCC path field, click the browse (…) button
and select the location in the SCC database where the custom include
files are to be saved (see the note above for assistance with this step).
Note The SCC database path field is automatically populated with
the database-path root of the source control system you specify in
the two above fields.
8 Click OK to exit the System Settings dialog. Source control integration is
now complete. SilkPerformer’s source control functions are now
available from the File menu (File/Source Control).
Placing Projects and Files Under Source Control
Note System include files can not be placed under source control.
Only custom include files can be placed under source control.
298 SilkPerformer User Guide

16 SOURCE CONTROL INTEGRATION
Placing Projects and Files Under Source Control
Placing individual files
under source control
Procedure To place a complete project under source control:
1 With a project loaded into SilkPerformer, select the project node in the
project tree.
2 Select the Add to Source Control command from the File menu
(alternately you can right-click the project node and select Add to Source
Control from the context menu).
3 In the subsequent SCC dialog, all scripts, include files, and data files
associated with the project are automatically selected for placement
under source code control. Uncheck any file type selections for file types
that should not be placed under source control.
Note Ensure that source control paths have been configured for each
project file type. As you have already configured source control
paths for custom include files and data files in SilkPerformer’s
system settings, you should now only have to specify source control
paths for project files and script files (e.g., For VSS, a ‘Click here to
specify link’ appears in the Source Control Project column of the
Add to Source Control dialog for any project file type for which a
source control path has not been specified). If required, log into your
SCC and specify directories for your project and script files.
4 Click OK to place the project’s files under source control
Procedure To place individual files under source control:
Note Project files (.ltp), script files (.bdh), custom include files, and
data files can be independently placed under source code control.
See above for information on placing complete project files (.ltp)
under source code control.
1 To place a file that is currently only available on your local disk under
source control in the specified SCC file location (thereby configuring a
connection between your local disk and the SCC), right-click a file in
SilkPerformer’s project tree and select Add to Source Control from the
context menu.
Note Alternately you can select the Scripts, Data Files, or Include
Files nodes to select all files of those types for placement under
source control.
2 The Add to Source Control dialog appears, showing the files upon which
the operation is to be performed.
Note The remainder of this process is handled via dialogs provided
by your SCC provider (StarTeam, PVCS, or VSS). Please see your
SCC documentation for full details.
SilkPerformer User Guide 299

16 SOURCE CONTROL INTEGRATION
Checking Files In
3 Using the checkboxes on the Add to Source Control dialog, select/
deselect files for placement into source control.
Note If not already defined, a source control project path must be
defined for each file that is to be placed under source control. If a file
marked for placement under source control does not yet have a
source control project path configured for it, you will be prompted to
configure a path (the source control project of the SilkPerformer
project file (*.ltp) is suggested by default).
Note Files added to source control are written to the corresponding
*.ltp file. Therefore, to add files to source control, the corresponding
*.ltp file must be checked out.
4 Click Add to place the file under source control.
Note A blue padlock appears next to files in the project tree that are
under source control and are currently checked in (read-only). This
is the default status of files that are newly placed under source
control.
Checking Files In
These instructions apply to the checking in of individual scripts, include files,
data files, and entire projects.
Procedure To check a file into source control:
1 In the project tree, right-click a file that is currently checked out
(indicated by a red checkmark icon) and select Check In from the context
menu.
2 The Check In dialog appears, showing the local path where the file is
currently saved. Click the Check In button to check in the file. The status
icon of the file now appears as a blue padlock, indicating that the source
control version of the file is now read-only.
Note To check in an entire project, right click the project node and
select Check In from the context menu.
Note Right-click the Scripts, Data Files, or Include Files nodes to
select all files of those types for check-in. If required, on the
subsequent Check In dialog, individual files of that type can be
deselected.
300 SilkPerformer User Guide

16 SOURCE CONTROL INTEGRATION
Checking Files Out
Checking Files Out
These instructions apply to the checking out of individual scripts, custom
include files, data files, and entire projects.
Procedure To check a file out of source control:
1 Right-click a file in the project tree that is currently checked in (indicated
by a blue padlock icon) and select Check Out from the context menu.
2 The Check Out dialog appears, showing the local path to where the file
will be copied.
Note If a single checked-in file is selected for check out, then only
that file will be listed on the Check Out dialog. If no file or a
checked-out file is selected for check out, then all currently checkedin
files of the project will be listed on the Check Out dialog.
Undoing file check out
Click the Check Out button to check out the file. The status icon of
the file now appears as a red checkmark, indicating that the source
control version of the file is available and writable.
Note To check out an entire project, right click the project node and
select Check Out from the context menu.
Note Right-click the Scripts, Data Files, or Include Files nodes to
select all files of those types for check out. If required, on the
subsequent Check Out dialog, individual files of that type can be
deselected.
Procedure To undo file check out:
Note ‘Undoing’ a file check-out replaces the file on your local disk
(regardless of whether or not it includes recent changes) with the
most recently checked-in version of the file in the SCC. The local
version of the file is then read-only.
1 In the project tree, right-click a file that is checked out (indicated by a red
checkmark icon) and select Undo Checkout from the context menu.
2 The Undo Checkout dialog appears, showing the local path to where the
file is located. Click the Undo Checkout button to undo the check out.
The status icon of the file now appears as a blue padlock, indicating that
the source control version of the file is read-only.
Note To undo checkout of an entire project, right click the project
node and select Undo Checkout from the context menu.
SilkPerformer User Guide 301

16 SOURCE CONTROL INTEGRATION
Getting the Latest File Versions
Note Right-click the Scripts, Data Files, or Include Files nodes to
select all files of those types at once. If required, on the subsequent
Undo Checkout dialog, individual files of that type can be
deselected.
Getting the Latest File Versions
A source control system ‘Get Latest Version’ function downloads the latest
version of a file to your local disk (by default, in read-only format) without
actually checking the file out.
Note See StarTeam, MS VSS, or PVCS documentation for full
details regarding use of your SCC’s ‘Get Latest Version’ function.
These instructions apply to the Get Latest Version function for individual
scripts, include files, data files, and entire projects.
Procedure To get the latest version of a file from your source control
system:
1 Right-click a file in the project tree and select Get Latest Version from
the context menu.
2 The Get Latest Version dialog appears, showing the local path where the
file will be copied. Click the Get Latest Version button.
Note ‘Get Latest Version’ operations do not change the checked-in/
checked-out status of files. For example, if the ‘Get Latest Version’
operation is performed on a file that is currently checked out, the file
remains checked out (if the file is open, you will be prompted as to
whether or not you want to have the file reloaded with the new
version).
Note To get the latest version of an entire project, right click the
project node and select Get Latest Version from the context menu.
Note Right-click the Scripts, Data Files, or Include Files nodes to
select all files of those types for a ‘Get Latest Version’ operation. If
required, on the subsequent Get Latest Version dialog, individual
files of that type can be deselected.
Checking In/Out Project Files
In addition to offering source control commands that operate on all the files of a
complete project (project file, custom include files, data files, and scripts), a sub-
302 SilkPerformer User Guide

16 SOURCE CONTROL INTEGRATION
Synchronizing Source Control Status
set of SCC commands that select out only the project file itself are available
from the File menu (File/Source Control/Project File Only) and the project
node of the project tree. These commands operate the same as their counterparts
for associated files do; the only difference is that they do not affect the
associated files (scripts, custom include files, and data files). Such functionality
is useful if, for example, you want to check in a project file without overwriting
the associated script, custom include, and data files that are currently stored in
the SCC.
The project-file only commands that are available from the File menu (File/
Source Control/Project File Only) and the project node of the project tree are:
- Get Latest Version
- Check Out
- Check In
- Undo Checkout
- Add to Source Control
Removing Source Control Linkage
The Remove Source Control Linkage command is the reverse of the Add to
Source Control command. This command clears all linkage information from
the project file that connects the version of a file on your local disk with the
version of the file in source control. This command is useful when working with
files (possibly received from a customer or support staff) for which the
associated source control system is not available.
Note Source control linkage can be removed for complete projects
only, not individual files.
Procedure To remove source control linkage from a project:
1 In the project tree, right-click the project node of a project that is under
source control and select Remove Source Control Linkage (or select
Source Control/Remove Source Control Linkage from the File menu).
2 Click OK on the subsequent dialog, confirming your understanding that
all source control connection data for the project will be lost.
Synchronizing Source Control Status
Under certain conditions, SilkPerformer may fall out of sync with source control
systems due to the corruption of linkage information. For example, if one were
to close SilkPerformer while SilkPerformer is loaded with a checked out file and
SilkPerformer User Guide 303

16 SOURCE CONTROL INTEGRATION
Known Issues and Workarounds
then check in that same file via the SCC, SilkPerformer would not be aware that
the file had been checked in. SilkPerformer’s Synchronize command looks at
the check-in/check-out status of files in the source control system and updates
the status of those files in SilkPerformer.
Note This function can be applied to complete projects only, not
individual files.
Procedure To synchronize the source control status of a project:
1 In the project tree, right-click the project node of a project that is under
source control and select Synchronize (or select Source Control/
Synchronize from the File menu).
2 Click OK on the subsequent dialog, confirming your understanding that
all source control connection data for the file will be lost.
Known Issues and Workarounds
SCC client's reconnect functionality is not supported
Issue: When SilkPerformer projects and/or files are linked to a defined SCC
directory, and the directory is changed via the SCC client, SilkPerformer does
not reconnect to the new directory.
Workaround: Remove the existing source control linkage (as explained in
“Removing Source Control Linkage”) then add the affected projects/files to the
new directory (as explained in “Placing Projects and Files Under Source
Control”).
Note It is generally not recommended to use SCC client
functionality for maintaining linked SilkPerformer projects.
MS VSS 8.0 requests username/password for each file check-in/check-out
operation
Issue: MS VSS 8.0 introduced a security setting that, when activated, requests
the user to enter their username and password each time they check in or check
out a file from VSS. This request also appears in SilkPerformer when
performing file operations with VSS.
Workaround: To disable MS VSS manual login:
1 Open SourceSafe Administrator.
2 Open the database where your SilkPerformer projects are saved.
3 Select the Options command from the Tools menu (Tools/Options).
4 Select the General tab.
5 Check the Use network name for automatic user login checkbox.
304 SilkPerformer User Guide

16 SOURCE CONTROL INTEGRATION
Known Issues and Workarounds
Note This procedure only works if the username in VSS is identical
to the network username.
SilkPerformer User Guide 305

16 SOURCE CONTROL INTEGRATION
Known Issues and Workarounds
306 SilkPerformer User Guide

17
Other 17 Chapter
Supported
Environments
Introduction
What you will learn
This chapter explains SilkPerformer’s support of other computing
environments.
This chapter contains the following sections:
Section
Page
Overview 307
Windows Vista 308
Multibyte Support 308
AJAX Support for Web Application Testing 311
SAP eCATT Support 312
Terminal Emulation Support 314
GUI-Level Testing Support 317
SilkPerformer Plug-In for Eclipse 325
Remedy Support 333
Adobe Flex AMF3 Support 337
Customizing Oracle Applications 12i Scripts for Replay 349
Overview
SilkPerformer supports testing within a variety of computing environments.
This chapter explains SilkPerformer’s support of Windows Vista, multibyte
character sets, SAP eCATT, and terminal emulation, and more.
SilkPerformer User Guide 307

17 OTHER SUPPORTED ENVIRONMENTS
Windows Vista
Windows Vista
With Windows Vista, Microsoft encourages users to use standard user accounts
for their daily work instead of administrator accounts. Several changes in
privilege management have been made to make this possible. Microsoft is also
educating users to follow the new paradigm by introducing User Account
Control (UAC). It is the UAC’s task to inform users when an application leaves
the standard user account privileges path and requires higher-level privileges,
even when the user account already has the required privilege. For this reason,
even administrators are prompted while performing tasks that require
administrator privileges.
More than ever before, software producers are encouraged to write Windows
applications that require as few privileges as possible for normal operation.
Although SilkPerformer does not require administrator privileges for most tasks
during normal operation, there are some tasks that do need them:
- Recording: SilkPerformer’s recording architecture requires that qaphooks.dll
be copied into the Windows\system32 directory in order so that it will be in
every application's .dll search path.
- Working Directories: The Project, Custom Include, and Custom Data
directories are located in the Program Files\Borland\... hierarchy. However,
these directories can be configured from within SilkPerformer.
- Automatic Agent launch via the launcher service.
To provide a smooth workflow, SilkPerformer has integrated an application
manifest that informs the system that it requires administrator privileges for
normal operation.
When SilkPerformer is launched on Windows Vista with the UAC enabled,
users are informed that applications require administrator privileges. Once users
opt to continue, they will no longer be interrupted by any OS-generated
privilege-related SilkPerformer message.
Without UAC enabled, SilkPerformer runs as usual without any interruptions
from the system.
Note With UAC enabled, Windows Vista can not be used as a
SilkPerformer Agent machine.
Multibyte Support
Overview
Unicode is a list off all known characters. It includes all alphabets of all spoken
and unspoken languages. Each character has its own unique index in the
Unicode list. The first 128 characters are known as ASCII characters.
308 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Multibyte Support
When data is stored or computed, the Unicode list is not used to represent the
characters. Instead, so-called character encodings (or character sets) define how
characters are represented on computers and within files. Numerous character
sets are used throughout the world.
Two of the more frequently-used character encodings that cover all Unicode
characters are:
• UTF-8: Requires 1-4 Bytes per character; this character set is widely
used for international Web sites and international text representation.
Strings in the Linux kernel and Java are encoded in UTF-8.
• UTF-16: Requires 2 or 4 Bytes per character; this encoding is mainly
used for string representation in the Windows NT kernel (Win NT and
newer).
Most character sets do not define a representation for all characters in the
Unicode list. Rahter, they define a subset of characters that are used in a specific
regional area of the world. These are commonly referred to as code pages (cp):
• ASCII: Single Byte encoding: 1 Byte per character
• Latin-1 (Windows Codepage 1252): Single Byte encoding: 1 Byte per
character
• Shift-JIS (Windows Codepage 932): Double Byte encoding: 1 or 2 Bytes
per character
• EUC-JP (Windows Codepage ): Includes 3 Japanese char sets: 1, 2, or 3
Bytes per character
When developing applications for Windows, programmers can choose between
Unicode (UTF-16) string representation or Multi-Byte-Character-Set (MBCS)
string representation. MBCS representation refers to a geographic regiondependent
code page encoding (for example, Shift-JIS for Japan; LAtin-1 for the
Americas and most European countries). This affects all GUI elements, as all
data needs to be displayed in the same string representation that has been
selected for the application.
Mutlibyte support in
SilkPerformer
SilkPerformer is a MBCS-based application, meaning that to be displayed
correctly, every string must be encoded in MBCS format. Because
SilkPerformer, and particularly TrueLog Explorer, visualize and customize data
that originates from Web servers with different encodings (UTF-8, EUC-JP,
SilkPerformer User Guide 309

17 OTHER SUPPORTED ENVIRONMENTS
Multibyte Support
Example
Codepage check
Replaying the correct
encoding
ISO-8859-1, etc.), many string conversion operations may be involved before
data can be displayed.
Sometimes when testing UTF-8 encoded Web sites, data containing characters
cannot be converted to the active Windows system codepage. In such cases,
SilkPerformer’s Web Recorder scripts all non-convertible strings into hex
format. If the data is only displayed, TrueLog Explorer will replace the nonconvertible
characters with a configurable replacement character (usually '?').
On an English Windows system with Japanese local settings (system codepage
is Shift-JIS) a UTF-8 application is recorded. For some reason a link name on a
particular Web page contains a Korean character. SilkPerformer’s Web Recorder
attempts to convert the string from UTF-8 to Shift-JIS to make it readable in the
script. The conversion fails because of the Korean character, which does not
occur in the Shift-JIS code page. The Web Recorder then uses hex notation for
the link name in the script.
One of the most important things to know about SilkPerformer's multibyte
support is that the system codepage of the recording machine must match the
system codepage of the replay machine for the replay engine to correctly
interpret the byte sequences in the script as characters and strings. Therefore
SilkPerformer’s Web Recorder scripts the @codepage annotation into the script
along with the currently active system codepage. Whenever this annotation is
present in a script the SilkPerformer runtime system checks to see if the system
codepage of the machine matches the one specified in the script. If they do not
match the runtime stops and an appropriate error message is displayed.
No matter which character encoding a Web site uses, SilkPerformer can only
visualize MBCS strings. However, there are several API functions which
require some string parameters to be encoded in the encoding that the server
expects. So that replay engines can know the requried encoding, SilkPerformer’s
Web Recorder scripts SetEncoding(), ToEncoding() and FromEncoding()
functions at appropriate locations. Typically, those parameters that refer to data
that needs to be interpreted by the server, require an appropriate encoding (for
example, link names). Those parameters that are not a part of client/server
communication do not require conversion (for example, timer names).
The SetEncoding() function sets an encoding for all subsequent ToEncoding()
and FromEncoding() functions. The encoding itself is passed as the name of the
encoding as it appears in the charset specification of the Web page, for example
"UTF-8", "SHIFT_JIS", "WINDOWS-1252", "WINDOWS-1255", or "EUC-
JP".
ToEncoding() converts the string that is passed as a parameter, which is encoded
in the system codepage (MBCS), to the encoding specified by the last
SetEncoding() function.
FromEncoding() converts the string that is passed as a parameter, which is
encoded in the encoding specified by the last SetEncoding() call, to the system
codepage (MBCS).
310 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
AJAX Support for Web Application Testing
Copying & pasting
When copying strings into SilkPerformer, the paste operation checks if the data:
• Needs to be converted
• Can be converted to the MBCS (system codepage)
• Can not be converted and so needs to be opened with SilkPerformer’s
Unicode Text/Hex Editor
If copied data is, for example, UTF-8 encoded and the paste operation fails to
convert it to the system codepage, SilkPerformer launches its Unicode Text/Hex
Editor. The tool is a Unicode application and is therefore able to display any
string, regardless if the characters exist in the system codepage. The editor can
also display the hex notation of strings in order to paste them into a BDL script.
AJAX Support for Web Application Testing
The SilkPerformer Recorder can record and replay Web applications that utilize
AJAX (Asynchronous JavaScript and XML) requests. This is possible because
SilkPerformer can recognize asynchronous AJAX requests and responses that
arrive in the form of either XML or JSON within HTML responses.
SilkPerformer scripts AJAX requests that it encounters as WebPageUrl calls.
Understanding AJAX
AJAX and script
customization
Pretty-formatted JSON
and XML data
According to Wikipedia (see www.wikipedia.com), AJAX is “...a group of
inter-related web development techniques used for creating interactive web
applications. A primary characteristic is the increased responsiveness and
interactivity of web pages achieved by exchanging small amounts of data with
the server "behind the scenes" so that entire web pages do not have to be
reloaded each time there is a need to fetch data from the server. This is intended
to increase the web page's interactivity, speed, functionality, and usability.
AJAX is asynchronous; in that extra data is requested from the server and
loaded in the background without interfering with the display and behavior of
the existing page. JavaScript is the scripting language in which AJAX function
calls are usually made. Data is retrieved using the XMLHttpRequest object that
is available to scripting languages run in modern browsers, or alternatively
Remote Scripting in browsers that do not support XMLHttpRequest. There is,
however, no requirement that the asynchronous content be formatted in XML.”
SilkPerformer and TrueLog Explorer support access to values values within
AJAX requests. This enables script customizations such as input data
parameterization, verification, parsing, and session-information customization
within AJAX responses.
JSON and XML are data-structure formats commonly used in AJAX
applications, REST techniques, and other environments. SilkPerformer supports
pretty-formatted viewing of XML and JSON-formatted byte streams in .bdf
scripts. Enhanced rendering of JSON formatted data enables easier
SilkPerformer User Guide 311

17 OTHER SUPPORTED ENVIRONMENTS
SAP eCATT Support
Understanding JSON
Enabling prettyformatted
JSON and
XML viewing
Formatted JSON and
XML data in Truelog
Explorer
customization of string values via TrueLog Explorer’s string customization
functions.
When JSON-formatted data is recorded or inserted into a .bdf script,
SilkPerformer displays the raw JSON byte stream. Once displayed in JSON
format, XML can easily be customized using SilkPerformer’s Parameter
Wizard.
Note SilkPerformer offers the option of viewing JSON data in either
pretty-formatted JSON-renderng view or as raw JSON byte stream.
According to JSON.org (see www.json.org), “JSON (JavaScript Object
Notation) is a lightweight data-interchange format. It is easy for humans to read
and write. It is easy for machines to parse and generate. It is based on a subset of
the JavaScript Programming Language...” “JSON is a text format that is
completely language independent but uses conventions that are familiar to
programmers of the C-family of languages, including C, C++, C#, Java,
JavaScript, Perl, Python, and many others. These properties make JSON an
ideal data-interchange language.”
Procedure To pretty-format JSON and XML data:
1 Within a .bdf script that includes JSON- or XML-formatted data, rightclick
within the screen data that you want to have rendered for enhanced
viewing.
2 Select Format As / JSON or Render As / XML from the context menu to
pretty-format the data.
You can then right-click formatted data strings to access SilkPerformer’s
standard string customization commands.
Alternative Select Auto Format from the context menu to have
SilkPerformer determine the best formatting option for screen data
types. By default, JSON and XML data is pretty-formatted in your
.bdf scripts.
Note Because formatted JSON data is integrated into the
SilkPerformer code editor, JSON-formatting can be undone/redone
using the Undo/Redo buttons on the toolbar.
Pretty-formatted JSON and XML data viewing is also available in Silk TrueLog
Explorer. See Silk TrueLog Explorer help for full details.
SAP eCATT Support
SAP eCATT has been integrated with SilkPerformer. SAP’s eCATT facility
allows you to create test scripts in SAP using the scripting language of your
312 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
SAP eCATT Support
Configuring a SAP
eCATT Connection
choice. eCATT allows you to use external test tools (i.e., SilkPerformer) while
utilizing eCATT as a repository for your test scripts. eCATT also serves as a
basic test management solution for triggering script executions. Not only can
both internal and external scripts be executed individually, they can also be
combined and executed in sequence.
eCATT offers import arguments, a mechanism for calling scripts with special
input values. Scripts can not only receive input values, scripts can also set output
values when they are executed—scripts can be executed in sequence, using
input values derived from the output values of earlier script executions.
For full details regarding SilkPerformer’s integration with SAP eCATT, please
refer to the SAPGUI Tutorial. For more information regarding eCATT, please
consult SAP documentation.
Connection details for SilkPerformer’s communication with SAP eCATT must
be specified in SilkPerformer system settings. There are two options for
connecting to SAP: you can either specify a ‘SAPLOGONID’ or you can
specify ‘AS Host’, ‘RFC Type’, and ‘SystemNr’ settings. With either option you
must specify client, language, username, and password details. Note that when
you select a SAPLOGONID the AS Host, RFC Type, and SystemNr fields are
grayed out.
Procedure To specify SAP eCATT connection data:
1 Select the System command from SilkPerformer’s Settings menu.
2 On the System Settings - Workbench dialog, select the SAPGUI group
icon.
3 The eCATT Connection tab is selected by default. From the SAPLogon
drop box, select your SAP login ID. This box is preconfigured with all
available SAP login IDs.
4 In the AS Host field, enter the combined router/application-server string
(e.g., H/195.61.176.22/H/194.117.106.130/S/3297/H/cpce801).
5 In the RFC Type field, enter either ‘3’ (for R/3) or ‘2’ (for R/2).
6 In the System NR field, enter the SAP system number.
7 In the Client field, enter the internal client ID number from the SAP
server (i.e., the value that must be entered on the SAP login screen).
8 From the Language drop box, select your language preference. The
values ‘EN’ (English) and ‘DE’ (German) are preconfigured, though you
can specify any other language abbreviation string if that language is
installed on the SAP system.
9 In the Username field, enter your SAP eCATT username.
10 In the Password field, enter your SAP eCATT password.
SilkPerformer User Guide 313

17 OTHER SUPPORTED ENVIRONMENTS
Terminal Emulation Support
Configuring eCATT
extended results
11 Once you have completed this dialog, click Test Connection to confirm
that you have specified accurate connection details. If your connection
attempt is unsuccessful, please confirm your settings.
12 Click the OK button once you have completed configuring SAP eCATT
connection settings.
To enable the viewing of SilkPerformer result files from within SAPGUI, you
can specify a UNC path to a public file share in which extended SilkPerformer
test results can be stored and accessed by users (e.g., \\fileserver\ecattresults).
SilkPerformer will use the specified directory to store the results of
SilkPerformer test executions initiated via SAP eCATT. Users can easily access
test results by clicking a link in the SAP eCATT GUI.
Procedure To configure eCATT extended results:
1 Select the System command from SilkPerformer’s Settings menu.
2 On the System Settings – Workbench dialog, select the SAPGUI group
icon.
3 Select the eCATT Results tab.
4 Select the Use SAP extended results checkbox.
5 In the SAP extended results directory field, browse to and select the
directory that is to be used for SAP extended results.
6 Click OK to save your settings.
Terminal Emulation Support
Testing IBM
mainframe protocols
SilkPerformer’s support for terminal emulation applications enables the
recording of terminal-emulator traffic based on the Telnet protocol (aka “green
screen” applications). Supported terminal types include VT100, VT200 etc.
(UNIX, IBM AS400) and IBM mainframes accessed via TN3270(E) &
TN5250.
A prerequisite for recording terminal-emulator traffic is an installed terminalemulator
for accessing host applications, while the recorder records the telnet
traffic which results. Many such emulators are available on the market (e.g.,
Rumba).
The Host Screen Info tab of terminal emulation TrueLogs offers controls for all
fields and statuses that appear in recorded and replayed screens. These controls
offer easy access to field-value and value-status parsing/verification functions.
See TrueLog Explorer User Guide for details on available parsing and
verification script-customization functions.
314 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Terminal Emulation Support
IBM terminals rely on EBCDIC code pages that vary based on locality.
SilkPerformer chooses an EBCDIC code page among the code pages installed
on the system. If national language characters are not displayed correctly it is
possible to change the code page during project configuration. See Creating a
SilkPerformer Project for details.
IBM code pages control translation of human-readable ASCII input data (e.g.,
search parameters) into and out of binary hexadecimal representation. This
translation process facilitates script customization.
Creating a terminal
emulation project
Configuring profile
settings
Procedure To create a terminal emulation load testing project:
1 Click the Start Here button on SilkPerformer’s Workflow bar.
2 On the Workflow – Outline Project dialog, define a name and optional
description for your project.
3 Expand the Terminal Emulation application type node.
4 Select the appropriate terminal type subnode.
5 Click OK.
Before beginning recording, you must configure profile settings for the
recording of terminal emulation applications via terminal emulators.
1 Select the Active Profile command from SilkPerformer’s Settings menu
to open the Profile dialog.
2 From the Record group tab (as these settings are only relevant to
application recording) select the Terminal Client icon to open the Telnet
tab.
3 In the Telnet Settings portion of the dialog, ensure that the Auto-detect
Telnet-mode radio button is selected (the other two options are for
trouble-shooting purposes).
4 In the Terminal Properties area, ensure that the specific terminal type
you are testing is displayed in the Terminal type drop-list box (this
setting was configured automatically when you selected the application
type for the project).
Note If you select a new terminal type on this dialog, you might be
presented with a dialog telling you that the default host code page
has been changed, and with a dialog telling you that the default
terminal configuration string has changed. Click Yes on these
dialogs.
5 Although the default dimensions (Rows and Columns) and Colors of the
terminal type are detected automatically for recording, if required you
can define these settings manually in the Default dimensions portion of
the dialog.
SilkPerformer User Guide 315

17 OTHER SUPPORTED ENVIRONMENTS
Terminal Emulation Support
Setting up an
application profile
Recording terminal
emulation applications
6 Ensure that the required host code page is displayed in the Host code
page drop-list box, if the preselected value default isn’t suitable (this
setting was configured automatically when you selected the application
type for the project).
Note The Host code page drop-list box includes all code pages that
are both installed and enabled. To review all installed code pages, or
to enable a specific code page, go to Start/Control Panel/Regional
and Language Options/Advanced. Only those host code pages that
are listed and enabled with a checkmark are available.
Procedure To set up an application profile to record a terminal emulator:
1 Select System from SilkPerformer’s Settings menu.
2 Click the Recorder icon.
3 Click the Add button.
4 On the Application Profile dialog, give the profile a name in the
Application profile field.
5 For the Application path field, browse to and select the .exe file of the
terminal emulator on your local system.
Note All other settings on this dialog are standard for all TCP/IP
based applications.
6 Select Custom Application as the Application type.
7 Select the TCP/IP protocol checkbox (note that SilkPerformer’s TCP/IPlevel
API has been enhanced to facilitate terminal emulation application
support).
8 Click OK. The new profile now appears in the Application Profiles tab.
Procedure To record a terminal emulation application:
1 Click the Model Script button on SilkPerformer’s Workflow bar to open
the Model Script dialog.
2 From the Select application profile drop-list box, select your newly
created application profile (as explained in “Setting up an application
profile”).
3 Click OK to begin recording the session.
4 Perform required interactions with the application under test. When
recording is complete, close the application and then close the recorder
and save the test script.
316 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
5 To confirm that your test script runs as expected, execute a TryScript run
in animated mode; you will then be able to view the results of the test run
in TrueLog Explorer. Please see TrueLog Explorer User Guide for details
about working with terminal emulation application TrueLogs.
Session customization of terminal emulation application test scripts is normally
not required. If session customizations are required, they must be partly
performed manually. However, TrueLog Explorer can be used to insert parsing
functions into test scripts.
In some cases the synchronization between replay and host application may not
work well, for example if the requested screen is not complete but the cursor
position already assumes a completed screen. A recorded cursor
synchronization function in the script will be satisfied with the first portion of
the screen and continues executing the script out of synchronization, which may
lead to replay errors. In this case you would have to replace the recorded
synchronization function(s) manually with a more suitable synchronization
function in the script (e.g., WebTcpipRecvPacketsUntilDataor
WebTelnetRecvRecordsUntilData). See the BDL Function Reference for
function descriptions.
GUI-Level Testing Support
SilkPerformer supports GUI-level load testing (load testing of fat-client
applications). To load test these applications, SilkPerformer uses Windows
Terminal Services to connect to and open client sessions on agent computers.
Via its own session manager, SilkPerformer uses SilkTest, Borland’s functional
testing tool, to communicate with and control sessions (SilkTest likewise
communicates with SilkPerformer via the session manager). SilkTest launches
clients, runs GUI-level tests, and communicates results back to SilkPerformer
via the session manager (see “Exploring SilkTest Results”). SilkPerformer
ultimately collects the results and makes them available to end-users.
A GUI-level testing session is opened through the Remote Desktop protocol.
This connection is established without a GUI. See “Configuring Windows” for
pre-requisite information. Virtual users rely on OrbServer/Client to connect to
remote sessions. SilkPerformer’s session manager is used as a proxy between
virtual users and SilkTest. The session manager includes an Orb client that
receives and sends Data to SilkPerformer virtual users and an Orb Server that
forwards requests to SilkTest, and reverses the reply data to the Orb Clients.
SilkTest can be started in local host mode. With this approach, virtual users use a
console session rather than a separate Windows session. See “GUI-Level
Testing options” for details.
SilkPerformer User Guide 317

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
Note GUI-level testing only works on Windows server edition
platforms. Windows Home or Professional editions can only be used
as agents for a single GUI-level testing VUser (see also “GUI-Level
Testing options” for information on agent configuration).
Functions provided by
SilkPerformer
SilkPerformer provides the following functions for SilkTest GUI-level testing
replay (see also BDL Functions Reference):
StInitSession
StInitSession(in sUser : string optional,
in sPassword : string optional) : boolean;
Initializes a connection to a session or starts the Session Manager on the
localhost. It also launches SilkTest. If sUser or sPassword are null in the
StInitSession, the VUser will look up the profile setting. If username and
password values have been set in the profile settings, they will be used.
Refer to “Specifying user credentials” for detailed information.
StSetExecutionTimeOut
StSetExecutionTimeOut(in nTimeOut : number);
Sets a timeout for SilkPerformer to wait for the end of each testcase
execution. The default timeout is 300 seconds. To define an unlimited
timeout, set the timeout value to ’0’.
StSetConnectionTimeOut
StSetConnectionTimeOut(in nTimeOut : number);
Sets the timeout for waiting for a session to respond to SilkPerformer. If
the timeout is reached, no session is established. The default timeout is 50
seconds. To define an unlimited timeout, set the timeout value to ’0’.
Table 29: GUI-level testing functions
318 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
StExecuteTestcase
StExecuteTestcase(in sScript : string,
in sTestCase : string, in sTestData : string,
in nTimer : string) : boolean;
This function starts the process of performing a full test in an alternate
session. The user decides if TryScript runs of the testcase are run on the
console or within a session. Sessions require a username and password
(these values should be provided to the StInitSession). If the username and
password are null, this function executes SilkTest on the localhost. If the
username and the password are not specified and the user tries to run more
than one VUser, the process will be stopped because it is not possible to
run more than one VUser on a local host. Refer to “Specifying user
credentials” for detailed information.
Note StExecuteTestcase returns True when a testcase is
complete. Returned values do not indicate the success of
testcases themselves. Testcase related errors are not handled at
this level.
SilkTest’s execution mode is retrieved through the extension of the script
file.
“.t” Execution Mode: SilkTest Script
“.g.t” Execution Mode: Data Driven SilkTest
“.pln” Execution Mode: SilkTest Plan (not yet implemented by SilkTest)
If the path to the script file is relative, the script must be added to the data
files of the project, or the working directory must be set.
When using StExecuteTestcase in combination with StLoadProject (see
below), the script file needs not be added to the data files of the project, but
the relative path of the script file within the SilkTest package must be
specified.
StSetOption
StSetOption(in sOption : string, in sValue : string)
: boolean;
Sets SilkTest options.
Table 29: GUI-level testing functions
SilkPerformer User Guide 319

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
StLoadProject
StLoadProject(in sProject : string) : boolean;
Sets specific environmental values that are required for replaying
testcases. Only SilkTest packages can be loaded with this function (.stp
files), and the packages may not be password-protected.
If the path to the SilkTest package is relative, the package must be added to
the data files of the project.
StResetSilktest
StResetSilkTest(in bResetConnection : boolean
optional) : boolean;
Restarts SilkTest; or shuts down SilkTest, disconnects from the session,
and starts a new session where SilkTest is running
(bDisconnect = VTRUE).
StLogoffSession
StLogoffSession();
Disconnects from the session.
Table 29: GUI-level testing functions
Configuring Windows
Before executing GUI-level testing with SilkPerformer, you must configure
your Windows operating system. Additionally, SilkTest version 2006 R2 or
higher needs to be installed on the agent computer (please refer to the SilkTest
documentation for detailed information about setting up SilkTest).
Procedure To configure Windows for GUI-level testing:
1 Go to Start/.../Control Panel/System.
2 Select the Remote tab.
3 Verify that the Enable Remote Desktop on this computer option is
enabled.
4 Go to Start/.../Administrative Tools/Computer Management.
5 Expand the tree and select Local Users and Groups.
6 Select the Users directory.
7 In the User field, double-click the virtual user that you want to use in
your test.
8 On the Properties dialog, select the Member Of tab.
320 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
9 Ensure that this user is a member of one of the following two groups:
Administrators and/or Remote Desktop Users (if the user is not a member
of one of these groups, click the Add button and add the user to one or
both of these groups).
Licenses for remote
desktop sessions
Windows Server editions offer multiple licenses for remote desktop sessions,
which means that you can connect multiple times during a single session.
Note Windows Server editions offer two free licenses for remote
desktop sessions. Windows Home or Professional editions can only
be used as agents for a single GUI-level testing VUser (see also
“GUI-Level Testing options” for information on agent configuration).
Procedure To obtain more licences for Windows Server editions:
1 Go to Start/.../Control Panel/Add or Remove Programs.
2 Click the Add/Remove Windows Components button.
3 Within the Windows Components Wizard, scroll to and select the
Terminal Server checkbox.
4 Click Next to enable application mode for your Windows system.
Note A license server is required for application mode. Contact
Microsoft for information about obtaining a license server.
5 Restart your computer.
Configuring Terminal
Services settings
Procedure To see how many sessions are available on your system:
1 Go to Start/.../Administrative Tools/Terminal Server Licensing.
2 Using the Terminal Server Licensing tool you can see which terminal
services licensing servers are available and which licenses you have.
3 For additional information about using the Terminal Server Licensing
tool, please refer to the integrated Help of the tool.
It is important to configure your terminal services settings correctly, otherwise
GUI-level test execution is likely to fail.
Procedure To configure your RDP settings:
1 Go to Start/.../Administrative Tools/Terminal Services Configuration.
2 Right-click the Remote Desktop Protocol-TCP (RDP-Tcp) icon.
3 Select Properties from the context menu.
4 Select the General tab.
- Verify that Security layer is set to RDP Security Layer.
- Verify that Encryption level is set to Low.
SilkPerformer User Guide 321

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
- Verify that all other settings are unselected/left blank.
5 Select the Logon Settings tab.
- Verify that the Always prompt for password check box is not
selected.
6 Select the Sessions tab.
- Verify that Override user settings is selected and that End a
disconnected session is set to 1 minute.
- Verify that all other settings are unselected/left blank.
7 Select the Environment tab.
- Verify that Run initial program specified by user profile and Remote
Desktop Connection or Terminal Services client is selected.
8 Select the Remote Control tab.
- Verify that Use remote control with default user settings is selected.
9 Select the Client Settings tab.
- Verify that all connection settings in the Connection section are
selected.
10 Select the Network Adapter tab.
- Verify that All Network adapters configured with this protocol is
selected in the Network adapter list.
11 Click OK.
Terminal Services
Manager
Procedure To configure Terminal Services settings:
1 Go to Start/.../Administrative Tools/Terminal Services Configuration.
2 Select the Server Settings folder.
3 Right-click any settings icon in the Server Settings directory.
4 Select Properties from the context menu and configure settings as
required. For additional information about using the Terminal Services
Configuration tool, please refer to the integrated Help of the tool.
Note Typically Licensing settings should be set to Per Device and
the Restrict each user to one session setting should be set to No.
With the Terminal Services Manager tool you can see which connections are
open to your computer. Here you can see the control application (“Console”)
denoted with an ID number. You can also see any virtual user connections
(denoted with different IDs). For additional information about using the
Terminal Services Manager tool, please refer to the integrated Help of the
tool.
322 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
Note No configuration settings are necessary to enable GUI-level
testing with SilkPerformer.
Executing Tests
Setting up projects
Modeling scripts
Procedure To setup a GUI-level testing project:
1 Select New Project from the File menu.
2 Enter a Project Name.
3 (Optional) Enter a Project Description.
4 From the Application Type list box, select GUI-Level Testing/SilkTest.
5 Click OK.
Procedure To model a testing script based on a SilkTest test case:
1 Click the Model Script button.
2 Select the Generate custom script radio button.
Alternative Select the New Script radio button if you intend to
model a testing script based on SilkPerformer’s GUI-level testing
template file. Then, manually edit the file’s BDL code to meet your
needs.
Note You can specify a SilkTest option file from the Select SilkTest
option file field by clicking the field’s browse (...) button and
selecting an option file. Option files are .opt files that contain all
SilkTest project options. These are the same option settings that are
available in SilkPerformer’s GUI-Level Testing profile settings (see
“Configuring project settings” for details). To edit a SilkTest options
file, right-click the file in the Include Files node of TrueLog
Explorer’s project tree. See the Silk TrueLog Explorer for SilkTest User
Guide for full details.
3 The Add Testcase dialog appears.
4 Select the Open a SilkTest test case file radio button if you plan to create
a test case directly from a test case file.
Alternative If a test case file you intend to work with is located
within a SilkTest package file, select the Open a SilkTest package
file radio button. Then browse to and select the package file (.stp)
using the browse (...) button next to the Select SilkTest package file
field.
SilkPerformer User Guide 323

17 OTHER SUPPORTED ENVIRONMENTS
GUI-Level Testing Support
5 Use the Browse (...) button next to the Select test case file field to browse
for a test case file (.t) on your local system.
6 Select a specific test case within the test case file by selecting it in the
Select test case drop-down list box.
7 (optional) You can add SilkTest test data to the selected test case, if
required. Enter test data into the Test data field using the format “”, (For example, "test", 10).
8 Click the Add button. The selected test case appears below in the Case
field.
9 Add more test cases to your project as required by repeating this
procedure.
10 Select the Use project attributes for session login checkbox to have
username/password strings taken from SilkPerformer’s pre-configured
project attributes (available at Project/Project Attributes) to enable login
for remote desktop sessions. These credentials will be added to your
script’s TInit transaction.
11 Click OK.
12 Give the test case a File name.
13 Click Save to save the test case.
14 Click Yes on the confirmation dialog to save the test case (.bdf ) to your
project. A BDL file that includes your newly created test case will be
generated for your project.
Configuring project
settings
Specifying user
credentials
To configure GUI-level testing profile settings, please refer to “GUI-Level
Testing options”.
User credentials for GUI-level testing can be specified in the following areas:
• Profile settings (see “GUI-Level Testing options”)
• Project attributes (Username and Password project attributes are
automatically defined when creating a GUI-level testing project)
• Plain text specified in the BDL script
• Imported from data files
If you want each VUser to connect with different login credentials, we
recommend that you specify the credentials with project attributes (see Project
Attributes Overview in the BDL Function Reference for detailed information) or
that you use script customization through data files (see “Data Files” for
detailed information).
If you want each VUser to connect with identical login credentials, we
recommend that you specify the credentials in the profile settings (see “GUI-
324 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Using timers
Understanding result
files
Exploring SilkTest
Results
Level Testing options”) or that you specify the credentials with plain text in the
BDL script (see BDL Function Reference).
Note that the user credentials specified in the profile settings are used only if the
other options listed above are not used. If no user credentials are specified in any
of the areas listed above, SilkPerformer connects to the console session without
using the remote desktop protocol. In this case you can only run one VUser per
agent.
For GUI-level load testing, timers are of core importance. You can add timers in
your SilkTest scripts which will be reported to SilkPerformer’s load test results.
See the SilkTest documentation for detailed information on how to create timers
within SilkTest scripts.
Note SilkPerformer automatically generates names for SilkTest
timers without a name.
You can find the most recent TryScript TrueLog files in the RecentTryScriptTest
directory within your SilkPerformer project directory. During GUI-level testing
load tests, temporary SilkTest TrueLog files with the extension .xlgs are written.
After each SilkTest test case execution, the results of the SilkTest .xlgs and the
results of the SilkPerformer .xlg files are merged into the SilkPerformer .xlg files
(per VUser) and the temporary .xlgs files are deleted.
The RecentTryScriptTest directory within your SilkPerformer project directory
also includes SilkTest .res result files. These are the files that are displayed in
SilkTest when you initiate the Explore SilkTest results command (as explained in
“Exploring SilkTest Results”).
Procedure To explore GUI-level test results:
1 Within SilkPerformer, right-click a virtual user profile.
2 Select Explore SilkTest Results from the context menu. SilkTest
launches, allowing you to analyze the corresponding .res SilkTest result
file.
Alternative You can also select Explore TrueLog from the context
menu to view a TryScript’s TrueLog in Silk TrueLog Explorer.
Alternative Select the Results tab to view test results directly in
SilkPerformer (see “Exploring Test Results” for detaild
information).
SilkPerformer Plug-In for Eclipse
SilkPerformer is packaged with a plug-in for Eclipse, the popular Java IDE that
is based on plug-in technology. The plug-in enables Eclipse users to be
supported with SilkPerformer’s load-testing features.
SilkPerformer User Guide 325

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Structure of the
SilkPerformer
TestClass
Installing the plug-in
The structure of the SilkPerformer test class supports Java developers in
creating test cases for SilkPerformer. The test class implements the ITestClass
interface that is packaged with the Eclipse plug-in. It contains a SilkPerformer
object that represents the interface to SilkPerformer’s Java Framework.
The test class includes two basic methods:
testInit - This method initializes the interface. It is called at the beginning of test
runs.
testEnd - This method closes the interface. It is called at the end of TryScript
runs and load tests.
Between testInit and testEnd there may be one or more test methods invoked by
a running load test. Each method that has a signature void testXY() throws
SilkPerformerException.
Example:
public void testMyMethod() throws
silk.performer.SilkPerformerException
To use the plug-in, you must have Eclipse SDK (version 3.2 or higher) and
SilkPerformer (version 2006 R2 or higher) installed on the machine where the
plug-in is to run. Additionally, SilkPerformer must have been run on the
machine at least once.
Procedure To install the SilkPerformer plug-in for Eclipse:
1 Using the standard procedure for installing Eclipse plug-ins, go to the
Eclipse SDK’s Help menu, select Software Updates/Find and Install.
2 Select the Search for new features to install option button.
3 Click Next.
4 On the Update sites to visit page, you must specify the site that contains
your SilkPerformer Extras folder.
Note The Extras folder is located on your SilkPerformer Installation
CD. If you have used the Web setup package rather than an
installation CD, the Extras folder is located in the folder where you
extracted the package to.
5 Click New Local Site.
6 On the Browse for Folder page, browse to and select the Eclipse Plug-in
directory in your SilkPerformer Extras directory (.../Extras/Eclipse
Plugin/).
7 Click OK.
8 Click OK on the Edit Local Site dialog to confirm your directory
selection. The directory now appears in the Sites to include in search
pick list.
326 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Configuring the plugin
9 Select the checkbox of your newly defined local site (.../Extras/Eclipse
Plugin/) and click Finish.
10 On the Search Results page, click the plus sign of the new local site to
expand its contents.
11 Select the checkbox of the plug-in feature (SilkPerformer Feature).
12 Click Next.
13 A license description appears on the Feature License page. If you agree
to the terms of the license agreement, select the I accept the terms in the
license agreement option box.
14 Click Next.
15 On the Installation window, click Finish.
16 Verify the features you are about to install (listed on the Feature
Verification window) and click Install.
17 On the Install/Update dialog, click Yes to restart the Eclipse Workbench.
18 Upon restart, a SilkPerformer menu entry will appear in the Eclipse
workbench.
Procedure To configure Eclipse settings for the plug-in:
1 Go to Window/Preferences.
2 Select the SilkPerformer node on the navigation tree to display
SilkPerformer settings.
3 Confirm that the installation directory of your SilkPerformer installation
is listed in the Install directory text box. If it is not listed, click Browse to
browse for and select your SilkPerformer installation directory.
4 Edit the Default project name as required.
5 Edit the Default test class name as required.
6 (experienced users only) Edit the default Timeout for connecting to Java
debugger setting as required. This is the amount of time allowed (in
milliseconds) for connecting to a SilkPerformer process during tests.
7 (experienced users only) Edit the default Shutdown after finish delay
setting as required. This is the amount of time (in milliseconds) that
Eclipse waits before retrieving results after the completion of
SilkPerformer tests.
8 Select the Autostart monitoring checkbox to have Performance Explorer
begin server monitoring when SilkPerformer tests are initiated.
SilkPerformer User Guide 327

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
9 Select a Log Level to specify when logs are to be written. Selecting
‘None’ results in no log files being written. Selecting ‘Errors’ results in
log files being written only when errors are encountered. Selecting
‘Debug’ results in all activity being logged.
10 Select the Java / Installed JREs node in the navigation tree to open the
Installed JREs page. The Java runtimes that are used by Eclipse on your
local system are listed here. Confirm the Location and Type settings for
your Java runtime environment.
- If you need to add a Java runtime profile, click the Add button and
proceed with step 11.
- If you need to edit an existing Java runtime profile, click the Edit
button and proceed with step 11.
- If your Java runtime settings are acceptable, proceed with step 13.
11 On the Add/Edit JRE dialog, click the Browse button to browse to and
select the directory that contains your JDK installation (specifically, your
Java compiler), then click OK.
Note You must select a JDK (version 1.5 or higher), rather than a
JRE, because SilkPerformer attempts to recompile Java files in its
test classes.
12 The fields on the Add/Edit JRE dialog are automatically populated with
values derived from your JDK installation. Click OK to save your
settings.
13 Back on the Installed JREs page, click OK to save your settings.
Projects
Setting up New
Projects
There are restrictions on concurrent use of Eclipse and SilkPerformer. Projects
are locked from use by Eclipse while SilkPerformer Workbench works on them.
If you attempt to open a project from Eclipse that is currently in use by
SilkPerformer, you will receive an error message. You must close SilkPerformer
to pass project control back to Eclipse.
Procedure To set up a new project:
1 Within a SilkPerformer-enabled Eclipse workbench, select File / New/
Project to launch the New Project wizard.
2 Expand the SilkPerformer node in the navigation tree.
3 Select the Java Project node.
4 Click Next.
328 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Importing a
SilkPerformer project
into Eclipse
Configuring
SilkPerformer project
properties from
Eclipse
5 On the New SilkPerformer load testing capable Java project page, edit
the Name of your new project as required.
6 Edit the Test class name (the name of the initial test class template) as
required.
7 Click Finish to have the plug-in generate the test class template in your
Eclipse workspace and link the test classes to a test case in your
SilkPerformer project.
A class path links the new test class into SilkPerformer’s workspace so that it
works with both the Eclipse editor and SilkPerformer. Behind each test class is
BDF stub code. This code is generated each time a test class is compiled, so any
changes you make to BDF stub code will be overwritten the next time the test
class is compiled.
Another way of linking your Eclipse and SilkPerformer workspaces is to import
an existing SilkPerformer project into your Eclipse workspace.
Procedure To import a SilkPerformer project into Eclipse:
1 Select Import from the File menu.
2 On the Import dialog, expand the Other node in the navigation tree.
3 Select Import Existing SilkPerformer Java Project.
4 Click Next to view all available SilkPerformer projects in your
SilkPerformer workspace.
5 Select the checkboxes of the projects that you want to import.
6 Click Finish to import the selected project settings and classpaths into
your Eclipse workspace.
Procedure To configure SilkPerformer project properties from Eclipse:
1 Go go the Project menu and select Properties.
2 Select the SilkPerformer Project node in the navigation tree.
3 At the top of the SilkPerformer Project Properties page the path to the
Targeted SilkPerformer Project is displayed. If required, edit the Name
that has been specified for the project when opened in SilkPerformer
workbench.
4 Edit the Description that has been defined for the project, as required.
5 You can define SilkPerformer Project Attributes for the project (Name,
value Type, Value, Default value, Description). See Project Attributes
Overview in the BDL Function Reference for full details regarding
configuring SilkPerformer project settings.
6 Click Apply to apply your changes.
SilkPerformer User Guide 329

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Configuring
SilkPerformer system
settings
Configuring
SilkPerformer project
settings
Redirecting Java build
paths
7 Click OK to save the project settings.
Note SilkPerformer project attributes can also be accessed directly
by selecting Project Attributes from Eclipse’s SilkPerformer menu.
To configure SilkPerformer system settings, select System Settings from
Eclipse’s SilkPerformer menu. For full details on configuring system settings
for SilkPerformer workbench, see “SilkPerformer Workbench”.
To configure profile settings for an active SilkPerformer project, select Profile
Settings from Eclipse’s SilkPerformer menu. You will be directed to the
General tab for your project profile Java settings, where you can view the Java
version, Java home, and classpaths have been set up for your newly created
project. For full details on configuring profile settings for SilkPerformer
projects, see “Project Profile Settings”.
Note When Eclipse controls a SilkPerformer project, all Java
settings defined in SilkPerformer are overruled by the settings you
defined in Eclipse.
From the moment that SilkPerformer projects are controlled from Eclipse, the
classpaths of projects are controlled from the Eclipse environment. For this
reason you may want to adjust the Java build path of your Eclipse projects.
Note There are restrictions for classpaths of Eclipse projects that are
used within SilkPerformer. To enable the distribution of libraries
during a loadtest, all libraries must be located either in the
SilkPerformer project directory (\Working\Projects) or they must be
available on a UNC path.
Procedure To redirect an Eclipse project’s Java build path:
1 Right-click a SilkPerformer-enabled project in Eclipse’s Package
Explorer.
2 Select Build Path/Configure Build Path from the context menu.
3 Select the Libraries tab on the Java Build Path page. Here you will see
all of the User libraries that are required for running tests.
4 Click Add Library.
5 On the Add Library page, select User Library from the pick list and click
Next.
6 Click User Libraries.
7 On the User Libraries page, click the New button.
8 Specify a name for the new library on the New User Library dialog, and
click OK.
9 Back on the User Libraries page, select Add JARs.
330 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
Generating multiple
load-testing Java
classes
10 Browse to and select one or more Java libraries that are available to your
local system (hold down your Shift or Ctrl keys to select multiple
libraries).
11 Click Open to add the JAR files to your new user library.
12 Back on the User Libraries page, click OK.
13 Back on the Add Library page, click Finish to migrate these settings to
the classpaths of both your SilkPerformer Workbench project and your
SilkPerformer-enabled Eclipse project.
You can generate multiple load-testing capable Java classes for SilkPerformerenabled
Eclipse projects.
Note SilkPerformer can handle user-defined JAR libraries. It can
only launch classes that are within the root package of Java projects
within SilkPerformer. Only external libraries such as JARs can be
used.
Procedure To add a SilkPerformer test class to an Eclipse project:
1 Select Other from the File menu.
2 On the New page, select SilkPerformer/SilkPerformer Test Class from
the navigation tree.
3 On the New SilkPerformer Test Class page, edit the Project name and
Test class values as required.
4 Click the Browse button.
5 On the SilkPerformer-Enabled Projects dialog, select the SilkPerformerenabled
Eclipse project to which this new test class template is to be
appended.
6 Click Finish to append the test class template to your Eclipse project.
TryScript Runs
Executing TryScript
runs for test classes
Procedure To execute a TryScript run for a test class:
1 In Eclipse’s Package Explorer navigation tree, select the test class for
which you want to run a TryScript.
2 Select Try Script from Eclipse’s SilkPerformer menu. Eclipse then
automatically switches to the Debug perspective and initiates a TryScript
run. The bottom pane of the Debug page includes SilkPerformer views
that monitor the status of the test.
The two main SilkPerformer views are:
SilkPerformer User Guide 331

17 OTHER SUPPORTED ENVIRONMENTS
SilkPerformer Plug-In for Eclipse
- SilkPerformer Virtual User - This view displays virtual-user status
messages that are passed from the latest TryScript test.
- SilkPerformer simulation Controller - This view displays
simulation-status messages that are passed by the SilkPerformer
Simulation Controller.
3 If your test class generates errors or leads to a failed test, you must edit
your test class in the Eclipse Editor. If your TryScript run is successful,
you can view the results of the test in the SilkPerformer log files and
result files, including TrueLog files. See “Viewing Results” for full
details.
Viewing Results
Viewing Virtual User
Reports
Viewing Virtual User
Log Files
Viewing Virtual User
Output Files
Viewing TrueLogs
Following a successful test or TryScript run, you can choose to view a Virtual
User Report file, a Virtual User Log file, a Virtual User Output file or a TrueLog.
To view a Virtual User Report, select Virtual User Report File from Eclipse’s
SilkPerformer menu. The report is then displayed in Eclipse’s SilkPerformer
Report File Viewer tab. See “Virtual User Report files” for full details
regarding Virtual User Reports.
To view log files, select Virtual User Log File from Eclipse’s SilkPerformer
menu. The log file of the most recent test run is then displayed.
To view output files, select Virtual User Output File from Eclipse’s
SilkPerformer menu. The output file of the TestClass code for the most recent
test run is then displayed. See “Virtual User Output files” for full details
regarding Virtual User Output files.
To view a TrueLog, select Show TrueLog from Eclipse’s SilkPerformer menu.
The TrueLog file of the most recent TryScript run is then displayed in Silk
TrueLog Explorer. See the Silk TrueLog Explorer User Guide for full details
regarding TrueLog files.
Launch Configurations
Defining a launch
configuration for
SilkPerformer
TryScripts
Procedure To set up a launch configuration for TryScript runs:
1 Select the Debug button on Eclipse’s main toolbar.
2 On the Debug page, select the SilkPerformer TryScript node in the
navigation tree.
3 Click the New button on the toolbar.
332 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Remedy Support
4 On the Create, manage, and run configurations page, the currently active
SilkPerformer project and its first test class are pre-selected. Select the
browse (...) button alongside the Project text box.
5 On the SilkPerformer Enabled Projects dialog, select the SilkPerformerenabled
project that is to be associated with this launch configuration.
6 Click OK.
7 On the Launch as SilkPerformer test case dialog, select the TestClass
that is to be launched with this launch configuration.
8 Click OK.
9 (optional) Edit the parameters of any pre-defined project attributes in the
Project Attributes section of the page.
10 Click Apply to apply your changes.
11 Click Debug to run a TryScript run based on your launch configuration
settings.
Baseline Tests
Once you are satisfied with the TryScript results generated by your TestClass, it
is time to pass your test project from Eclipse to SilkPerformer and run a
baseline test. See “Establishing Baseline Performance” for full details
regarding baseline tests.
Running baseline
tests
Procedure To run a baseline test:
1 Select Open in SilkPerformer from Eclipse’s SilkPerformer menu.
SilkPerformer workbench launches, with your Eclipse project loaded in
the Project tree. All TestClasses are appended to the project under the
Scripts node.
2 Define and run a baseline test, as explained in “Establishing Baseline
Performance”.
Remedy Support
The testing workflow for testing Remedy with SilkPerformer is the same as the
workflow for testing Web-based applications, with the following exceptions:
• On the Workflow - Outline Project dialog, you must select the Remedy
version under test (these settings are grouped under the ERP/CRM node).
• On the Workflow - Model Script dialog, you enter the commandline path
to the Remedy installation under test.
SilkPerformer User Guide 333

17 OTHER SUPPORTED ENVIRONMENTS
Remedy Support
BMC tutorial
Licensing
TrueLog support
Data representation in
SilkPerformer
• Recorded scripts include a number of WebRemedy functions, which are
modified WebPagePost and WebPageUrl calls.
• In the Forms section of test scripts, each Web form contains an XML
fragment that is a visualization of the Remedy traffic that was generated
by the browser. This XML visualization of the browser traffic is easier to
read than actual Remedy traffic. It also facilitates SilkPerformer script
customization. The actual, originally recorded Remedy traffic is
commented at the end of each Web form XML fragment.
BMC has created an excellent tutorial that illustrates how to load test BMC
Remedy IT Service Management (ITSM) applications with BMC Remedy
Action Request System (AR System). To obtain the tutorial, contact BMC and
request the following white paper: Performance Benchmarking Kit: Using
Incident Management with SilkPerformer.
A premium virtual user license is required to test Remedy with SilkPerformer.
TrueLog functionality is not supported for Remedy testing.
Remedy uses a proprietary format for data that is sent from the client to the
server. This format includes length prefixes and makes customization difficult.
Example:
157/GetTableEntryList/11/smi-web-00126/
SMI:TEL35:SHR:AgentConsole12/Windows_View9/53687098211/
smi-web-00127/SMI:TEL35:SYS:IncomingEmail0/1/01/02/0/0/
2/0/2/0/2/0/
Such data is sent in HTTP requests to the URL http:///arsys/
BackChannel.
Remedy versions 6.3 patch level 2 and earlier send such data in the body of
HTTP POST requests. Versions 6.3 and later send such data in the query string
of HTTP GET requests.
The main benefit of SilkPerformer’s Remedy add-on is that this proprietary data
format is translated into an XML-based representation that eliminates the length
prefixes and introduces names to the individual data items. This makes scripts
easier to understand and customize. During script replay, this XML-based
format is transparently translated back into the original format.
This backward transformation is performed by the functions
WebRemedyBackChannelUrl and WebRemedyBackChannelPost.
WebRemedyBackChannelUrl is a replacement for the function WebPageUrl,
while WebRemedyBackChannelPost is a replacement for the function
WebPagePost.
Example:
Without the Remedy add-on the following API call would be recorded as:
334 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Remedy Support
WebPagePost("http://213.131.176.85/arsys/BackChannel",
"157/GetTableEntryList/11/smi-web-00126/
SMI:TEL35:SHR:AgentConsole12/W"
"indows_View9/53687098211/smi-web-00127/
SMI:TEL35:SYS:IncomingEmail0/1"
"/01/02/0/0/2/0/2/0/2/0/",
STRING_COMPLETE,
"text/plain; charset=UTF-8",
"arsys/BackChannel");
With the Remedy add-on the API call is recorded as follows:
WebRemedyBackChannelPost("http://213.131.176.85/
arsys/BackChannel",
"\r\n"
"\r\n"
" smi-web-001\r\n"
" SMI:TEL35:SHR:AgentConsole\r\n"
" Windows_View\r\n"
" 536870982\r\n"
" smi-web-001\r\n"
" SMI:TEL35:SYS:IncomingEmail\r\n"
" \r\n"
" 0\r\n"
" 0\r\n"
" \r\n"
" \r\n"
" \r\n"
" \r\n"
" \r\n"
"",
STRING_COMPLETE,
"text/plain; charset=UTF-8",
"BackChannel - GetTableEntryList");
The key differences are:
- Use of the function WebRemedyBackChannelPost instead of
WebPagePost
SilkPerformer User Guide 335

17 OTHER SUPPORTED ENVIRONMENTS
Remedy Support
Additional benefits
Remedy SilkEssential
- Data is in XML format instead of original format
- XML format does not contain length prefixes, which would
otherwise have to be considered during script customization.
- XML format specifies data type and meaningful names for each data
item.
In addition to the different representation of the data there are other benefits to
using the Remedy add-on:
- Remedy.bdh: A bdh file that implements the functions
WebRemedyBackChannelPost, WebRemedyBackChannelUrl and
WebRemedyInit.
- The recorder places a call to the function WebRemedyInit() into the
TInit transaction.
- The function WebRemedyInit() installs verifications to catch
application level errors.
- The recorder detects all timestamps which would otherwise be hard
coded into the script, and records the function GetTimeStamp()
instead.
- The recorder records parsing functions for IDs of newly created
entities and appropriate substitutions for all occurrences of such IDs
in the script with the parsed variable.
The Remedy application type is triggered through a SilkEssential package. It
contains a set of recording rules and a remedy.bdh. The remedy.bdh wraps the
WebPageCalls:
// wrapper function for WebPagePost
// param nDataLength will be ignored
function WebRemedyBackChannelPost(sUrl :
string;
sData :
string;
nDataLength : number
optional;
sContent : string
optional;
sTimer : string
optional;
formUrl : form
optional) : boolean
begin
WebRemedyBackChannelPost := WebPagePost(sUrl,
RemedyBackChannelData(sData), STRING_COMPLETE, sContent,
sTimer, formUrl);
336 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
end WebRemedyBackChannelPost;
and
// wrapper function for WebPageUrl for use with
BackChannel requests
function WebRemedyBackChannelUrl(sUrl : string;
sTimer : string
optional;
formUrl : form
optional) : boolean
var
sXmlData : string(100000);
sDummy : string;
begin
WebFormExpand(formUrl, sDummy, 1, true);
WebFormValueGet(formUrl, sXmlData, sizeof(sXmlData),
"param");
WebFormValueSet(ARSYS_BACKCHANNEL_INTERNAL_HELPER,
"param", RemedyBackChannelData(sXmlData));
WebRemedyBackChannelUrl := WebPageUrl(sUrl, sTimer,
ARSYS_BACKCHANNEL_INTERNAL_HELPER);
end WebRemedyBackChannelUrl;
dclform
ARSYS_BACKCHANNEL_INTERNAL_HELPER :
"param" := "";
Protocol detection and
conversion
Remedy protocol detection is not available by default. However, it is already
built into the SilkPerformer recorder and is enabled by choosing the appropriate
Remedy application type when creating a new project.
The Remedy data conversion into XML and back out of XML is implemented
in the ProxyEngine for recording and in the perfRemedy.dll, which is loaded into
the perfRun by the Remedy.bdh.
Adobe Flex AMF3 Support
AMF3 (Action Message Format, version 3) is a binary protocol that was created
by Adobe. It is primarily used for remote procedure calls by Flex applications.
A primary goal in the design of AMF3 is that transmitted data be as small as
possible. This requirement presents a challenge for protocol customization with
SilkPerformer. To make the AMF3 protocol customizable and human-readable,
SilkPerformer transforms AMF3 traffic into XML during script recording, and
SilkPerformer User Guide 337

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
Understanding Flex
Understanding
Advanced Flex Data
Services
transforms the XML back into AMF3 format during replay. This means that you
do not have to content with binary AMF3 data in hex string form. Rather, in
recorded scripts and TrueLogs, you see only XML-formatted data.
AMF0 and AMF3 support are tightly integrated (in fact AMF3 uses AMF0’s
message structure). Every AMF3 message begins as an AMF0 message (even
when they contain the version “3” in the first two bytes of the data stream). A
special AMF0 data type marker indicates where AMF0 encoding ends and
AMF3 encoding begins.
According to Adobe (see www.adobe.com), “Flex is a free, open source
framework for building highly interactive, expressive web applications that
deploy consistently on all major browsers, desktops and operating systems. It
provides a modern, standards-based language and programming model that
supports common design patterns. MXML, a declarative XML-based language,
is used to describe UI layout and behaviors, and ActionScript 3, a powerful OO
programming language is used to create client logic. Flex also includes a rich
component library with over 100 proven, extensible UI components for creating
RIAs, as well as an interactive Flex application debugger.
Rich Internet applications created with Flex can run in the browser using the
ubiquitous Adobe Flash ® Player software or on the desktop on Adobe AIR .
This enables Flex applications to run consistently across all major browsers and
across operating systems on the desktop. And using Adobe AIR, the crossoperating
system runtime, Flex applications can now access local data and
system resources on the desktop.
You can accelerate application development with Adobe Flex Builder 3, a
highly productive, Eclipse based development environment, and Adobe Live
Cycle Data Services ES, a set of advanced data services that can be used in Flex
development. Both of these products are available for purchase.”
According to Adobe, “Adobe LiveCycle Data Services ES provides a
comprehensive set of data-enabling features for using data in RIAs. It enables
RIAs to talk to back-end data and business logic in a faster, more efficient
operating model. LiveCycle Data Services ES also enables seamless integration
with LiveCycle ES business processes and document services.
BlazeDS is a free, open source project providing Flex Remoting and Messaging
to all developers. Flex Remoting provides a binary, serialized data transport
format called the ActionScript Message Format (AMF) to provide a fast,
efficient means of transporting data to your RIA which accelerates application
performance. Flex Remoting also makes it fast and easy for developers to
connect to back-end business logic and data. Flex Messaging adds realtime data
push and publish/subscribe, both powerful capabilities now made easy. Using
BlazeDS, you can start using these powerful Java server integration features for
free, then subscribe to the LiveCycle Data Services Community Edition for
certified builds and support, or upgrade to the full LiveCycle Data Services ES
Enterprise edition for a complete server solution.”
338 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
AMF3 Data Types
AMF3 provides a number of data types.
Simple Data Types
Integer
Double
String
Here are four simple data types that are provided by AMF3. In the binary
protocol, each of these data types is transmitted as a 1-byte type marker. The
corresponding XML tags for each data type are shown within brackets.
• Undefined ()
• Null ()
• False ()
• True ()
Integers in AMF3 are serialized as variable-length 29-bit unsigned integers. The
variable-length format ensures that integers take up as little space as necessary.
In XML format, integers look like this:
123456
The integer data type is encoded as a 1-byte type marker, followed by up to 4
bytes that encode the integer using the variable length format. Besides their use
as a data type, variable-length integers are used in other places within the
protocol.
A double stores an 8-byte double precision floating point value as defined in
IEEE 754. This is the data type that many programming languages refer to as a
double. In XML format, doubles look like this:
3.14159
In the binary format, doubles are encoded as 1-byte type markers followed by 8
bytes that contain the double value.
Strings are used to store all types UTF-8-encoded string data. Besides their use
as a data type, strings are also used to specify string data, such as classes or
member names. All string values in an AMF3 data stream are put into a
reference table which is typically generated when AMF3 or XML data is parsed.
A string can either be encoded as a string value or as a string reference. A string
reference refers to a previous occurrence of its number in the string reference
table. This feature is used to save space when the same string value occurs
several times in a single AMF3 data stream.
In XML format, a string looks like this:
Hello, World!
When string references are used, the referenced string is marked with a
reference ID (in the form of an attribute), and the string reference refers to the
reference ID:
This string is referenced
SilkPerformer User Guide 339

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
XML and
XMLDocument
Date
Array
AMF3 provides two data types to encode XML data, XMLDocument, which
represents a Flash legacy data type, and XML, which is a newer XML type. Both
data types encode XML as strings. The difference is that these strings are not
registered in the string reference table. They are registered in the object
reference table.
Note With the XML format, the actual XML content must be XMLencoded.
Due to a limitation with SilkPerformer's XML parser, it is
currently not possible to use XML CDATA sections to notate such
data.
&lt;Data&gt;some data&lt;/Data&gt;
&lt;Data&gt;some more data&lt;/Data&gt;
As with strings, XML data can be referenced to save space:
&lt;Data&gt;foobar&lt;/Data&gt;
In AMF3, the date is encoded as the number of milliseconds elapsed since
January 1st 1970 00:00, in the UTC timezone. The number of milliseconds is
encoded as a double. Dates also support references, and each date is registered
in the object reference table.
2007-02-18 11:29:00.000
2000-01-01 00:00:00.000
Arrays in AMF3 cover two types of data structures: (1) standard arrays, where
each element is identified by an index number, and (2) associative arrays, where
each element is identified by a key (in the case of AMF3, this is always a string).
In the binary representation of an array, the associative portion of the array is
listed first. This is followed by a separator (an empty string) and finally the
dense portion of the array (the standard array elements).
In XML representation, the elements of an associative array are marked by their
name attributes. Generally, arrays are not bound to a specific type, which means
that you can put any data type in them, including different data types.
23
42
EMEA\johndoe
2008-02-17 13:47:13.000
In the above example there is an array with three elements that have a name
attribute (meaning, they belong to the associative portion of the array) and two
340 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
Object
elements that do not have a name attribute, which means that they belong to the
dense portion of the array.
As with other data types, arrays can be referenced; their elements can be
referenced; even their elements’ names can be referenced:
bar
In this example, there is an array with two string elements. The first element has
the name “foo” and the value “bar”. The second element has no name, but
references the string with the string reference table ID #0 (the string “foo”).
Arrays themselves are registered in the object reference table.
Objects are the most powerful data type within AMF3. There are several
subtypes of classes that you need to be aware of. What all object types have in
common is that they bear a class name.
Object Traits
The traits of an object are its member names, plus information indicating if an
object is dynamic or externalizable.
An object with traits contains a list of member names followed by the
corresponding members. In binary format, member names are encoded as
strings without type markers. The number of member names is encoded in an
additional field of flags that also contains information about the object’s sub
type. Following the members, an optional, additional list of dynamic members,
each preceded by their respective member name, is included. The list of
dynamic members is only included if the dynamic flag is set to true and
terminated by an empty string.
strFirstName
strSurname
John
Doe
42
Object References
Objects can be directly referenced because they are registered in the objects
reference table.
42
SilkPerformer User Guide 341

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
Traits References
Each object of subtype object traits or externalizable traits is not only added to
the objects reference table, but also to the traits reference table. Other objects
can then refer to the traits of these objects. This means that such an object does
not need to come with the list of members and the object flags by itself, but only
with the actual members.
userid
firstName
lastName
jdoe
John
Doe
mmustermann
Max
Mustermann
Externalizable Traits
Externalizable traits contain the binary representation of an object, an
indeterminate number of bytes serialized in an unknown format. The format
depends on the class that serializes and deserializes this data. The class also has
to know how many bytes it has to consume from the byte stream.
SilkPerformer provides support for the three most common classes:
• flex.messaging.io.ArrayCollection
• flex.messaging.io.ArrayList
• flex.messaging.io.ObjectProxy
According to Adobe, these three classes are the most common, and are encoded
as AMF3. This means that these three classes are parsed like normal arrays or
objects, respectively.
In case a class other than these three is found, it is assumed that all bytes up to
the end of the stream belong to the externalizable traits. The content of such an
unknown externalizable trait is stored as Base64-encoded data.
1
2
3
4
342 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
hugo
hugo2
555-55-5555
Tyler
5
SGVsbG8sIHdvcmxk
ByteArray
References
A ByteArray holds an array of bytes. In XML representation, binary data is
encoded in Base64 format.
ByteArrays are registered in the object reference table.
SGVsbG8sIHdvcmxk
Three different reference tables are used. The following list offers an overview
of which data types are registered in which reference table.
• String Reference Table:
- String
- Class names (not an actual data type)
- Member names (not an actual data type)
- Value name of elements of the associative portion of an array (not an
actual data type)
• Object Reference Table:
- Object
- Date
- Array
- ByteArray
- XML
- XmlDocument
• Traits Reference Table:
- Object traits
SilkPerformer User Guide 343

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
- Externalizable traits
• Not registered in any reference table:
- Undefined
- Null
- False
- True
- Integer
- Double
Reference IDs are generated by the AMF3 parser. However in hand-written
XML documents, you can use custom reference IDs:
EMEA\johndoe
Warning Custom reference IDs are lost when an XML document is
transformed from XML to AMF3 and back to XML because the
AMF3 generator simply looks up the reference table position using
the reference ID string and only encodes the position in the AMF3.
When parsing such a generated AMF3 data stream, the parser no
longer has the information about the custom reference IDs. It
generates reference IDs that begin with “#” and a continuous number
that directly reflects the position in the string reference table.
344 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
AMF3-specific XML
tags at a glance
Here is an overview of all the AMF3-specific XML tags, including their
attributes. See the list following the table for reference ID explanations.
XML Nodename
Value Location
1: Optional
Additional
attributes
Amf3 Subnodes ---
Null --- ---
Undefined --- ---
False --- ---
True --- ---
Integer Content of node ---
Double Content of node ---
String Content of node val-refId (1;2), valref
(1;2)
XMLDocument Content of node refId (1;3), ref (1;3)
Notes
Date Content of node refId (1;3), ref (1;3) Unlike AMF0,
AMF3 dates are
shown in UTC time
format.
Array Subnodes refId (1;3), ref (1;3) Subnodes may
have a name
attribute (1;2)
Object Subnodes dynamic (1;5),
externalizable (1;5),
classname,
classname-refId
(1;2), classname-ref
(1;2), traits-refId
(1;4), traits-ref (1;4)
Subnodes may have
a name attribute
(1;2) if object is
dynamic
Member Content of node refId (1;2), ref (1;2) May only appear as
a subnode of an
object
XML Content of node refId (1;3), ref (1;3)
ByteArray Content of node refId (1;3), ref (1;3) Content is Base64-
encoded
SilkPerformer User Guide 345

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
2: String reference
3: Object reference
4: Traits reference
5: Boolean value (“true” or “false;” default is “false.”)
Creating AMF3 SilkPerformer Projects
To configure scripting for an AMF3 Application, as with AMF0 projects, create
a project based on the Web Browser -> Flash Remoting project type. Due to the
tight integration of AMF0 and AMF3 support, this project type works for both
AMF0 and AMF3.
Recording
Common patterns in
AMF3 traffic
To record a script for an AMF3 Application, as with AMF0 projects, use the
Model Script dialog.
You will typically only encounter a few object types that are sent as remote
procedure calls with their corresponding responses.
The two most common types of objects that are sent from a Flex application to
the server are:
• flex.messaging.messages.CommandMessage (sent by the Flex
application on startup)
• flex.messaging.messages.RemotingMessage (directly associated with
user actions such as button clicks)
A CommandMessage’s XML structure typically looks like this:
operation
correlationId
body
clientId
timeToLive
destination
messageId
headers
timestamp
5
346 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
0
93C26579-723C-8375-6A63-0E2F350D5C02
nil
0
The basic structure (the AmfXml and Msg tags) is the same as AMF0, and the
AMF3-specific part is encapsulated within the AMF3 tags. Below that, you can
see the flex.messaging.messages.CommandMessage object.
RemotingMessages are quite similar:
source
operation
body
clientId
timeToLive
destination
messageId
headers
timestamp
getCollection
0
mathservice
7966B5A2-EF2C-B8C3-8FE3-0E2F34BFAD68
my-amf
SilkPerformer User Guide 347

17 OTHER SUPPORTED ENVIRONMENTS
Adobe Flex AMF3 Support
BB7EB7E4-DA1C-4595-61B1-
AF035EA08054
0
Responses to CommandMessages and RemotingMessages are usually objects of
the type flex.messaging.messages.AcknowledgeMessage:
;jsessionid=FD9408A5721C2FA76BDABBA5F5C56A09
destination
headers
correlationId
messageId
timestamp
clientId
timeToLive
body
DSId
BB7EB7E4-DA1C-4595-61B1-AF035EA08054
93C26579-723C-8375-6A63-0E2F350D5C02
BB7EB7E4-DA3E-FADA-E6D5-1DEDDD67C705
1202834762618
BB7EB7E4-DA2D-84CA-B275-089186BC83A6
0
348 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Customizing Oracle Applications 12i Scripts for Replay
Customizing AMF3
scripts
Warning Do not change the overall structure of XML objects
within AMF3 scripts. It is okay to parse values or insert verification
function, but deleting or rearranging AMF3 XML elements will
destroy a call. The order of an object’s member values directly
depends on the member names. Also do not change the order of
elements within arrays.
Customizing Oracle Applications 12i Scripts for Replay
Prerequisite for
recording
Customizing recorded
scripts
Before recording Oracle Applications 12i you must disable ports 9000-9005.
Before you can replay a recorded Oracle Applications 12i script you need to
customize the script.
The parameter icx_ticket contains session information for Oracle Applications
12i. The SilkPerformer recorder generates code to automatically parse the value
of icx_ticket:
WebParseDataBoundEx(gsIcxTicket, STRING_COMPLETE,
"icx_ticket='", 1, "'", WEB_FLAG_IGNORE_WHITE_SPACE, 1);
WebPageParseUrl("IMG SRC", "IMG SRC=\"", "or", WEB_
FLAG_IGNORE_WHITE_SPACE);
WebPageUrl("http://myserver.com:8005/OA_HTML/
runforms.jsp", "Oracle Applications R12", OA_HTML_
RUNFORMS_JSP003);
The WebParseDataBoundEx function tells the SilkPerformer runtime to parse
the value of icx_ticket from the next HTML page call into the gsIcxTicket
variable.
In this example, the next HTML page call is http://myserver.com:8005/OA_
HTML/runforms.jsp.
This page takes input parameters from the form OA_HTML_RUNFORMS_
JSP003. If this form already contains a recorded value for icx_ticket, the value
has to be set to an empty value.
SilkPerformer User Guide 349

17 OTHER SUPPORTED ENVIRONMENTS
Customizing Oracle Applications 12i Scripts for Replay
For all subsequent usages of the icx_ticket parameter, the value stored in the
gsIcxTicket variable must be used instead of the recorded value.
Note The SilkPerformer recorder scripts customized code for the
OraFormsConnect function by default. Verify that the customization
has been performed correctly.
Additional session
information
customization
Additional customizations are necessary after you have customized the icx_
ticket parameter. TrueLog Explorer can assist you with these customizations.
Procedure To customize Oracle Applications 12i session information:
1 Ensure that the icx_ticket parameter is customized correctly.
2 Try Run the script in SilkPerformer with TrueLog generation enabled.
3 Open the resulting Try Script TrueLog with TrueLog Explorer. Typically,
an HTML page (one appearing soon after login) will show errors.
4 Click the Analyze Test button on the Workflow Bar.
5 Select Compare your test run from the Workflow - Analyze Test wizard
page.
6 Close the Step through Truelog dialog.
7 Select the TrueLog node that corresponds to the recorded application’s
login page.
8 Right-click the node and select Synchronize Truelogs from the context
menu.
9 Select Edit / Truelog Type / Web from the TrueLog Explorer menu bar.
350 SilkPerformer User Guide

17 OTHER SUPPORTED ENVIRONMENTS
Customizing Oracle Applications 12i Scripts for Replay
10 Select the Source Differences tab.
11 Right-click any rows that show differences between the recorded
TrueLog and the replay TrueLog that also occur in the script and
therefore likely contain session information (highlighted in yellow) that
should be customized.
12 Run another TryScript run in SilkPerformer to verify that your session
information customizations are successful.
SilkPerformer User Guide 351

17 OTHER SUPPORTED ENVIRONMENTS
Customizing Oracle Applications 12i Scripts for Replay
352 SilkPerformer User Guide

18
Trouble 18 Chapter
Shooting
Introduction
What you will learn
This chapter explains known issues and trouble shooting tips for SilkPerformer.
This chapter contains the following sections:
Section
Page
Overview 353
Contacting technical support 353
Controller Agent Communication Troubleshooting 354
Error Message Overview 356
Overview
All Borland customers whose products qualify for maintenance and all
prospective customers who are evaluating Borland products are eligible for
Technical Support.
Contacting technical
support
If you have a question on recording Web traffic, load testing scenarios or any
other product or implementation issue, our highly trained technical support staff
will respond to your requests as quickly and professionally as possible.
To reach Borland Technical Support:
- Visit http://support.borland.com
Required information
Please have the following information ready when you call:
- Product name and version number
- Operating system name and version
SilkPerformer User Guide 353

18 TROUBLE SHOOTING
Controller Agent Communication Troubleshooting
- Exact wording of any error messages involved
- Customer ID found on your shipping receipt for SilkPerformer or on
an index card found in your software box
Controller Agent Communication Troubleshooting
The following issues may occur when a controller computer attempts to connect
to agent computers:
“Load Test Controller - 3211: Could not connect agent control service.”
This problem may be caused by one of the following reasons:
The agent control service is not running.
Solution: Start the agent control service.
A port mismatch occurs (i.e., the port that the agent control service is listening to
and the port at which the controller wants to connect the agent control service
are different).
Solution: Adjust the agent connection settings.
The agent is connected through a proxy and the proxy is down.
Solution: Check the agent connection settings and restart the proxy.
“Load Test Controller - 3212: Could not authenticate the Agent Control”
Service."
The agent requires a password and encryption is enabled. The agent’s server
certificate could not be found or a required dynamic link library (.dll) is missing
(e.g., "ssleay32.dll"). However, note that the agent control service does not
report an error when it is started, if this is the problem.
Solution: Your best option is to re-install the remote agent, because missing
certificates or dynamic link libraries are indications that the agent was not
installed correctly.
“Load Test Controller - 3213: The agent control service could not create
the agent.”
This is the same type of error as 3212, except that the agent does not require a
password.
“Load Test Controller - 3233: The agent's version does not match the
version of its controller.”
Solution: Install the correct remote agent version.
“Load Test Controller - 3303: The local results directory could not be
created.”
This is an access rights problem, that is, the user account of the agent process
does not have the necessary access rights to create a directory; the parent
354 SilkPerformer User Guide

18 TROUBLE SHOOTING
Controller Agent Communication Troubleshooting
directory of the desired directory is probably a read-only version for all
accounts.
Solution: Grant the necessary access rights.
“Load Test Controller - 3304: The local results directory could not be
cleared.”
A file within the local results directory is still open and being used by another
process (most likely the perfLtcAgent.exe and perfRun.exe). This is usually a
hint that either the agent or the runtime processes did not shutdown at the end of
a previous load test, which should not happen.
Solution: Check for pending perfLtcAgent.exe or perfRun.exe files and
terminate them manually.
If this is not the case, check that other programs are not using an open file from
the local results directory. If none of the above options solve the problem, reboot
the remote agent’s machine.
“Load Test Controller - 3305: The local data directory could not be
created.”
The reason and solutions to this problem are the same as for error 3303.
“Load Test Controller - 3306: The local data directory could not be
cleared.”
The reasons and solutions to this problem are the same as for error 3304.
“SilkPerformer - 32: Access denied, authentication required.”
The agent expects a password, but the controller did not send a password,
meaning the controller did not authenticate itself.
Solution: Adjust the agent connection settings.
“SilkPerformer - 33: Authentication failed (wrong password).”
The agent expects a password but the password sent by the controller is
incorrect.
Solution: Adjust the agent connection settings.
“SilkPerformer - 25: Limit of concurrent instances exceeded.”
Two agents cannot run simultaneously. Most likely, the agent is involved in a
load test started by another controller.
Solution: Wait for the other controller to finish its load test and then try again.
If this is not the problem, the agent did not terminate correctly at the end of a
previous load test and you must terminate the agent process manually.
“System - 2: The system cannot find the file specified.”
The agent executable (perfLtcAgent.exe) could not be found or there is a
different entry for the agent executable in the performerLocal.ini file.
Solution: Check the performerLocal.ini file. If it is not there or if the
perfLtcAgent.exe is not present, then re-install the agent.
SilkPerformer User Guide 355

18 TROUBLE SHOOTING
Error Message Overview
“Load Test Controller - 3422: Command timed out.”
Usually the controller has 30 seconds to:
- Connect to the remote agent control service.
- Pass authentication information to the remote agent control service
(if required).
- Ask the agent control service to create the agent.
- Connect to the agent.
- Exchange version information with the remote agent.
These tasks may take more time than the default timeout of 30 seconds,
especially if the agent is connected through a proxy or if the connection is slow.
Solution: Specify a higher timeout in the system settings (System Settings/
Control/Remote Agent Connection).
In the remote performerLocal.ini file, the "port" entry is missing in the agent's
AcsClient/type section. That is, the agent was started by the agent control
service, but it is not listening to any port and therefore cannot be connected.
Solution: Specify the "port" entry in the agent’s AcsClient/type section in the
remote performerLocal.ini file.
“Load Test Controller - 3223: Agent connection lost.”
This occurs when the connection is broken between the controller and agent or
when the underlying communication units of the load test controller report
connection problems. Furthermore, connecting the agent through a proxy might
be the cause of the error if the proxy itself is buggy.
Solution: There should be log files named orbClient_.log,
orbServer_.log in the controller’s or agent’s home directory,
respectively, where logging information related to the error can be found.
Error Message Overview
SilkPerformer system-related errors occur either during script compilation or
during load test execution and can cause load tests to fail. As they occur,
SilkPerformer reports system-level errors on the screen. Severe system-level
errors are reported to the Windows NT Event Log.
SilkPerformer distinguishes among the following types of system-level errors:
• “Compiler errors”
- “Lexical errors” (LEX errors)
- “Semantic errors” (SEM errors)
- “Syntax errors” (SYN errors)
• “Internal errors” (COM errors)
356 SilkPerformer User Guide

18 TROUBLE SHOOTING
Error Message Overview
• “System limitation errors” (RES errors)
• “Runtime errors” (RUN errors)
Each error message is identified by a sequential error number (unique within an
error category) plus a three-character code that indicates the category of error,
followed by a string providing a detailed description of the error.
See the BDL Function Reference for details on error-handling functions.
Compiler errors
Lexical errors
If an error occurs during compilation, the SilkPerformer compiler stops at the
position where the error was detected. The error message is displayed and the
line causing the error is highlighted. The compiler distinguishes among three
categories of errors:
- “Lexical errors” (LEX)
- “Syntax errors” (SYN)
- “Semantic errors” (SEM)
A lexical error (error type LEX) occurs if the compiler detects lexical symbols
that are not allowed in Benchmark Description Language.
Example
If a script contains an invalid character such as '{', compilation is aborted and an
error message is printed:
1 benchmark WebTest
2
3 dcluser
4 user {
Output
compile error LEX 13: not a lexical symbol of BDL
--- compilation was not successful ---
Syntax errors
A syntax error (error type SYN) occurs if the syntax of a test script does not
agree with the syntax of Benchmark Description Language.
Example
If you do not specify the keyword then within an if statement, an error message
appears:
8 transaction TMain
9 var
10 j : number init 1;
11 i : array [500] of number;
12 s : array [5] of string(10);
13 begin
14 if j=1
15 j:
Output
SilkPerformer User Guide 357

18 TROUBLE SHOOTING
Error Message Overview
compile error SYN 38: 'then' expected
--- compilation was not successful ---
Semantic errors
System limitation
errors
Runtime errors
Internal errors
A semantic error (error type SEM) occurs if semantic conditions do not comply
with the requirements of Benchmark Description Language.
Example
If you use an undeclared variable (“i” in the example below), compilation is
aborted and an error message is displayed:
46 dcltrans
47 transaction TMain
48 var
49 artname : string(40);
50 artno, price : number;
51 result set c1;
52 begin
53 writeln; write("transaction TMain:"); writeln;
54 c1: SelArticle(out artno, price, artname);
56 write("rows found: "); write(rows(c1));
57 fetch c1 all;
58 for i:
Output
compile error SEM 12: variable is not declared
--- compilation was not successful ---
System limitation errors (error types RES) can occur while either compiling a
script or running a load test. If they occur during a test execution, they will cause
the test to fail. These errors indicate that a system limitation has been exceeded.
Only the SilkPerformer simulator will produce run-time errors (error type
RUN). Run-time errors indicate programming errors in test scripts that cannot
be recognized during compilation (for example, using array indexes outside the
scope of an array). If a run-time error occurs, an error message is printed on
screen and to a .rpt file.
SilkPerformer can detect internal script errors (error type COM). Please contact
Borland technical support if you encounter one.
358 SilkPerformer User Guide

Index
Symbols
.NET 19, 20, 104, 128, 269
.NET Explorer 269
.NET Framework 269
Remoting objects 269
@codepage 310
A
Abandon option 85
Accepted Baseline Reports 182
ADO 19, 114, 128
Adobe Flex 338
Advanced Flex Data Services 338
Agents 121
Agent Pool 32
Connection properties 51
Monitoring 195
Multiple 165
Pool settings 33, 48
Remote 48
AJAX 311
Alphanumeric characters 91
AMF3 337
Creating projects 346
Data types 339
References 343
XML tags 345
Animated checkbox 136
Application profiles 39, 40, 41
Application types, available 17
Arguments, program 42
ARM options 73
Arrays 170
ASCII 308
Authentication, browsers 85
B
Bandwidth 75
Baseline performance 177, 179
Establishing 25
Baseline Reports 26, 180, 182
Benchmark Description Language 213
Borland Optimizeit ServerTrace Plug-In 265
Boundaries, fixed 170
Bounds
Variable, upper 170
Browsers 144
authentication 85
Browser-level API 89
Options 79
Settings 83
C
Candle graphs 227
Certificates, client 52
Character encodings 309
Charts 214
Custom 214
Options 225
CIDR 145
Citrix 44, 102, 128
Citrix MetaFrame 19
Measurements 243
Client certificates 46, 52, 78
Exporting 54
Client-side real-time measures 75
Code completion 249
Code List, Code completion 249
Codepage 310
COM 19, 44, 128
COM/ADO 114
Command line
Options 109
Commandline
Launching tests 205
Component testing, remote 273
Computers, agent 121
Concurrent connections 232
Configuring services 40
Configuring SilkPerformer 31
Connect timeout 76
Connection parameters 197
Content-encoding 81
Control settings 36
Controller connection, configuring 48
Cookies 148, 150
persistent recording 89
CORBA 63, 105, 241
Measurements 241
Counter logs 259
CPU utilization 232
CSV files
Converting to Time Series Data format 259
SilkPerformer User Guide 359

csv2tsd.exe 259
Customizing test scripts 139
D
Data files 120
Data files, multi-column 254
Data Source Wizard 196
Data types
Structured 172
Data verification 93
Databases 18, 63, 238
Measurements 242
DB2 CLI 44
Digest verification 92
Directories 32
Settings 34
DLLs 166
DNS lookup 88
Document cache 87
Document download time 239
Downstream traffic 76
Duplex 76
dynaTrace Diagnostics plug-in 267
E
eCATT, SAP 312
Eclipse plug-in 325
Installing 326
Viewing test results 332
Working with projects 328
EJB 18, 43
Emailing projects 247
Embedded objects data 239
Emulate DNS lookup 88
Encodings, character 309
Encryption 49
Enterprise JavaBeans 269
ERP/CRM 17
Error code 66
Error tolerance 85
Error-handling 65
Errors, SilkPerformer 353
EUC-JP 309
Evaluating SilkPerformer 4
Exploring TrueLogs 136
Exporting projects 246
External functions 166
F
File locking 248
Files, data 120
Firewalls 49
HTTP 50
Multiple 52
SOCKS 51
TCP/IP 49
Fixed boundaries 170
Flash Remoting 17
Flex 337
Creating projects 346
Data types 339
References 343
XML tags 345
FTP 45, 128
Passive semantics 76
Function declarations 141
Functions
External 166
Matrix 174
Parsing 153
G
Graphs 222
Green screen support 314
GUI-level testing support 317
Configuring Terminal Services settings 321
Configuring Windows 320
Functions 318
H
Host name 46
Host options 77
HTML 175
Page Timers 137
HTTP 45, 81, 234
Firewalls 50
HTTP body transformation 93
HTTPS 46
I
IBM 314
IDL files 108
IIOP 44, 63, 105, 128, 241
Measurements 236
Importing projects 246
Indention, smart 251
Informational 66
Input data, customized 252
Interfaces
Server 44
360 SilkPerformer User Guide

Internet Explorer 42, 54
Internet options 75
IP addresses 46, 144
Issues, known and workarounds in
SilkPerformer 304
J
Jacada 43
Java 19, 20, 129, 249, 269, 273
Advanced options 109
API 43
Compiling files 258
Configuration 55
EJB 43
Integration with SilkPerformer
Java Framework 256
Java Explorer 271
Java Framework 271
Java Virtual Machine 109
JNDI 43
Logging options 110
Options 108
Projects in SilkPerformer 257
RMI 43
JIT compiler 109
Jolt 128
JSON
Pretty-formatted 311
JUnit 19, 273
K
Keep-alive semantics 81
Known issues in SilkPerformer 304
L
LAN 33, 49
Latin-1 309
Layout
Reports 217
Settings 35
LDAP 20, 45, 128
Licensing
SilkPerformer 4
Link checking 92
Listen port 46
Load tests
Defining workload 27
Executing 28
Launching from commandline 205
Starting 194
Loading times tolerance 84
Locking, files 248
Log files 136
M
Manual script writing 131
Mapi 20
Measurements 230
Memory 232
Memory footprints by application type 7
Message output 264
MIBs 196
Middleware measurements 242
Modems 75, 144, 178
Compression rate 76
Module 65
Monitoring
Data 221
Requirements 199
Servers 183
Multi-Byte Character Set (MBCS) 309
Multi-column data files 254
N
Netscape 42, 55
Network options 75
NUnit 19, 273
O
OCI 44
ODBC 44, 111
OID 197
Options
Recording 107, 114
Verification 90
Oracle 17, 128
Oracle Forms 18, 43, 97, 128
Oracle Applications 12i 17
Customizing scripts for replay 349
ORB features support 107
Outlining projects 16
Output files 138, 178
Output messages 264
Overlay series 222
Overview Report 229
Overview Summary Reports 213
SilkPerformer User Guide 361

P
PAC 81
Page data 239
Page timers 138
Page-based browser-level API 88
Parameter Wizard 252
Parameters
Passing 168
Test 5
Parsing functions 153
Passwords 86
Authentication 49
Performance
Monitoring 195
Test baseline 179
Performance and Scalability Matrix 5
Performance counter logs 259
Performance Data Collection Engine 213
Performance Explorer 75, 212, 213, 229, 259
Converting CSV data for import 259
Performance, baseline 25, 177
Persistent-cookie recording 89
Plug-Ins
Initializing 264
POP3 46
Port settings 48
Preprocessor arguments 108
Printable characters 91
Profile settings 60
.NET options 104
Administering Profiles 60
Citrix options 102
COM/ADO options 114
CORBA options 105
Database options 110
IIOP options 105
Internet options 75
Java options 108
Oracle Forms options 97
Recording options 62
Replay options 64
Result options 66
SAP GUI options 99
Script options 62
Simulation options 64
Terminal Client options 94
WAP options 117
Web options 79
Profile settings, projects 60
Profiles
Applications 41
Users 144
Program arguments 42
Project profile settings 60
Projects
Configuring 60
Emailing 247
Importing/exporting 246
Outlining 16
Uploading to SilkCentral 247
Protocol selection 42
Protocol-specific recording options 63
Proxies 39, 86
Options 82
Servers 50
Settings 40, 45
Proxy Auto-Config 81
PVCS 297
Q
Queuing workload 64
R
Random variables 253
Real-time measures 75
Recorder 45
Matrix 174
Protocol-specific options 63
SilkPerformer 39
Recording
Options 62, 88, 107, 114
Reload times 84
Remedy 129, 333
SilkEssential 336
Remote agents, configuring 48, 165
Remote-component testing 273
Rendezvous functionality 164
Replay 77
Errors 66
Options 64
Via a scripting function 54
Within the GUI 54
Reports 215, 229
Custom 216
Files 137
Layouts 217
Request data 240, 243
Requests sent 232
Response data 175, 240
Response data verification 158
Response time 243
Thresholds 182
Transactions 137
Responses failed 233
362 SilkPerformer User Guide

Responsiveness 232
REST 311
Results 33
Analyzing 29
Custom 219
Exploring, test 209
Options 66
Overview 209
Settings 38
Viewing facilities 210
RMI objects 269
Root cause analysis, visual 212
S
Sample scripts 19, 131
SAP 17
eCATT 312
SAPGUI 17
SAP GUI 99
Scalability
Performance and Scalability Matrix 5
Scaling options 223
Scripts 119
Customizing 24, 139
Modeling 21
Samples 19, 131
Trying out 23, 132
Validating 22
Writing manually 131
Security options 77
Security, encryption 49
Server busy time 239
Server interfaces 44
Server monitoring 183
Requirements 199
ServerTrace 266
Services
Configuring 40
Session handling 149
Session IDs 152
Settings
Browser 83
Severity 65
Shift-JIS 309
Siebel 17
Silk Performance Explorer 259
SilkCentral
Test Manager 211, 287
Uploading projects to 247
SilkEssential
Remedy 336
SilkPerformer
Configuring 31
Evaluation version 4
Licensing 4
Overview 1
Recorder 39, 45
Upgrading to 4
What’s new 2
Workbench 32
SilkPerformer 2006 R2
What’s new 2
SilkTest
GUI-level testing support 317
Simulation 65
Options 64
Settings 83
Smart indention 251
SMTP 46, 128
SNMP 197
SOCKS 42, 46, 47, 49, 128
Firewalls 51
Proxy server 51
Source code control integration
SilkPerformer 295
Source control integration 33
MS VSS 296
PVCS 296
SilkPerformer 295
StarTeam 296, 298, 299
SQL 63, 237
SQLSTATE 110
SSL 49, 78
StarTeam
Source control integration 296, 298, 299
Stress test 64
Structured data types 172
Subnet masks 145
Summary Reports 213
Summary tables 137
Synchronization 163
Source control status 303
System classpath 109
System Configuration Manager 48, 49, 145, 160
T
Table verification 92
TCP/IP 44, 46
Firewalls 49
Templates
Monitoring 196
Custom 184
Reports 217
Terminal Client 94
Terminal emulation 17, 314
Terminal Emulation Support
SilkPerformer User Guide 363

TN3270 314
TN5250 314
VT100 314
VT200 314
Test baseline
Establishing 26
Performance 179
Test Manager
Integration 210
Test Manager, SilkCentral 33, 287
Test parameters 5
Test results
Exploring 209
Overview 209
Viewing facilities 210
Test scripts 119
Customizing 24, 139
Modeling 21
Samples 131
Trying out 23, 132
Validating 22
Writing manually 131
Tests
Executing, load 28
Launching from commandline 205
Remote components 273
Results, analyzing 29
Starting 194
Text and grid color 223
Think time
Limited 64
Random 64, 177
Thresholds
Response time 182
Time series data 226
Traffic 76
Upstream 76
Transaction declarations 141
Transaction response times 137
Transactions measurement group 238
Trouble shooting, SilkPerformer 353
TrueLog
Exploring 136
Setting options 70
TrueLog Explorer 22
TrueLog Explorer 135, 140, 150, 157
Function matrix 174
TrueLog On Error 190, 212
TSD (Time Series Data) 259
TSD-Writer 228
Tuxedo 18, 44, 128, 243
U
Unicode 308
Unicode Text/Hex Editor 311
Unit testing 19, 273
Uploading projects to SilkCentral 247
Upper bounds, variable 170
Upstream traffic 76
User distribution 190
User groups 142
Customization 143
User profiles 144
User tolerance 84
User types 137, 178
UTF-16 309
UTF-8 309
V
Validating test scripts 22
Variables, random for virtual users 253
Verification
Response data 158
Verification under load 211
Verifications
Data 93
Digest 92
Options 90
Table 92
Virtual User Summary report 135
Virtual users 137, 142
Monitoring 195
Output files 229
Parameter Wizard 252
Stopping 65
Vista, Windows 308
Visual analysis of TrueLogs 135
Visual SourceSafe 297
Visual Studio .NET, Add-In 270
W
WAN 49
Agents 51
Networks 33
WAP 117
Measurements 240
Warning 66
WChar 106
Web 136, 249
Settings 42
Web application testing, AJAX 311
Web form
364 SilkPerformer User Guide

Measurements 239
Web forms 138, 147, 181
Web page measurements 239
Web Services 18, 20
WebDAV (MS Outlook Web Access) 17
WebSphere Application Servers 3.5 / 4.0 200
Windows Vista 308
WinSock 42, 128
Workbench, SilkPerformer 245
Workload 187
Available models 185
Customizing 142
Defining 27, 121
Queuing 64
Workspace 33
Settings 38
Writing test scripts manually 131
WString 106
X
XML 20
Pretty-formatted 311
Verification 92
XML/SOAP 18
Z
ZIP archives 246
SilkPerformer User Guide 365

366 SilkPerformer User Guide