19.06.2013 Views

DB2 UDB for z/OS Version 8 Performance Topics - IBM Redbooks

DB2 UDB for z/OS Version 8 Performance Topics - IBM Redbooks

DB2 UDB for z/OS Version 8 Performance Topics - IBM Redbooks

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Table 8-1 shows extracts from <strong>DB2</strong> PE statistics reports <strong>for</strong> an Online Transaction Processing<br />

workload (OLTP). The two columns of numbers within each z/<strong>OS</strong> represent values per<br />

transaction <strong>for</strong> data sharing member 1 and member 2.<br />

Table 8-1 CF Request Batching - <strong>DB2</strong> PE extract (OLTP)<br />

It can be clearly seen from Table 8-1 that CF request batching was not used in z/<strong>OS</strong> 1.3. The<br />

counters <strong>for</strong> “Write & Register Multi”, “Pages Write & Reg Multi” and “Read <strong>for</strong> Castout Multi”<br />

are all zero in z/<strong>OS</strong> 1.3.<br />

CF request batching was used under z/<strong>OS</strong> 1.4:<br />

► An average 0f 6.9 pages were written per WARM command<br />

“Pages Write & Reg Multi” / “Write & Register Multi”=4.09 / 0.59<br />

= 6.9<br />

► An average of 7.7 pages were read per RFCOM command.<br />

(“Pages Castout” - “Read <strong>for</strong> Castout”) / Read <strong>for</strong> Castout Multi”<br />

= (4.61 - 0.01) / 0.60<br />

= 7.7<br />

Table 8-2 shows extracts from RMF reports <strong>for</strong> the same OLTP workload.<br />

Table 8-2 CF Request Batching - RMF extract (OLTP)<br />

As we can see from Table 8-2, the test using z/<strong>OS</strong> 1.4 results in significantly fewer requests to<br />

the coupling facility (9,880 compared to 14,082 <strong>for</strong> this workload). This is the direct result of<br />

CF request batching. We can also see the service times <strong>for</strong> both synchronous requests and<br />

asynchronous requests to the coupling facility have increased slightly. On average, the<br />

coupling facility is doing more work per each request, as a number of page movements may<br />

need to be processed with each request. We also notice the coupling facility utilization has<br />

322 <strong>DB2</strong> <strong>UDB</strong> <strong>for</strong> z/<strong>OS</strong> <strong>Version</strong> 8 Per<strong>for</strong>mance <strong>Topics</strong><br />

z/<strong>OS</strong> 1.3 z/<strong>OS</strong> 1.4<br />

Member 1 Member 2 Member 1 Member 2<br />

Pages Castout 3.91 4.35 3.71 4.61<br />

Write & Register 6.28 6.29 2.22 2.22<br />

Write & Register Multi 0.00 0.00 0.59 0.59<br />

Pages Write & Reg Multi 0.00 0.00 4.09 4.09<br />

Read <strong>for</strong> Castout 3.91 4.35 0.00 0.01<br />

Read <strong>for</strong> Castout Multi 0.00 0.00 0.48 0.60<br />

z/<strong>OS</strong> 1.3 z/<strong>OS</strong> 1.4<br />

Requests / sec 14,082 9,880<br />

Sync requests<br />

Serv time<br />

(usec)<br />

Async requests<br />

Serv time<br />

(usec)<br />

% of all req 17.7 95.5 19.5 86.6<br />

% of all req 173.9 4.4 192.8 13.3<br />

CF utilization (%) 21.7 21.0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!