Cisco Unified Contact Center Enterprise Solution Reference ...
Cisco Unified Contact Center Enterprise Solution Reference ...
Cisco Unified Contact Center Enterprise Solution Reference ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Load Balancing <strong>Cisco</strong> <strong>Unified</strong> CallManager<br />
Load Balancing <strong>Cisco</strong> <strong>Unified</strong> CallManager<br />
10-10<br />
<strong>Cisco</strong> <strong>Unified</strong> <strong>Contact</strong> <strong>Center</strong> <strong>Enterprise</strong> 7.x SRND<br />
Chapter 10 Sizing <strong>Cisco</strong> <strong>Unified</strong> CallManager 4.x and 5.x Servers<br />
An additional benefit of using the 1:1 redundancy scheme is that it enables you to balance the devices<br />
over the primary and backup subscriber pairs. Normally (as in the 2:1 redundancy scheme) a backup<br />
server has no devices registered unless its primary is unavailable.<br />
With load balancing, you can move up to half of the device load from the primary to the secondary<br />
subscriber by using the <strong>Cisco</strong> <strong>Unified</strong> CallManager redundancy groups and device pool settings. In this<br />
way, you can reduce by 50% the impact of any server becoming unavailable.<br />
To plan for 50/50 load balancing, calculate the capacity of a cluster without load balancing and then<br />
distribute the load across the primary and backup subscribers based on devices and call volume. To allow<br />
for failure of the primary or the backup, the total load on the primary and secondary subscribers should<br />
not exceed that of a single subscriber. For example, MCS-7845 servers have a total server limit of 500<br />
<strong>Unified</strong> CCE agents. In a 1:1 redundancy pair, you can split the load between the two subscribers,<br />
configuring each subscriber with 250 agents. To provide for system fault tolerance, make sure that all<br />
capacity limits are observed so that <strong>Unified</strong> CCE agent phones, <strong>Unified</strong> IP phones, CTI limits, and so<br />
on, for the subscriber pair do not exceed the limits allowed for a subscriber server.<br />
<strong>Cisco</strong> recommends the equal distribution of all devices and call volumes as much as possible across all<br />
active subscribers. For instance, distributing the <strong>Unified</strong> CCE agents, CTI ports, gateways, trunks,<br />
voicemail ports, and other users and devices among all subscribers equally, minimizes the impact of any<br />
outage.<br />
For additional information on general call processing topics such as secondary TFTP servers and<br />
gatekeeper considerations, refer to the <strong>Cisco</strong> <strong>Unified</strong> Communications <strong>Solution</strong> <strong>Reference</strong> Network<br />
Design (SRND) guide, available at<br />
www.cisco.com/go/srnd<br />
Upgrading <strong>Cisco</strong> <strong>Unified</strong> CallManager<br />
The 1:1 redundancy scheme allows upgrades with only the failover periods impacting the cluster. The<br />
1:1 redundancy scheme enables you to upgrade the cluster using the following method:<br />
Step 1 Upgrade the publisher server.<br />
Step 2 Upgrade dedicated TFTP and music-on-hold (MoH) servers.<br />
Step 3 Upgrade the backup subscribers one at a time. This step will affect some users if 50/50 load balancing<br />
is implemented.<br />
Step 4 Fail over the primary subscribers to their backups, and stop the <strong>Cisco</strong> <strong>Unified</strong> CallManager service on<br />
the primaries. All users are on primaries and are moved to backup subscribers when the <strong>Cisco</strong> <strong>Unified</strong><br />
CallManager service is stopped. CTI Manager is also stopped, causing the Peripheral Gateway (PG) to<br />
switch sides and inducing a brief outage for agents on that particular node.<br />
Step 5 Upgrade the primaries, and then re-enable the <strong>Cisco</strong> <strong>Unified</strong> CallManager service.<br />
With this upgrade method, there is no period (except for the failover period) when devices are registered<br />
to subscriber servers that are running different versions of the <strong>Cisco</strong> <strong>Unified</strong> CallManager software. This<br />
factor can be important, because the Intra-Cluster Communication Signaling (ICCS) protocol that<br />
communicates between subscribers can detect a different software version and shut down<br />
communications to that subscriber.<br />
OL-8669-05