How does load balancing work in Cognos 8?
Load balancing spreads processing among all available processors which means that processing requests are distributed appropriately among all the available Cognos servers.
By default, Cognos balances request load automatically, so each server dispatcher processes the same number of requests. The automatic load balancing is weighted round robin.
Load balancing can be also customized by setting process capacity settings using server administration options. The CPU speed is a good indicator to set the balancing weight.
For instance, when there are 2 servers with following configuration:
a dual P4 3,8GHz with 2048MB RAM server
a single CPU 1Ghz with 1024MB RAM machine
Then it would be reasonable to assing a weight of 4 to the first server and a weight of 1 to the second one. Then Cognos will submit most of the request (80%) to the more powerful server.
In order to disable the Cognos 8 internal load balancing mechanism, the mode needs to be set to Cluster compatible.
Load balancing can be also customized by setting process capacity settings using server administration options. The CPU speed is a good indicator to set the balancing weight.
For instance, when there are 2 servers with following configuration:
Then it would be reasonable to assing a weight of 4 to the first server and a weight of 1 to the second one. Then Cognos will submit most of the request (80%) to the more powerful server.
In order to disable the Cognos 8 internal load balancing mechanism, the mode needs to be set to Cluster compatible.
Back to the IBM Cognos COG-122 certification materials