Processing Capacity
1.0The appropriate value for this setting is a function of the number of CPUs, CPU clock rate, memory, etc in relation to other servers in the topology. It is a relative setting. This setting dictates the number of BI requests that will be handled by a particular a server in the Cognos 8 topology.
By default, requests will be distributed so that each server is expected to handle an equivalent number of requests.
While it may be difficult to estimate a precisely correct set of values for heterogeneous server environments, there are several methods that can improve the relative accuracy of this task.
1) Use the relative performance ratings of a server (provided by hardware vendors).
2) Estimate the process capacity for a server using one variable at a time in the order of (#of CPUs, clock rate of CPUs)
It is important to monitor the request rate through the Server Administration tool of Cognos Connection. Isolate the Report services to monitor interactive usage. This will indicate whether the applied process capacity settings lead to expected behavior. Key metrics to observe include Latency and Requests per second. If latency is building up on one report service while others are functioning freely, reduce the process capacity on the high latency server. In general, monitor the requests per second metric so that the number is proportional to the capacity of each server in the topology.
