Tuning Server for Performance

[Ravisha]

Hello Sir/Madam,

I referred the documentation and also different articles on the below question. But, still I'm unable to understand the differences between them  :'(. I would really appreciate it if anyone could educate me on this :).

What are High Affinity connections (what is meant by "Connections"), Low Affinity Connections, Report Processes, Threads and BiBus. How are they different from one another? What's the logical flow?

Thanks in advance !  ;D

[sunosoft]

Hi,

Affinity means we can say like how efficiently a particular request will be executed by a particular server/process.
Lets take example of page down request when report executed in HTML format.
Till you get the HTML output your request might have executed by some process p1....now when u do page down if that request goes again back to process p1 then it will be executed more efficiently....if not..lets say it goes to some process p2 then process p2 has to perform all tasks that p1 have done till now and then ur next page request..here your request might take more time.
Such type of requests are called as the high affinity requests. On the other hand low affinity request will be executed with the same efficiency by any process/server. For example executing a report.

Now processes and BiBus are generally related to reportservice and batchreportservice. There will be certain number of processes per reportservice and batchreportservice. And certain number of connections per process. These settings depend on your hardware configurations. I am not sure though but what I understand is one process for reportservice or batchreporservice will yield one BiBus process in task manager. And when you set number of low affinity(2) and high affinity(1) connections for process it means that particular process can execute 3(2 low affinity+1 high affinity) requests simultaneously. Now thread what I understand is more related to operating system where one process in task manager can have multiple threads, which can execute different parts of request simultaneously and return result to process.

[Ravisha]

Hello Sir,

Thanks a lot for your time and patience for replying to my query. I really appreciate it.

We have a standalone installation of Cognos 10.1.1 running on windows server with dual core and having 8GB of RAM. I have a trouble of understanding
1) How many connections should I set for both Low and High Affinity
2) No of Processes for both Low and High Affinity request during Peak and Non-Peak hours

I would appreciate any help that you can offer.

Thanks

[sunosoft]

Hello Sir,

Thanks a lot for your time and patience for replying to my query. I really appreciate it.

We have a standalone installation of Cognos 10.1.1 running on windows server with dual core and having 8GB of RAM. I have a trouble of understanding
1) How many connections should I set for both Low and High Affinity
2) No of Processes for both Low and High Affinity request during Peak and Non-Peak hours

I would appreciate any help that you can offer.

First of all small request, please do not call as Sir.

Regarding your question about settings it all depends on usage of particular environment. Having said that if I would have been at your place with given hardware configuration I would have started with basic settings as below

Report Service Processes -2
High Affinity connections-1
Low affinity Connections-2

BatchReportSrvice Processes -2
High affinity connections-1
Low affinity connections-2

And I would have monitored the environment performance over the time and made changes accordingly.