Industry Standards and Performance Benchmark of Cognos Reports

[Suraj]

We are discussing on a topic of 'Industry Standards and Performance Benchmark of Cognos Reports'.
Different reports run have different performance depending upon report design, database, server configuration etc...
Regardless of those conditions, are there any standards or benchmark that we should go after?
Please discuss your current report performance/benchmarks. What should be the target performance?
Should we set 20s maximum for any report to run or 1minute?

What is the normal report run time in your company and is that satisfactory?
Cognos audit reports give execution time of a report, but is there a way to find out what part is taking long to process when a report is running slow?
Is execution time only for query execution time or does it include report presentation/rendering time?
Thank you all for your input.  ???

[almeids]

Suraj-
   When developing a new interactive report my goal is 20 seconds, though when facing complex requirements and/or heavy volume or a lot of runtime aggregation I'll consider an overall runtime of 40 seconds acceptable (including time to present the prompts).  In all cases I include pdf (our preferred format) rendering time.
   For pregenerated output we don't worry so much as long as the runtimes don't hit our radar - which generally would necessitate a noticeable delay (in starting our next batch cycle, for example).
   I'm not sure to what degree the statistics subdivide the processing as I mostly use them to monitor what's being used and what isn't.  When it comes to the detailed analysis I do it the hard way on a report by report basis, by running the tabular models separately (still on ReportNet!) or running the sql directly against the database, by comparing html vs. pdf vs. excel runs, and so forth.
   HTH...
Steve

[almeids]

I just took a look at our average runtimes for some of the interactive reports and though many are still in the under-30-seconds range, some have crept up and are 50-60.  We haven't had any complaints but that may be because of a gradual degradation (as some of of our tables have built up more history for example), or perhaps because the big hitters that are pulling up the average have lower expectations due to the volumes they are looking at...

[Suraj]

Thank you for replying Steve.
What I am trying to do is give a target performance criteria to our authors so that when they develop a report, they should be able to run under that time. If that is not the case, then they should explain the reasons behind it. That way we can fix the back end or create new tables/calculations etc after making sure the reports themselves are good.
But the whole target for performance is a gray area as it depends upon multiple areas.
It'll be easier to set a target after knowing what is the current situation in various companies.
Another, area of this is querying the audit tables for execution time. Does this time only refers to query execution time in the db or does it include everything from the request received time to report presentation in the browser?
thanks for opinion.