how to improve the performance of mainframe jobs

HariD · Post by **HariD** » Tue Mar 23, 2010 4:40 am

Hi,

I have data coming from mainframe systems and we r going to improve the performance of the job (i.e Parallel job),is there logic to apply for improve the performace.

Any answer can be appreciated.

ArndW · Post by **ArndW** » Tue Mar 23, 2010 4:42 am

How is the data coming from the mainframe - that step might be your bottleneck, in which case any improvements to your DataStage jobs wouldn't increase throughput.

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Tue Mar 23, 2010 5:05 am

You have been posting open questions on "how to improve performance" in such-and-such environment without detailed information on your situation.

It will be useful if you supply what you have tried, tested and identified as potential bottlenecks as part of your queries.

ray.wurlod · Post by **ray.wurlod** » Tue Mar 23, 2010 4:08 pm

Define "performance" in an ETL context.

Unless you know what "performance" is, there's no point trying to improve it.

For example, in a mainframe environment where you're being billed for MIPS, the elapsed time metric may be secondary to using the smallest number of MIPS.