DataStage server job performance on Unix Vs Windows

[dpsahoo]

I was just analysing on a generic basis what OS will be better for processing high volume jobs(> 1 mill rows). Unix or Windows?

I have never worked on Windows before, hence would greatly appreciate if anybody who has worked on Windows can point out the drawbacks on working on a Windows server while processing server jobs with huge vols of data.

Thanks.

[ray.wurlod]

You're not comparing apples with apples so most comparisons are probably unfair.

The thing I most like about UNIX over Windows is the infrequency with which the machine has to be re-booted - that is, its "up time".

On identical hardware there theoretically should be little difference between Windows and UNIX, since the underlying software is, to all intents and purposes, identical. Parallel jobs may be more costly on Windows because of the extra layer of software (MKS Toolkit) and because of the fact that Windows is a messaging operating system.

[Ultramundane]

The Windows version will give you a SQL Enterprise stage which can be used for doing bulk loads to SQL Server. That might be important for doing a lot of SQL Server work.

Maybe what OS the developers are most familiar with could be an issue as well?

Other than that, the Unix version will probably be more stable and faster.

[lstsaur]

Server jobs cannot use the SQL Enterprise stage. For server jobs to load huge volumes of data, take a look at Inter-Process stage, Link-Partioner stage, and Link-Collector stage.