Differences between Server and Parallel Extender

dsx · Post by **dsx** » Tue Aug 30, 2005 3:50 pm

Just curious in terms of performance about some of the basic differences between the Server edition and the Parallel Extender version of DataStage ?

ray.wurlod · Post by **ray.wurlod** » Tue Aug 30, 2005 4:23 pm

Define "performance".

Read chapter 2 of the Parallel Job Developer's Guide

ray.wurlod · Post by **ray.wurlod** » Thu Sep 01, 2005 4:53 pm

It depends. Given that one of your criteria is to use all the partitioning and parallelism available in server jobs, I'd be happy to assert that a highly competent server job developer could create a set of jobs that would finish more quickly than a set of parallel jobs created by a less competent developer.

Complex lookups are a bummer in either edition; you always seek ways to pre-select a lookup set. It's like the difference in SQL between a subquery (done once) and a correlated subquery (done for every row) - you try to avoid the latter.

However, that said, the parallel job is automatically scalable. Right now you are only processing four million rows. What happens in two years time when you need to process ten million rows. Against this scenario parallel jobs seem to be the way to go.

Also, there's some degree of FUD (fear, uncertainty and doubt) being put about - unreasonably, in my opinion - about the future of server job technology.