Page 1 of 1
server vs parallel
Posted: Thu Jan 15, 2004 8:03 am
by nag0143
When can we decide whether to use a server job or a parallel job ,
does parallel job does everything as server job can do ...any exceptions...
can anyone throw some light on this.....
Thanks a lot guys for your time
Re: server vs parallel
Posted: Thu Jan 15, 2004 8:10 am
by raju_chvr
Well Nag,
Server jobs does everything that Parallel jobs can. But the main difference is Parallel jobs are used for vast amounts of data. They are highly powerful. The way they have to be designed is differed from the Server jobs design.
For true Parallelism you need to have Parallel Extender which is also a product of Ascential. I assume you already know about this. For Server jobs you just need regular DataStage server.
You can search on this Forum under 'Parallel Extender' for parallel jobs help.
Posted: Thu Jan 15, 2004 8:41 am
by kcbland
Your post shows Windows as your DS OS type. So Server is your only choice currently.
Posted: Thu Jan 15, 2004 3:03 pm
by ray.wurlod
Just to make it clear, Parallel Extender is an additional cost, and not (yet) available for the Windows environment.
Parallel Extender does more than allow DataStage to run parallel jobs. It becomes part of the underlying architecture so, for example, INTEGRITY (now Quality Stage) processes can also take advantage of parallel processing where partitioning of data is handled automatically.
Posted: Thu Jan 15, 2004 4:50 pm
by nag0143
yeah... i am not not trying to either buy or use it now. :D .. i am just curious... wht makes it different from server jobs and what will be the criteria to choose between server and parallel jobs.. i know everything (tool supported processses)can be done with server jobs... it this vice versa
Posted: Thu Jan 15, 2004 8:32 pm
by ray.wurlod
nag0143 wrote:yeah... i am not not trying to either buy or use it now. :D .. i am just curious... wht makes it different from server jobs and what will be the criteria to choose between server and parallel jobs.. i know everything (tool supported processses)can be done with server jobs... it this vice versa
Everything can also be done in parallel jobs. But it's a different set of stage types, a different programming language, and a different way to use "routines" (buildops, etc).
The
Parallel Job Developer's Guide is in the DataStage documentation set (installed with your client software). Why not read through it?