in server jobs we have few stages and its mainly logical intensive and we r using transformer for most of the things and it does not uses MPP systems

in paralell jobs we have lots of stages and its stage intensive and for particular thing we have in built stages in parallel jobs and it uses MPP systems

In server we dont have an option to process the data in multiple nodes as in parallel. In parallel we have an advatage to process the data in pipelines and by partitioning, whereas we dont have any such concept in server jobs.

There are lot of differences in using same stages in server and parallel. For example, in parallel, a sequencial file or any other file can have either an input link or an output ink, but in server it can have both(that too more than 1).

server jobs can compile and run with in datastage server but parallel jobs can compile and run with in datastage unix server.

server jobs can extact total rows from source to anthor stage then only that stage will be activate and passing the rows into target level or dwh.it is time taking.

but in parallel jobs it is two types

1.pipe line parallelisam

2.partion parallelisam

1.based on statistical performence we can extract some rows from source to anthor stage at the same time the stage will be activate and passing the rows into target level or dwh.it will maintain only one node with in source and target.

2.partion parallelisam will maintain more than one node with in source and target.