Please understand the use of DAC.
1)DAC order the many of the informatica taks and issues command to run one by one.
2)When there are more then 300+ tasks (typically for any one module of BI APPS) it is impossible to control the tasks flow and order of execution by informatica scheduler.
3)Also when there are many conforned dimension like one exmaple (Organizations table is getting loaded in HR,SCM,FIN etc) and controlling incremental and full load for conforned dimesion done by DAC.
If you are using Informatica scheduler for your own custom grown applications you need not to use DAC for this.If you are using Custom ETL tasks along with BI APPS then I would say DAC comes handy.
Mark helps if it answers.
Here are a few I could think of:
1. DAC allows you to execute based on a "functional" subject area/module. This allows you to run a load for a functional module. INFA sched does not have this.
2. There is parallellism logic pre built based on the ETL stages, tasks, etc. You will have to manually set this up in INFA scheduler which is a pain.
3. DAC has pre built functionality to add/drop indexes/run stats, etc. INFA sched you will have to manually set this up. For example, suppose multiple tasks load the same table, the DAC will automatically add indexes after the last load. Dac can also differentiate ETL versus Query indexes
4. DAC has pre built logic for FULL/INCREMENTAL loads vua refresh dates. You can do this in INFA but again..its more work than is needed.
5. DAC allows you to configure key parameters for BI Apps loads. It dynamically uses a set of parameter files during run time to pass key values based on OBIA configuration.
6. In general, if you plan to use OBIA, using DAC will save you a lot of time and effort. Perhaps for custom INFA ETls, you can use INFA scheduler. For complex pre built loads for OBIA, it makes more sense to use DAC.
7. Suppose OBIA has a patch or new set of mappings or if you upgrade, Oracle will provide the DAC metadata whereas using INFA Scheduler you are on your own for any updates or variations in ETL load processes. Hope this helps. Pls mark if correct/helpful