IBM WebSphere DataStage Implementation Considerations

The following considerations should be noted before you begin DataStage implementation:

1. Perform a detailed analysis of your project creation strategy.

   You should decide whether you would like a single project for the whole EPM application or have separate projects for each data mart.

2. Create separate DataStage projects for development, QA, and production.

   PeopleSoft recommends that the production project reside on a separate DataStage server.

3. Classify your jobs as high, medium, and low volume.

   Provide project defaults for array size, transaction size, IPC buffer and other performance parameters. Any exceptions and special cases must be handled by changing the value at the job level.

4. Open a sample job from each category and familiarize yourself with the filter conditions in the source, update strategy, job design, job parameters and other transformations.

5. Review the master run utility and create appropriate sequential file inputs.

   Analyze this feature and decide on the different categories that you want to run using this utility.

6. Review the master sequencers and familiarize yourself with them.

7. Open one of the business process and identify all the jobs that are required to run it.

   Run this as an example to learn how the jobs are ordered, the interdependencies, the hash file usage, and so forth.