Encyclopedia functions that can run together within the same model are easily run during peak processing hours. A few examples of these types of processes are code generation, testing, reporting, and defining subsets.
Processes that cannot run concurrently, but that are needed during peak hours, can be single streamed by running them in batch using the same job names. The best example of this is upload and download.
Other encyclopedia functions that have a tendency to cause contention should be broken into small jobs (where applicable). For Example, migration and adoption. Both can cause contention when they do not complete in a short time and both can be broken into small pieces. Running 10 migrates that each complete in 10 minutes causes much less contention than one migrate that runs for 100 minutes.
|
Copyright © 2013 CA.
All rights reserved.
|
|