Implementing Siebel Business Applications on DB2 UDB for z/OS > Migrating Data Using Siebel Enterprise Integration Manager > About Setting Up EIM for DB2 >

Optimal EIM Batch Size

You can regulate EIM commit frequency in several ways, including:

• Setting COMMIT EACH TABLE = TRUE.
• Setting COMMIT EACH PASS = TRUE.
• Adjusting batch size.

Of these, the batch size most directly influences EIM performance.

The size of EIM batches can significantly affect throughput. This issue occurs as a result of the number of commits that EIM must execute. While each commit is CPU-intensive, commits release page locks, thereby releasing memory and avoiding lock escalation and timeouts.

While every installation is different, perform initial testing with large batch sizes (except DELETE EXACT, which normally runs better with smaller batch sizes). You might want to use this large-batch testing as a starting point for your own processes and modify batch size based on results.

NOTE:  Processes that populate child tables might violate the maximum number of locks allowed by the system or cause contention for resources in those child tables. To avoid this, reduce the batch size of these processes.