Performance best practices with IBM InfoSphere Change Data Capture

For my experience with this data migration/synchronization tool there are a couple of best practices to speed up the transfer of data from a source database into a target database that I've used with a DB2 database. Before to start the refresh or mirroring operations of the data evaluate to:

• remove all the triggers present on the target database
• remove all foreign keys on the target database 

Removing the triggers boosts a lot the speed because if you have a huge transfer of rows from a source table into a target table that has a trigger defined this can cause many consequentials insert and update to be executed.

Removing the foreign keys you can transfer without taking care of the orderm because if you are transfering all tables content involved with the foreign keys, the referencial integrity is respected anyway.

If you are transferring the database from an environment A into an environment B in general after moved the data you have probably to recreate all triggers and foreign keys after synchronized the source and target database.

Instead, for example, if you are using this tool for mirroring (that is to copy changes continuously from source to target) a primary database into a secondary database for data warehouse purposes, like for reporting, then you do not need the triggers neither the foreign keys at runtime usage of the database, then removing them it is not a problem.