The informational databases are populated with data that originates from the various productions databases. Typically, the data is copied or...
The informational databases are populated with data that originates from the various productions databases. Typically, the data is copied or extracted using one of two techniques:
- Refresh replaces the entire target with data from the source. This works well when you are moving small amounts small amounts of data, which have low requirements for frequency of update (that is, the data has low volatility). It is also used for doing initial bulk loads to the target database.
- Update only sends the changed data to the target. Updates can be either synchronous, which means that the target copy is updated in the same commit scope as the source table, or they can be asynchronous, which means that the target table is updated in a separate transaction that the one updating the source. Synchronous updates are useful in production environments for creating replicated databases that provide high availability. Asynchronous updates are useful in data warehousing situations. You get to specify the level of synchronization that you want to maintain between the source and the target and the interval of updates. This means that you get to control the level of data obsolescence you can tolerate.