Oracle GoldenGate supports cascading synchronization, where Oracle GoldenGate propagates data changes from the source database to a second database, and then on to a third database. In this configuration:

• A primary Extract on the source writes captured data to a local trail, and a data pump sends the data to a remote trail on the second system in the cascade.

• On the second system, Replicat applies the data to the local database.

• Another primary Extract on that same system captures the data from the local database and writes it to a local trail.

• A data pump sends the data to a remote trail on the third system in the cascade, where it is applied to the local database by another Replicat.

  Note:

  See Creating a Reporting Configuration with a Data Pump on an Intermediary System if you do not need to apply the replicated changes to a database on the secondary system.

Use this configuration if:

• One or more of the target systems does not have a direct connection to the source, but the second system can connect in both directions.

• You want to limit network activity from the source system.

• You are sending data to two or more servers that are very far apart geographically, such as from Chicago to Los Angeles and then from Los Angeles to servers throughout China.

When considering this topology, take note of the following:

• This configuration can be used to perform data filtering and conversion if the character sets on all systems are identical. If character sets differ, a data pump cannot perform conversion between character sets, and you must configure Replicat to perform the conversion and transformation on the target.

• On the second system, you must configure the Extract group to capture Replicat activity and to ignore local business application activity. The Extract parameters that control this behavior are IGNOREAPPLOPS and GETREPLICATES.