Latency between the time data is captured and the time it is published and applied to downstream targets can occur for a variety of reasons. Extreme Latency is however often caused by very large Transactions.
To prevent a huge transaction from affecting the replication latency of other transactions, consider excluding this transaction from replication by using one of the parameters provided for that purpose in the Capture configuration CAB file:
[--exclude-plan=<value>]
[--exclude-correlation-id=<value>]In order to maintain consistency between the source and Target it may then be necessary to execute the same transaction against the Target. One example might be a huge purge job that would be scheduled at both sites and excluded from replication.