This section describes scenarios that IBM Replication was not designed to support. In these cases, you must look for other solutions. Hints are given as to which products or techniques may be helpful.
IBM Replication does not support synchronous replication. However, lack of support is often the benefit of using IBM Replication as the base application is not affected. If the network connection from the base application to the target is unavailable, the base system is still be available if you use IBM Replication, and data changes are not lost.
In many cases, organizations that start out with an initial requirement for immediate availability of copy data find that the benefits of staged delivery (such as, better network use, less database contention, and the opportunity to enhance the data) are more attractive, so they reexamine the original requirements. IBM Replication supports continuous timing of subscriptions.
If synchronous data delivery is essential to the application, the recommended approach is to use DRDA two-phase commit within the application.
Many planners are tempted to consider IBM Replication for hot site recovery because of its ability to replicate tables to other sites. However, IBM Replication was not designed for such recovery because the underlying architecture is asynchronous. There is no guarantee of how much data has been captured or applied at any one instant.
Hot site recovery is a complex and specialized task and is supported by several IBM and vendor features and tools. For example, in the System/390 area, the peer-to-peer remote copy (PPRC) hardware feature and the extended recovery component (XRC) are available.