4.1.1 Failover and caching
The most well known use of DRS is to replicate persistent HTTP Session data. If an application server fails, the request is routed to another application server, and the session data will be available there.
In order to minimize the impact of a failure, DRS coordinates with the WebSphere Cluster Workload Management routing algorithm to ensure requests and data end up in the same place. For example, a new capability of WebSphere Application Server v6 is to capture a stateful session bean and enable failover to another instance of that bean in a different WebSphere Application Server.
Figure 4-3 WebSphere supports failover in a clustered environment
The Data Replication Service caters to two scenarios: failover and caching.