Apparatus for routing database requests from application middleware.
Original Publication Date: 2005-Oct-04
Included in the Prior Art Database: 2005-Oct-04
Discloses a system for routing database requests from application servers
Apparatus for routing database requests from application middleware .
Discloses a system for routing database requests from application servers. In the proposed system, the Application Server of a J2EE environment will spray requests of to a single datasource to multiple "mirrored" databases. Since the databases contain the same info, any of the mirrored databases can service the database request.
The administrator creates a cluster of mirrored databases. These databases are kept in synchronization somehow (Database synchronization is not part of this disclosure, but we assume that there is some underlying database replication system in place). The application server administrator will then be able to associate a datasource to the cluster of database servers. This is similar to the way a J2EE application server can associate a weight (a term meant to decide how much load should a particular cluster member should handle) to each application server member (i.e., clone) of a cluster, we will be able to associate a weight to each database clone. When a datasource request is made, this weight can be taken into consideration as a routing mechanism. As one clone of the database cluster might be busy servicing other requests (or for some other reason might have its CPU/memory utilization up), there can be an ongoing communication back to the app server via a call back adapter.
The Sequence Diagram below illustrates the runtime flow.
J2EE Application Server...