You can configure an ObjectGrid server to start within an application server. WebSphere Application Server detects the ObjectGrid component and automatically starts the ObjectGrid server.
If you are using cluster configurations that enable replication, the high availability manager is required. ObjectGrid servers use the high availability manager differently than normal application servers. When the ObjectGrid server is in an application server, the ObjectGrid server does not configure, initialize or create the high availability manager service, but uses the existing high availability service in the application server. For replication between ObjectGrid servers, the ObjectGrid servers must be running in application servers that are members of the same core group.
All other functions of the ObjectGrid server are the same when the server runs in WebSphere Application Server. If your ObjectGrid cluster specification includes three servers, any three application servers in a single core group can host these ObjectGrid servers. The application servers can also span clusters, as long as the clusters belong to the same core group. The most important step is to correlate the server TCP/IP host name and port information in the cluster.xml file.