Oracle® Business Intelligence Enterprise Edition Deployment Guide > Clustering, Load Balancing, and Failover in Oracle Business Intelligence >
Deploying Oracle Business Intelligence Publisher for High Availability
Figure 3 depicts the deployment of BI Publisher for high availability.
Figure 3. Deployment of BI Publisher for High Availability
BI Publisher J2EE Engine
The BI Publisher Engine is deployed in a web container in a J2EE server. Any number of BI Publisher Engines can be deployed simultaneously, each serving requests. For session state to be preserved in case of failure of any one particular engine, session state replication should be enabled.
Data sources are the raw sources used to compile the report data. This can be any JDBC data source. When integrated with Oracle BI, a JDBC data source named OracleBI EE is configured to point to the Oracle BI Server as a data source. When the multiple BI Servers participate in a BI clustered environment, the Oracle BI EE data source must point to the Primary and Secondary Cluster Controllers.
Repository Services and Scheduler Data
BI Publisher can store its Report metadata in either an XML DB or in a file system. All BI Publisher processes must have access to this Repository. In the case of a file system repository, a shared file system solution such as NAS must be used.