3.7.3 Load balancing with the CORBA Naming Service (in the case of Session Beans and Entity Beans)
This subsection describes the configuration where load is balanced by using the round robin search functionality of the CORBA Naming Service present on the J2EE server (in-process), when the access point component is either a Session Bean or an Entity Bean.
- Organization of this subsection
(1) Features of the system configuration
You can use this configuration when the access point of the application running on the J2EE server is either a Session Bean or an Entity Bean. The EJB client application is the client. The EJB client application distributes the request distribution destinations by looking up the object reference in the round robin format from the logical Naming Service registered in the system properties. The Session Bean and Entity Bean, however, need to be started with the same name (same optional name) in the respective J2EE servers.
The EJB client application accesses the same J2EE server from the time the Naming Service of the J2EE server is looked up in the round robin format until the Naming Service is looked up again.
The following figure shows an example configuration of load balancing with a Session Bean and an Entity Bean as the access points:
|
Note: For other legend items, see 3.2 Description of the system configuration.
- Features
-
-
Scalability and availability of a Session Bean and an Entity Bean can be ensured.
-
You can prepare multiple Application Server instances and then select and access the CORBA Naming Service in the round robin format from the EJB client, and thus distribute the load.
-
When an error occurs in a specific Application Server or when maintenance is required, the corresponding Application Server is not accessed after the EJB client application detects the termination of the J2EE server. As a result, reduced system operation is possible.
-
- Flow of requests
-
Requests are sent from the EJB client application to a Session Bean and an Entity Bean forming the access points on the J2EE server. At this point, the EJB client application selects the Naming Service of the J2EE server in the round robin format and looks up the object reference.
(2) Required software and the processes to be invoked on the respective machines
The required software and the processes to be invoked on the respective machines during load balancing with the CORBA Naming Service are the same as in the configuration that uses Session Beans and Entity Beans as the access points. See 3.4.3 Configuration with Session Beans and Entity Beans as access points.