Naming is used by clients to reference application objects bound into the name space. To implement naming:
- Develop an application using either JNDI or CosNaming interfaces.
- Assemble the application using the Assembly Toolkit or Ant.
If the application is a client to an application running in another process, qualify the jndiName values in the deployment descriptors. If the objects have fixed qualified names configured for them, use them so that the jndiName values do not depend on the other application's location within the topology of the cell.
- Deploy your application
Put your assembled application onto the appserver. If the application you are assembling is a client to an application running in another server process, be sure to qualify the jndiName values for the other application's server objects if they are not already qualified.
- Configure name space bindings.
This step is necessary if:
- The deployed application is to be accessed by legacy client applications running on previous versions of WAS. In this case, configure additional name bindings for application objects relative to the default initial context for legacy clients.
- The application requires qualified name bindings for such reasons as:
- It will be accessed by J2EE client applications or server applications running in another server process.
- It will be accessed by thin client applications.
In this case, you can configure name bindings as additional bindings for application objects. The qualified names for the configured bindings are fixed, meaning they do not contain elements of the cell topology that can change if the application is moved to another server. Objects as bound into the name space by the system can always be qualified with a topology-based name. You must explicitly configure a name binding to use as a fixed qualified name.
- Troubleshoot any problems that develop.
If a Naming operation is failing and you need to verify whether certain name bindings exist, use the dumpNameSpace tool to generate a dump of the name space.