The common precautions for setting a communication timeout are explained below:

• Precautions for when the definition of communication timeout overlaps
  If a communication timeout is set up for both A and B in Figure 2-14 and Figure 2-15, and a communication timeout is set up for A and C, you perform the operation as follows:
  
  • Operations when a communication timeout is specified for both A and B
    When a communication timeout is specified for both A and B, the operations will be as shown in the figure below:
    

    Figure 2-16 Operations when a communication timeout is specified for both A and B

    

  • Operations when a communication timeout is set for both A and C
    When a communication timeout is specified for both A and C, the communication timeout setting for C is enabled.
    
• Precautions during client implementation
  Along with the RMI-IIOP communication timeout and the naming service communication timeout, if a response is not returned within the specified time for a request sent from the client, the corresponding request will be cancelled as the timed out request. In such a case, the exceptions java.rmi.RemoteException (such as org.omg.CORBA.TIMEOUT) and javax.naming.NamingException will be thrown. For implementing a client that uses a communication timeout, give consideration to the fact that these exceptions might occur during invocation of the methods in Enterprise Beans or in JNDI APIs.
  
• Precautions concerning operation at the server side after the occurrence of a timeout
  When a request from the client reaches the server (naming service and Enterprise Bean) and a timeout occurs during the processing at the server side, an exception is returned to the client. However, the processing continues normally at the server side even after the occurrence of a timeout, and therefore, the instances of the Enterprise Bean are not destroyed and the resources such as the resource connection are not released.