NCL Concepts › Communication Between Processes › Dependent Request Queue

Dependent Request Queue

While the INTQUE command can be used to place messages on the dependent response queue of a target process, it is sometimes confusing if messages arrive from an INTQUE command when the process is also expecting the results of a command it has issued. The INTQUE messages might arrive in the middle of the messages generated by the command.

To avoid this and to provide a method of isolating messages generated by commands issued by a process (which are predictable) from messages arriving from INTQUE commands issued externally (which are unpredictable), a second message stream is available. This is the dependent request queue.

Example: Dependent Request Queue

If the example in the previous section is changed so that PROC1 executes:

&WRITE DATA=&ZNCLID READY FOR WORK.
&INTREAD ARGS TYPE=REQ

and the operator enters:

INTQUE ID=357 TYPE=REQ DATA=BEGIN

then the &INTREAD issued by PROC1 completes as before with &1 = BEGIN, but the communication flow takes place on the request flow rather than the response flow.

In a more complex scenario, PROC1 can use this differentiation in message flows to keep separate the results of its own commands (the response flow) and the arrival of messages from operators or other processes that arrive on the request flow.