[zeromq-dev] Point-to-Point Client/Server Request/Response

[Alex Silvera]

Studying the excellent design documentation. Wanted to get an understandng how a conventional point-to-point conversation would be mapped in 0MQ solution space with practicality left aside. Assuming a number n of private (firewalled) clients each sending 'requests' to a single public server and waiting for their appropriate 'responses', we would have:

1. Server with Global Queue Q-REQ
2. Each Client would create a Local Exchange E-REQ
3. Each Client would bind their E-REQ to Q-REQ to send their request.

It is the next steps on how the server will handle the respones to the each of the clients that is not so clear. Would we have:-

4. Server creates an instance of a Global Exchange E-RESi...n for each client
5. Each  Client would create a Local Queue Q-RES
6. Each Client would bind their Q-RES to their corresponding E-RESx to recieve their response.

Obviously, it feels like there should be a more elegant solution. Or is it a case of trying to fit a square peg in a round hole.

Alex


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20090426/2e03ba6c/attachment.htm>