Usually destinations are used for different 'services' or 'roles'; so one queue for requests to any server is a good idea - then either a topic for responses to all clients or a queue (or temporary queue) per client for responses.
This may help... http://incubator.apache.org/activemq/how-should-i-implement-request-response-with-jms.html On 1/10/07, rousseau <[EMAIL PROTECTED]> wrote:
I have a quick basic question about handling Client-server comms with queues and topics. If I want to pass data back and forth between processes is it a more generally accepted practise to: 1) have different queue/topic names for sending and subscribing to messages on both the client and server side? e.g. server starts subscribing to queue "client" and send messages on queue "server". Client starts subscribing to queue "server" and sending messages on queue "client". or 2) have a single queue/topic for all the messages? eg. both client and server send and subscribe to a single queue eg. "clientservercomms". Are there benefits/drawbacks to either approach? Also with approach #2, is there any built-in way to determine whether messages on the queue originated from the client or the server? thx /Steve -- View this message in context: http://www.nabble.com/Client-Server-comms-w.-MQ-tf2950766.html#a8252470 Sent from the ActiveMQ - User mailing list archive at Nabble.com.
-- James ------- http://radio.weblogs.com/0112098/
