Hi, I'm currently trying to develop an application where a couple of agents want to query several hundreds of clients for information. This happens individually for every client and not that frequent - there's no need for 'all clients, please do whatever' functionality. But I'd like to assure that the query reaches the client - or get the information that no one was listening for my query. Basically, this is just request-response using a temporary reply queue while the clients react on what the agents want to query.
As far as I see, there are these main options: * Create a topic exchange and use subscribtions like "client.1", "client.2", etc. In this case messages get dropped if the client is down in the moment of delivery. Is it possible to find out whether someone got the message or if it just went to /dev/null? * Create a queue for each client. These are many queues, but the message will reach its destination. Is there a limitation on how many queues I can create? Are there other recommended ways to achieve the desired setup? Thanks! Cajus --------------------------------------------------------------------- Apache Qpid - AMQP Messaging Implementation Project: http://qpid.apache.org Use/Interact: mailto:[email protected]
