On 8/30/06, Lennon Day-Reynolds <[EMAIL PROTECTED]> wrote:
All in all, it sounds like the underlying protocol really isn't going to be as important as the network connection polling and deferrred delivery, which basically requires either running a stripped-down ActiveMQ instance on every client system, or rolling your own client/server protocol.
Actually, you don't need individual servers to do the queueing. That's part of what JMS is deisgned for. The server will queue the messages until the client is ready to receive them, that is unless you tell the server otherwise.
JMS, and other enterprise messaging systems, are really designed for high-volume server-to-server communication, not intemittent client/server traffic.
Actually, that's not entirely true. JMS was designed for distributed, asynchronus messaging. It's actually a relatively lightweight messaging layer and the servers can be as light or as heavy as the designers intend them to be. ActiveMQ, in particular, was designed to be flexible enough to run in a heavyweight configuration or in very lightweight instances, depending on your needs. Of course, there are tradeoffs with each approach, but as long as you don't need things like failover, high throughput and database persistence (actually, you can use Derby, which is pretty light) you can run it with little overhead. _______________________________________________ PDXRuby mailing list [email protected] IRC: #pdx.rb on irc.freenode.net http://lists.pdxruby.org/mailman/listinfo/pdxruby
