I have a java server (jdk 1.6) that pulls in data from a tibco ems topic through one connection. I am starting to see a back log on being able to pull message from tibco. How well does tibco scale if i break the data into multiple topics and each topic gets a separate connection (same JMS server) to my server.
Also I can't break up the server into multiple servers.
Tibco is by principle made to be used in an enterprise with MASSIVE amounts of data transiting the message server. Instead of pulling messages using the server, you can also try to use tibco the other way around: Publishing the messages and using your server to subscribe to the publisher. You can use the observer/observable design pattern as an analogy. I wouldn't worry about the scaling.
