Our IBM MQ Series queue Q1 will be receiving XML messages that we need to consume, invoke XSLT and place the resulting XML on another queue Q2. The size of the messages is small (up to 5KB). We can have 50 different types of XML messages. They are independent of one another
The system is close to real-time, so we can have only small latency up to 1-2 mins. The worst case scenario is that all 50 files can come to the queue at the same time so the program needs to be relatively robust.
We use IBM Websphere App Server. Eclipse is used for development.
Question: Should we write a JMS program that implements MessageListener interface or a Message Driven Bean that would consume the message from queue Q1, call XSLT and place the resulting XML on Q2? What are the benefits, pros/cons of JMS vs MDB? I have not used JMS in years so would appreciate any thoughts, tips, and ideas on configuration, development and testing.