How is message exchange implemented in IBM MQ: push or pull?

Question

I'd like to understand what is the implementation of the JMS drivers for WebSphere MQ. I'm not interested in the JMS sync/async methods directly but rather in their implementation.

If an application is deployed on a J2EE server and performs asynchronous calls (by implementing onMessage()), does the J2EE server still pull to check for the availability of a message or does MQ actively call back the client (J2EE server) through the opened socket to notify the presence of the message?

Morag Hughson · Accepted Answer · 2018-12-19T03:43:50.813

3

It depends on what version of IBM MQ you are using.

Prior to MQ V7, the JMS layer would poll the queue manager for a message - essentially pull.

After MQ V7, the queue manager natively understands and implements asynchronous get where it will notify the getter of a message so it no longer has to poll.

Further Reading

What is new in WebSphere MQ classes for JMS - about half-way down the page.

edited Dec 19 '18 at 03:43

answered Mar 02 '18 at 10:31

Morag Hughson

7,255
15
44

Thanks for the insight! Are there any links to the corresponding sections in documentation? – Pavel Vasilev Dec 18 '18 at 13:27
Hi @PavelVasilev - I have updated my answer with a link which I hope helps. – Morag Hughson Dec 19 '18 at 03:44

How is message exchange implemented in IBM MQ: push or pull?

1 Answers1