My experience covers RabbitMQ but not IBM MQ, so I am not able to make a direct recommendation I would, however, say that it would probably be impossible to recommend one over the other as a general rule; the right solution will almost certainly depend upon the precise…

I found your comment about clustering interesting, because it echoes my own experience of RabbitMQ clusters in production (with mirrored / HA queues). What problems did you have, specifically?