Is pooling sessions / producers a good strategy to get multi-threaded throughput for producers?
I have a loop that produces message and needs to complete quickly. It’s parallel now but still not as fast as I want. I was thinking that perhaps using a pool of sessions/producers and then using one per thread would be faster: 1. The sessions/producers won’t have to block send()s from other threads, so I would get more parallelism. 2. I get more protocol level parallelism and use more cores within AMQ on the server side since I’m using multiple connections. Just worried this might be a JMS anti-patter and I also have to spend some time refactoring code (which is no fun). Additionally, I introduce a new component which could fail (my pooling code). Kevin -- Founder/CEO Spinn3r.com Location: *San Francisco, CA* blog: http://burtonator.wordpress.com … or check out my Google+ profile <https://plus.google.com/102718274791889610666/posts> <http://spinn3r.com>
