tag:blogger.com,1999:blog-3237724005744642470.post1987071795050105400..comments2020-08-21T15:18:26.763+01:00Roger Hugheshttp://www.blogger.com/profile/07042290171112551665[email protected]Blogger3125tag:blogger.com,1999:blog-3237724005744642470.post-68126741690756736842014-01-03T14:56:10.115+00:002014-01-03T14:56:10.115+00:00This is a nice article on how to use pub/sub within Hazelcast. I'm curious though, does anyone have any experience on how it compares from a performance perspective to a JMS and or AMQP mechansism such as active mq or rabbit mq?Rob Schierholzhttps://www.blogger.com/profile/03155833544710121849[email protected]tag:blogger.com,1999:blog-3237724005744642470.post-86888308643680357642014-01-03T14:53:41.469+00:002014-01-03T14:53:41.469+00:00This is a nice article on how to use Hazelcast pub/sub. I did have one question, does anyone know of any bench marks and or how performance would compare to say a rabbit mq. We've used Hazelcast successfully as a distributed cache, however have stuck with the move traditional queuing mechanisms for this type of work. I'm just curious if anyone has pros and or cons.Rob Schierholzhttps://www.blogger.com/profile/03155833544710121849[email protected]tag:blogger.com,1999:blog-3237724005744642470.post-38301094725683669762014-01-02T15:10:28.939+00:002014-01-02T15:10:28.939+00:00If you want to make this really high performing just use DataSerializable (better IdentifiedDataSerializable) instead of Serializable and write the data yourself into the stream. That way you can get max performance (up to 2-5x higher than Serializable).Chris Engelbert (noctarius)https://www.blogger.com/profile/09320653447815178755[email protected]