In this chapter, we discussed various performance tuning features related to JMS and HornetQ—the JMS provider in WildFly. HornetQ is a very competent MOM that has high performance as one of its major design goals.
The following are some of the general rules for performance-tuning JMS:
Keep messages as small as possible
Avoid durable messages (if possible)
Use JTA to be able to batch handling messages
If losing messages due to failures is okay, then there are some optimizations available
Tuning HornetQ includes looking at the storage of persistent messages, where dedicated and separate disks should be used for journals and large messages. Paging sizes, flow control and message delivery options are also major factors that affect performance. As HornetQ is an advanced and large project, we urge you to read its documentation in detail to make the most of all its bells and whistles.
We will shortly look more into HornetQ as we now will engage a cluster of WildFly instances.