Controlling the memory buffers used by a messaging engine
Every messaging engine manages two memory buffers that contain
messages and message-related data. You can improve the interaction of a messaging
engine with its data store by tuning the properties that set the sizes of
the two buffers.
Increasing the number of data store tables to relieve concurrency bottleneck
Service integration technologies enables users to spread
the data store for a messaging engine across several tables. In typical
use this is unlikely to have a significant influence. However, if
statistics suggest a concurrency bottleneck on the SIBnnn tables
for a data store, you might try to solve the problem by increasing
the number of tables.
Tuning one-phase commit optimization
If you have configured your messaging engine to use a data store,
you can achieve better performance by configuring both the messaging engine
and container-managed persistent (CMP) beans to share the same data source.
Tuning the detection of database connection loss
If a messaging engine is configured to use a data store
and cannot connect to its data store, for example because the database
that contains the data store is not running, the messaging engine
does not start. You can tune your system to increase the chance of
a successful start of the messaging engine.