On 8/23/06, Markus Joschko <[EMAIL PROTECTED]> wrote:

That means that the creation of a connection for the vm transport is so
efficient that pooling is not required?

Yes - plus you can't pool consumers without using them. Consumers are
not like JDBC connections which if you don't use them they have no
effect on the system, since consumers start to consume messages.


Why is it necessary to create a new connection for each new session?

Its not - only a JMS session/consumer is required. One connection for
the entire JVM is fine.

So to summarise; 1 JMS connection for the JVM, each HTTP session has
its own JMS session and producer/consumer.
--

James
-------
http://radio.weblogs.com/0112098/

Reply via email to