MQSeries.net :: View topic

hewitsd · Posted: Fri Nov 02, 2001 2:41 am Post subject:

I'm a bit confused about connection pooling, and how I should structure my application for scalability.

I'm working on a proof of concept that uses the P2P model using JMS.

I have coded a singleton that creates a single connection object, and all client requests create their own session and associated objects to send or receive a message. (Because the're not thread safe)

I've read all the stuff about JMS connection pooling, and it looks like it's automatically handled, but I would like to know how.

For now my application sends a message, gets the ID and does a receive for that specific message with a timeout - it's an interactive app, so I have to wait for a reply.

If say, 5 clients execute a send/receive together, 5 sets of objects get created in a thread safe way, ie the objects are member variables, but they all use the same connection.

1. Are the sessions being pooled?

2. Is there a limit to the number of concurrent users that can be doing this, ie do I need to design the application to share say 10 connections to spread the load?

3. As connections are thread safe, does this mean that each send/request will block before processing the next one.

The documentation on this lot is a bit sketchy for JMS, MQ Base classes are explained more fully.

Thanks in advance for any help

Dave Hewitson.