| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| MQ binding vs. client connection |
« View previous topic :: View next topic » |
| Author |
Message
|
| czaszek |
 Posted: Thu Jan 18, 2007 4:35 am Post subject: MQ binding vs. client connection |
 |
|
Apprentice
Joined: 30 May 2006
Posts: 36
Location: Poland
|
Hi All
I am to prepare an architecture for a new system which uses MQ as a transport.
On the server site will be about 5 gateway servers with queue managers and two servers with business logic behind them.
The first version of the architecture (not mine) tells not to install queue managers on these two servers but instead to connect using JMS and MQ clients to the gateway servers.
My question is:
lets assume that the servers are located next to each other, use fibre channels and gigabit ethernet with cisco switches, what will be a peformance impact of using MQ client connection instead of using MQ binding (install queue managers in all servers)?
Does anyone has any performance tests, saw any documents describing similar problems?
Thanks in advance for any help.
Best Regards
Lukasz |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Thu Jan 18, 2007 5:02 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Start with business requirements before you start with performance requirements.
Does the business logic require two-phase commit between transactions with MQ and transactions with databases?
If so, you probably want to put queue managers on the business logic servers. You can also use the Extended Transactional Client - but you're going to pay the same fees and many people opt to put a real queue manager down instead.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| czaszek |
| Posted: Fri Jan 19, 2007 3:26 am Post subject: |
|
|
Apprentice
Joined: 30 May 2006
Posts: 36
Location: Poland
|
As far as I know two phase commit will be required.
Thank you for your info, it will be most importand during negotiations with a management. |
|
| Back to top |
|
|
| czaszek |
| Posted: Mon Jan 22, 2007 8:51 am Post subject: performance tests from IBM |
|
|
Apprentice
Joined: 30 May 2006
Posts: 36
Location: Poland
|
Hi All
In the IBM page with support packs
http://www-1.ibm.com/support/docview.wss?rs=977&uid=swg27007205
under a link Performance Reports you can find several analisys regarding performance in many platforms (MQ 5.3 and 6.0).
For my configuration Linux x86, the performance difference (taking into consideration MQI application) between installing local qmgr and use binding mode and using client connection is about 10%. When the application uses JMS it can be more, up to 30% (but the same configurations in JMS tests was not tested, so it is my estimation). |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
