This week's giveaway is in the Android forum.
We're giving away four copies of Android Security Essentials Live Lessons and have Godfrey Nolan on-line!
See this thread for details.
The moose likes Websphere and the fly likes A doubt on WebSphere design Big Moose Saloon
  Search | Java FAQ | Recent Topics | Flagged Topics | Hot Topics | Zero Replies
Register / Login


Win a copy of Android Security Essentials Live Lessons this week in the Android forum!
JavaRanch » Java Forums » Products » Websphere
Bookmark "A doubt on WebSphere design" Watch "A doubt on WebSphere design" New topic
Author

A doubt on WebSphere design

Suttan Pym
Ranch Hand

Joined: Jan 30, 2003
Posts: 46
Hi,
In the case of WAS5.1, when I create Queue, there is a requirement that I have to add the queue name in the jmsserver also. This is when I am using Embedded JMS Server.
Now my doubt is that, as this is required, is there any particular reason why it is implemented so, i.e. why could not WebSphere transparently add the the Queue name to the list in the jmsserver also?
Thanks in advance,
Suttan
Roland Barcia
author
Ranch Hand

Joined: Apr 15, 2004
Posts: 181
Because the notion of the JNDI name and the physical queue that implements it are different. I could define the JNDI name to a Queue on WebSphere MQ instead of embedded messaging, then I would need to create the Queue using WebSphere MQ admin. They chose the route of consistency between JMS provider implementations.


Roland Barcia: IBM Distinguished Engineer, CTO Mobile for Lab Services
Kyle Brown
author
Ranch Hand

Joined: Aug 10, 2001
Posts: 3892
    
    5
Welcome to the JavaRanch, Roland! Pretty soon I hope you'll be doing a book promotion for your new book here...

Kyle


Kyle Brown, Author of Persistence in the Enterprise and Enterprise Java Programming with IBM Websphere, 2nd Edition
See my homepage at http://www.kyle-brown.com/ for other WebSphere information.
Roland Barcia
author
Ranch Hand

Joined: Apr 15, 2004
Posts: 181
Thanks Kyle, I will be. I did not even realize this forum existed till this week. Seems like a cool place to hang out.
I will be promoting my book here when officially released
http://www.pearsonptg.com/book_detail/0,3771,0131468626,00.html
Especially since you created the book opportunity for us.
Suttan Pym
Ranch Hand

Joined: Jan 30, 2003
Posts: 46
Hi Ronald,
The explanation binds good in the case of usage of WebSphere MQ. But when I am using an Embedded JMS Server would it not be easy in the point of view of usability? It is not only us here, I found that this is the common error done by many programmers and you can find it any user groups.
Yes, one can say go through the document or InfoCentre or get yourself trained. But yes, this is my point of view usability wise.
Thanks in advance,
Suttan
Roland Barcia
author
Ranch Hand

Joined: Apr 15, 2004
Posts: 181
Suttan,
I understand what you are saying. I could also understand the frustration because it is not consistent with Topics where you do not need the extra step. The root of it is Embedded Messaging is a light weight MQ and so shares the same client code base. But the goal of JNDI is to seperate the naming interface and the real objects. For example, would you consider automatic Table creation for defining a Data Source? Most App Servers follow this model. I know on WebLogic, you need to add the Queue to the JMS server as well. I would say the Application Server is well designed, but perhaps there can be a couple of utilities that ease the administration. If they were tightly coupled in the App Server (Name space and object), I would actually call it a bad design because it is not pluggable.
All I can say is there are things coming out in the future that should simplify the experience.
Suttan Pym
Ranch Hand

Joined: Jan 30, 2003
Posts: 46
Hi Roland,
OK. As a design consideration from the IBM point of view having a trimmed down version of MQSeries to be used for Embedded JMS Server, this seems fine.
FYI:
In the case of WebLogic there is one point creation of Queue and Topic i.e. within the JMSServer. And any configuration in any other locations is not required. Similar approach to that of IBM is seen only in the case of Borland where it is using SoniqueMQ as the JMSServer.
The only point I am tring to make is as IBM knowns at runtime that during Queue creation it is creating a resource on an Embedded Queue or WebSphereMQ Queue. As the start point in both the cases are different (one is embedded and other is MQ), so based on that it could create an extra entry also.
Your statement related to DataSource holds valid in the case of creating resource using an external JMS provider, where two point configuration is required like in the case of webSphereMQ.
And as we do not have an option of configuring an Embedded JMS Server to be configured to an exteranl server so being tied up with it would not be a bad consideration considering the usability.
Thanks in advance,
Suttan
[ April 17, 2004: Message edited by: Suttan Pym ]
Ibnul Haque
Greenhorn

Joined: Apr 22, 2004
Posts: 10
I think the idea of separate jndi name and jmsqueue is a good one.
As for our case, we use a cluster MQ, when we create the queues and queue connection factory we need to do it in both places. This separation is needed for distributed J2EE paradigm.
Thanks
Suttan Pym
Ranch Hand

Joined: Jan 30, 2003
Posts: 46
Hi Ibnul,
Yes, I fully agree with your point and it is valid in the case of usage of WebSphereMQ. The demarcation also makes sense in this case. My suggestion was only for the case of Embedded JMS server. In the case of WebSphereMQ it is an external JMS provider and it makes sense to create the resource there also.
But in the case of Embedded JMS server where evrything is internal my suggestion is there should not be requirement that the user should make two entries.
Thanks in advance,
Suttan
 
 
subject: A doubt on WebSphere design
 
Similar Threads
Doubt in this()
Doubt regarding Listeners
doubt
Generics doubt
Doubt in Using the PriorityQueue Class