| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Aggregate Control |
« View previous topic :: View next topic » |
| Author |
Message
|
| tmmgsf |
 Posted: Wed Mar 26, 2003 11:29 pm Post subject: Aggregate Control |
 |
|
Newbie
Joined: 13 Nov 2002
Posts: 6
|
Hi all
We looking at developing a generic solution that will fan-out messages to different destinations and fan them in and send them back to the requesting application.
In the examples given in the business scenarios, they use compute nodes to extract parts of the message and then send them to different destinations i.e. Hotel, Car Rental etc.
<Sorry I could not insert the image, its on page 76>
Figure from IBM Service Pac document (Business Scenarios) IC03
The question I have is, do I have to use more than 1 compute node to extract the information from the message and then send it to different destinations? Is there generic solution in doing the fan-out (e.g. without using 1..n compute nodes depending on the number of destinations I have)? If so, what about the Aggregate Request Nodes, don't I need to send each message send to destination to different Aggregate Request Nodes (Folders)?
Any suggestion or comments will be greatly appreciated.
Thank you in advance.[/img]  |
|
| Back to top |
|
 |
| lung |
| Posted: Thu Mar 27, 2003 7:28 pm Post subject: |
|
|
Master
Joined: 27 Aug 2002
Posts: 291
Location: Malaysia
|
Nope, you can use one single compute node IF your message is similar for each requests. In this case, you can use the 'propagate' esql command.
And nope, you can use one single aggregateRequest node If your message is similar for each requests. In this case, when the replies return, they will be grouped in folders of the same name but different cardinality. Eg. hotelReq[1], hotelReq[2], hotelReq[x], ... where x is the number of cardinality.
If each request is different, however, then you will need to use multiple compute nodes and multiple aggregateRequest nodes. Correct me if I'm wrong, please. 
Hope this helps.
_________________
lung |
|
| Back to top |
|
|
| tmmgsf |
| Posted: Fri Mar 28, 2003 7:31 am Post subject: |
|
|
Newbie
Joined: 13 Nov 2002
Posts: 6
|
Thanks Lung
We currently trying out the propagate 'esql' function. I am not too sure if the Propagate function creates different number of messages in a compute node and then sends each out one by one. But this is currently what we are trying to achive.
We also looking at using the propagate function with destinationList. The idea is to send each message created in the compute node with the Propagate function to different destination depending on the infomation of the incoming message.
OK, First we have a message coming into a compute node, then we want to split (in a sense) it up into 1..n message going to different destinations.
Coming to your point of using multiple Aggregate Request nodes. The different message that will come out of the compute node must be transformed before send to different destination, which means they will be different messages and therefore "night need" different AggRequest Nodes. I stand corrected though.
???
Thanks again. |
|
| Back to top |
|
|
| lung |
| Posted: Fri Mar 28, 2003 7:28 pm Post subject: |
|
|
Master
Joined: 27 Aug 2002
Posts: 291
Location: Malaysia
|
The 'propagate' function will create a new message to the out terminal of the compute node each time it is invoked. Read up on the redbooks for more details. Bear in mind the 'propagate' function is only available to wmqi 2.1, if I'm not mistaken.
What you mentioned in your previous post is correct. If you need to use multiple aggregateRequest nodes, then you will need to have multiple compute nodes as well cause unlike DestinationList, the aggregateRequest nodes have no way to differentiate between incoming messages. As such, you may need to use multiple output nodes as well...
_________________
lung |
|
| Back to top |
|
|
| tmmgsf |
| Posted: Wed Apr 09, 2003 3:58 am Post subject: AggregateControl/DestinationList |
|
|
Newbie
Joined: 13 Nov 2002
Posts: 6
|
Hi all
>
> Is it possible to pass the DestinationList from the AgggregateControl Node
>
> DESTINATIONLIST
> - ComIbmAggregateControlNode
> - AggregationName
> - GroupID
> - TimeOut
>
> With a DestinationList that I have created in a Compute node immediately
> after the AggControl Node
>
> DESTINATIONLIST
> - Destination
> - MQ
> - DestinationData
> - queueName
> - queuemManagerName
>
>
> It seems as if I can only pass one of them not both in the Compute Node.
>
> Here is me flow
>
> InputNode -----> AggregateControl Node --------> Compute Node --------->
> OutputNode -------------> AggregateRequest Node
>
> The Compute node is where all the action is taking place.
> Note: I have set the properties in the advance of the Compute Node tab to
> ALL. |
|
| Back to top |
|
|
| tmmgsf |
| Posted: Wed Apr 09, 2003 10:54 pm Post subject: Aggregation |
|
|
Newbie
Joined: 13 Nov 2002
Posts: 6
|
Thanks for all your replies.
I found out that both DestinationList infomation is passed in the compute node but only one of them is displayed during the debug session.
Tip: I always use trace/usertrace and those times I live them out, I always bump into problems. Point being, ALWAYS USE YOUR USERTRACE AND TRACE NODES.
Thanks again |
|
| 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
|
|
|
|
