|   | 
	 
  
    | 
RSS Feed - WebSphere MQ Support
 | 
RSS Feed - Message Broker Support
 |   
 
  
	     | 
	 | 
   
 
  
	|  Working with several message flows | 
	« View previous topic :: View next topic »  | 
   
  
  	
	  
		
		
		  | Author | 
		  Message
		 |  
		
		  | MQDummy | 
		  
		    
			  
				 Posted: Wed Jul 03, 2002 12:17 am    Post subject: Working with several message flows | 
				     | 
			   
			 
		   | 
		 
		
		   Novice
 
 Joined: 11 Jun 2002 Posts: 23
  
  | 
		  
		    
			  
				Hi,
 
 
Assuming if i have several message flows, is it possible to make the data go from one message flow to another if the message flows are not on the same 'screen' (How do i put it    )
 
 
What i meant was if i have :
 
 
3 message flows
 
 
flow_1, flow_2, flow_3 under the Message Flows tree.
 
Is it possible to route messages in flow_1 to flow_2?
 
I know it can be done if 3 of them are in one message flow definition, but was wondering it can be done for diff definitions
 
 
Best regards! | 
			   
			 
		   | 
		 
		
		  | Back to top | 
		  
		  	
		   | 
		 
		
		    | 
		 
		
		  | cronydude | 
		  
		    
			  
				 Posted: Wed Jul 03, 2002 12:53 am    Post subject:  | 
				     | 
			   
			 
		   | 
		 
		
		    Voyager
 
 Joined: 11 Nov 2001 Posts: 85 Location: US 
  | 
		  
		    
			  
				Hi,
 
   if u mean passing info from one flow to another...its simple.....let the outputQ of sending flow be same as the inputQ of the recieving flow.the prerequisite is that both the flows are deployed to the same broker.
 
 
regs,
 
crony | 
			   
			 
		   | 
		 
		
		  | Back to top | 
		  
		  	
		   | 
		 
		
		    | 
		 
		
		  | MQDummy | 
		  
		    
			  
				 Posted: Wed Jul 03, 2002 1:05 am    Post subject:  | 
				     | 
			   
			 
		   | 
		 
		
		   Novice
 
 Joined: 11 Jun 2002 Posts: 23
  
  | 
		  
		    
			  
				Thanks a bunch!
 
 
Its a pity that its hard to extract such information from the documentation that comes along with MQSI... | 
			   
			 
		   | 
		 
		
		  | Back to top | 
		  
		  	
		   | 
		 
		
		    | 
		 
		
		  | MQDummy | 
		  
		    
			  
				 Posted: Wed Jul 03, 2002 1:14 am    Post subject:  | 
				     | 
			   
			 
		   | 
		 
		
		   Novice
 
 Joined: 11 Jun 2002 Posts: 23
  
  | 
		  
		    
			  
				Another question
 
 
What if there are :
 
 
Flow_Decide,
 
Flow_A,
 
Flow_B
 
 
Flow_Decide will pass the message to either Flow_A or Flow_B, however can Flow_Decide accept messages of different MQMD? Because Flow_A and Flow_B have different message format | 
			   
			 
		   | 
		 
		
		  | Back to top | 
		  
		  	
		   | 
		 
		
		    | 
		 
		
		  | dhaksr | 
		  
		    
			  
				 Posted: Wed Jul 03, 2002 2:08 pm    Post subject:  | 
				     | 
			   
			 
		   | 
		 
		
		   Apprentice
 
 Joined: 20 Feb 2002 Posts: 37
  
  | 
		  
		    
			  
				1. Flow_decide show take the message as Blob.
 
(Actually driven by the "Decide" criteria)
 
Flow_A and Flow_B should use ResetContentDescriptor to set the
 
appropriate message format. 
 
 
2. To answer your first question, you may also use input terminal and
 
output terminal nodes. for eg. Flow_decide can have two output 
 
terminals and flow_a and flow_b each have one input terminal. | 
			   
			 
		   | 
		 
		
		  | Back to top | 
		  
		  	
		   | 
		 
		
		    | 
		 
		
		  | mpuetz | 
		  
		    
			  
				 Posted: Fri Jul 05, 2002 2:16 pm    Post subject:  | 
				     | 
			   
			 
		   | 
		 
		
		   Centurion
 
 Joined: 05 Jul 2001 Posts: 149 Location: IBM/Central WebSphere Services 
  | 
		  
		    
			  
				Hi,
 
 
from a design point of view, chaining of message flows
 
using external queues as 'gateways' between flows is a
 
two-edged sword:
 
On the plus side, it's a very modular and flexible design,
 
on the other side it costs quite a bit of performance, because
 
you have to parse a message serialize it into a message again
 
with each flow. All the worse if your message is persistent, then
 
the additional IO may really kill you.
 
 
If you are performance critical, you should avoid such a strategy
 
and use the master(routing)- and sub-flow design principle.
 
I.e. in you master flow you detect the message type and then
 
simply branch into a sub-flow without going through a queue.
 
Ideally in your overall architecture you should include a tiny
 
header in every message (most suitably the MQRFH2 header
 
which is directly supported by MQSI) which describes the message
 
format of the message body. This way you can save yourself
 
the costly ResetContentDescriptor, which forces the parser to
 
serialize the message tree and reparse it.
 
If you can't use an MQRFH2 header consistently the best way
 
is to read the message as BLOB, since the BLOB parser doesn't
 
parse at all. Then check the initial bytes of the message to detect
 
the message type and do a reset content descriptor in your sub-flow
 
to set the correct message type. Used in this way the overhead of
 
using a ResetContentDescriptor is usually acceptable. _________________ Mathias Puetz
 
 
IBM/Central WebSphere Services
 
WebSphere Business Integration Specialist | 
			   
			 
		   | 
		 
		
		  | 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
  | 
  		 
	   
	 | 
   
 
  	 | 
	  |