Re: [wsbpel-implement] Face-to-Face scenarios

[rkhalaf <rkhalaf@watson.ibm.com>]

Hi Rajesh,
I've posted the bpel/wsdls, but have not made diagrams. There is a 
readme describing each scenario in detail (i hope there's enough detail 
in there ;) if not please feel free to ask).  Diagrams would be helpful 
but I haven't gotten around to that yet.
 
About splitting up loan approval, for now the services used could be on 
other machines. I was thinking those may be implemented as BPELs or as 
something else (Java, etc). I included sample (simple!) Java 
implementations from the BPWS4J samples; we could add BPELs at some 
point to replace those and get another example of BPEL-BPEL interaction. 
What do you think?

As for infrastructure, I believe we would bring our own. Diane may be 
able to clarify that further.

regards,
rania

Rajesh Pradhan wrote:

>Hi Rania,
>
>This sounds good to me. The Loan Approval example is our starting scenario
>too:). How do you envision the processes split over multiple engines ?
>
>Should we create some diagrams to make sure we are all on the same page ? Do
>you already have sample BPELs, WSDLs etc ?
>
>One of the questions I forgot to ask during the call was about the
>infrastructure. Do we get our own or will we have a setup at the F2F ?
>
>Thanks,
>
>Rajesh.
>
>-----Original Message-----
>From: rkhalaf [mailto:rkhalaf@watson.ibm.com]
>Sent: Tuesday, October 28, 2003 7:47 AM
>To:
>Subject: [wsbpel-implement] Face-to-Face scenarios
>
>
>Hi all,
>
>I'd like to start up a discussion of what people would like to try out
>at the face-to-face. The idea there is to meet and try out a few BPEL
>files on different engines to exercise the spec.
>
>I would suggest starting with a few BPELs that show different behavior:
>correlation with a multi-start example, calling other services and the
>use of links such as in  the loan approval example, and  more than one
>BPEL interacting, perhaps we can run one on each engine. For this we
>could start with a chaining echo: one client starts it off, then a chain
>of processes do rcv/inv/reply where the invocations kick of an instance
>of the same process running on a different engine until finally one just
>invokes a process that is the plain echo (rcv/reply).
>
>I also propose having an additive scheme to a simple rcv/inv/reply
>process: creating different versions of it to concentrate on exercising
>different areas without getting overloaded with complexity.
>
>That's the approach I would take in getting this off the ground.  I'm
>looking forward to everyone's suggestions/input/discussion.
>
>thanks,
>Rania
>
>  
>