OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

soa-rm-ra message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: [soa-rm-ra] [soa-ra] why or why not composition needs to addressed

Ken et al,

Sorry for the delayed response on this - I was out of town visting my
native NJ the second half of last week (catching up now). Some partial
answers below.

Kind Regards,
Joseph Chiusano
Booz Allen Hamilton
700 13th St. NW, Suite 1100
Washington, DC 20005
O: 202-508-6514  
C: 202-251-0731
Visit us online@ http://www.boozallen.com

-----Original Message-----
From: Ken Laskey [mailto:klaskey@mitre.org] 
Sent: Wednesday, March 15, 2006 3:16 PM
To: Chiusano Joseph
Cc: soa-rm-ra@lists.oasis-open.org
Subject: RE: [soa-rm-ra] [soa-ra] why or why not composition needs to

OK, Joe, questions just for you (unless anyone else wants to pipe in):

- What is the hang-up for getting BPEL out?  I remember one of the
rationales for submitting BPEL to OASIS instead of W3C was W3C takes too
long.  However, WS-CDL finishes Candidate Rec testing at the end of this
month (but I have no insight whether any problems were uncovered to slow
further advancement).

[JMC] Here are some articles on the BPEL delay from last November that
may help:

"BPEL 2.0 is delayed"
"BPEL gets bopped... again"
I noted that the link to Dave Chappell's (the Sonic Dave Chappell) blog
post is broken - not sure why.

- Part of the WS-CDL spec is a formal model based on pi calculus that is
the basis of the spec content.  I believe BPEL had a formal model in the
beginning.  Does it still?

[JMC] Not sure what we really mean by "formal model" (could have
different interpretations), but I believe the answer is yes, at least
according to my definition of "formal model".

- Your write-up links BPEL to WSDL 1.1 constructs.  Is there any thought
of mapping to WSDL 2.0 (also in CR)?

[JMC] Good question - I checked the latest draft
ation-draft.doc) from 3/16/06 and it appears that the answer is no at
this time (search on "WSDL 1.1" and "WSDL 2.0").

- Is there a more detailed write-up that goes into the type of process
control that BPEL is intended to represent and what is outside its
scope?  For example, must the decision making be done within the BPEL
engine or can it make external calls to more sophisticated decision
support capabilities?

[JMC] I would recomend searching on "An opaque expression is a
placeholder for a corresponding Executable WS-BPEL expression" in the
latest WS-BPEL draft, which discusses decision points. The answer
may/may not be clear there.

- Orchestration represents central control while choreography is more
peer-to-peer.  Do you know of anything that discusses when you would use
one rather than the other?  For example, I could imagine a choreography
where some peers were more equal than others and the limit would look
like orchestration.  OTOH, I could imagine an orchestration where the
main engine delegates control and, in the limit, you really lose site of
a central control.  Are these naive suppositions that may be doable in
theory but prohibitive to implement?

[JMC] I would recomend looking at the latest WS-CDL draft
(http://www.w3.org/TR/ws-cdl-10/), where the relationship between
choreography and orchestration (the W3C view) is discussed. Search on
"The figure below demonstrates a possible usage of WS-CDL". If anyone
would like me to answer questions on the WS-CDL spec that are relevant
to our work, just let me know - I've absorbed the spec in its various
forms over the past several years and know it pretty well.

Interestingly, the terms "orchestration" and "choreography" do not
appear in the latest WS-BPEL draft (not that they have in the past

Just getting started but have to leave for another meeting :-)

[JMC] Hope this helps.



At 02:15 PM 3/15/2006, you wrote:
>Here's an article from May 2005: (I think you will recognize the
>L .xml&title=Articles%20:%20Get%20Ready%20for%20WS-BPEL
>Please note that the schedule for BPEL has since changed - the first 
>paragraph of this article reflects the schedule as conveyed by the BPEL

>TC chairs at the time of the article.
>Joseph Chiusano
>Booz Allen Hamilton
>700 13th St. NW, Suite 1100
>Washington, DC 20005
>O: 202-508-6514
>C: 202-251-0731
>Visit us online@ http://www.boozallen.com
>-----Original Message-----
>From: Ken Laskey [mailto:klaskey@mitre.org]
>Sent: Wednesday, March 15, 2006 2:06 PM
>To: Francis McCabe
>Cc: soa-rm-ra@lists.oasis-open.org
>Subject: Re: [soa-rm-ra] [soa-ra] why or why not composition needs to 
>I haven't gone through any details of BPEL in a long time but my 
>uninformed instinct is that BPEL is a small and insufficient subset of 
>what is needed.
>Is there a good BPEL reference (or a stable spec) to look at to confirm

>or disprove my assumptions?  I'd like to look at some BPEL info and 
>then try to further flesh out the traffic cop in an RA context.
>At 01:51 PM 3/15/2006, Francis McCabe wrote:
> >Ken:
> >  What you are talking about is orchestration. The traffic cop is (of
> >course) the key driver for BPEL.
> >  At the moment I am not I would want to mandate that all SOAs have 
> >this kind of organization. At the same time we should be able to 
> >clearly identify where in the RA such a cop would sit (stand) 
> >directing
> >the traffic.
> >  There are some echoes here with intermediary processing too..
> >Frank
> >
> >On Mar 15, 2006, at 7:49 AM, Ken Laskey wrote:
> >
> >>My apologies for falling back to email rather than the wiki but I am

> >>trying to get some percolating ideas down quickly in a familiar 
> >>format.
> >>
> >>Composition needs to be looked at in the context of not just how you

> >>would specify the combination of more atomic services into a higher 
> >>level one, but also the behavior models and the 
> >>orchestration/choreography that results in generating real world 
> >>effects.  An implemented SOA needs a "traffic cop" process/service 
> >>to
> >>- accept requests,
> >>- decide what needs to be done with the request,
> >>- rout the request or derived requests to other services,
> >>- collect the results of other services,
> >>- decide on next actions based on behavior model and responses to 
> >>routed requests,
> >>- continue this until the initial request is satisfied or 
> >>terminated,
> >>- package and send results to receiver designated by original 
> >>requester (where receiver possibly *not* the requester).
> >>
> >>The traffic cop will make use of known compositions to make sure 
> >>routing is done properly and for error recovery.  Whatever the 
> >>traffic
> >>cop decides, it will probably be captured internally as a service 
> >>composition and corresponding behavior model.  This is likely a 
> >>useful
> >>format for logging and future audits.  It also gives a basis for 
> >>evaluating levels of service and identifying bottlenecks.
> >>
> >>So in summary, I think composition is a fundamental part of the 
> >>reference architecture but in the context of how the composition 
> >>description is used rather than the description itself.
> >>
> >>Ken
> >>
> >>
> >>--------------------------------------------------------------------
> >>--
> >>--------------------
> >>Ken Laskey
> >>MITRE Corporation, M/S H305     phone:  703-983-7934
> >>7515 Colshire Drive                        fax:        703-983-1379
> >>McLean VA 22102-7508
> >
>    /   Ken
>Laskey                                                                \
>   |    MITRE Corporation, M/S H305    phone:  703-983-7934   |
>   |    7515 Colshire Drive                    fax:      703-983-1379
>    \   McLean VA 22102-7508

   /   Ken 
Laskey                                                                \
  |    MITRE Corporation, M/S H305    phone:  703-983-7934   |
  |    7515 Colshire Drive                    fax:      703-983-1379   |
   \   McLean VA 22102-7508

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]