RE: [soa-rm-ra] RE: Report on Use of Orchestration/Choreography in NCOIC

Michael,

I think we all agree that choreography should not indicate a tight coupling of tasks, even if across trust boundaries. I thought your modification of the choreography model made sense, although I wasn’t sure we need the proliferation of interfaces you show.

I like the idea in your proposal of new models that highlight the conductor. I’ve got some vague ideas that I will hopefully get on paper this weekend, but I encourage others to post their ideas.

The question to Rex will be how much of the old models need to be preserved to maintain continuity with our audience.

Ken

---------------------------------------------------------------------------

Dr. Kenneth Laskey

MITRE Corporation, M/S H305 phone: 703-983-7934

7515 Colshire Drive fax: 703-983-1379

McLean VA 22102-7508

From: soa-rm-ra@lists.oasis-open.org [mailto:soa-rm-ra@lists.oasis-open.org] On Behalf Of Mike Poulin
Sent: Friday, January 06, 2012 10:19 AM
To: soa-rm-ra@lists.oasis-open.org
Subject: Re: [soa-rm-ra] RE: Report on Use of Orchestration/Choreography in NCOIC Members

I'd like to comment on Rex's post becuase a) I expected to hear from him about the military audience as the major consumer of choreography model, and b) existing practice of choreography does not necessary fit with SOA ecosystem and service orientation.

1. I agree that
business collaboration (when the shared task becomes the primary task of the participant and the latter is willing to adjust its internal business logic / business services just for the sake of solving this task) ==> choreography
and I assume that my comment about willingness to change is equally applicable to choreography

business process AND aggregation (when the shared task becomes an additional task for the participant who intents to solve it by existing means with minimal or no changes in its internal business logic / business services) = business cooperation = orchestration

2. I can accept that in many cases, especially in military or legacy B2B integration, internal business processes of the participants were coupled for the choreography. Though, this couplong is unacceptable in SOA. This is what I tried to illustrate in the diagrams by placing interfaces between participants.

If Rex insists on understanding of collaboration/choreography as on possible copling of internal processes, the only what is left for me is to say that collaboration/choreography is not suitable model for SOA.

If Rex agree that business services / participants participate in the choreogrpagy as services (with no coupling of internal implementation/processes) AND that the only principal difference of choreography from orchestration is a rule-holder orchestration conductor, I 100% agree with him.

3. If collaboration tasks come infrequently and last relatively long, it may worth changing business services for the purpose of collective work [legistics planning for a lont-time deployment of troops, e.g.].
However, if collaboration tasks come frequently and the entity needs/wants to participate in several of them simultaneously or if tasks of collective work (collaboration ?) appear dynamically depending on the changes in the business execution context [the same legistics but during active movements of the troops in unpredicted situations], the changes of the internal business logic / business services becomes inefficient and impractical.

Propoal:

instead of existing diagrams or the ones I've drawn, let's draw a new pair of diagrams where one would reflect the role of conductor in orchestration (with no references to business process) AND another one would reflect an independent (not-changed) participnts that hold its own rules integrated a priori.

"Our clarification of Activity as a collection of actions and joint actions is pretty certain to please that audience" can nicely work with thses new diagrams.

The model "picture" for orchestration may look like a star-network layout while the "picture" for choreography may look like a mesh-network layout.

- Michael

----- Original Message -----

From: Ken Laskey

Sent: 01/05/12 04:01 PM

To: 'Rex Brooks', soa-rm-ra@lists.oasis-open.org

Subject: [soa-rm-ra] RE: Report on Use of Orchestration/Choreography in NCOIC Members

Rex,

Thanks for the quick input.  Some questions come to mind:

1. How do the NCOIC folks interpret Figures 30 and 31?  Michael had some

particular concerns when applying this in his work.  Boris also had issues.

Can you get us any insight into how the NCOIC folks make use of this?

2. We didn't finish going through Michael's choreography mods but answer to

1 will help us determine what changes (either Michael's or otherwise) may be

useful and what needs to stay the same

3. My realization yesterday (maybe later than some of the group) that

Figures 30 and 31 may be useful in showing examples of compositions but I'm

not sure they really provide a useful illustration of the distinction we try

to make in the text.  If we need to keep Figures 30 and 31 pretty much as

they are, do we need additional models?

Ken

---------------------------------------------------------------------------

Dr. Kenneth Laskey

MITRE Corporation, M/S H305              phone: 703-983-7934

7515 Colshire Drive                                    fax:

703-983-1379

McLean VA 22102-7508

-----Original Message-----

From: Rex Brooks [mailto:rexb@ncoic.org]

Sent: Thursday, January 05, 2012 9:39 AM

To: soa-rm-ra@lists.oasis-open.org; Ken Laskey

Subject: Report on Use of Orchestration/Choreography in NCOIC Members

Hi Ken, Everyone,

I wanted to get word to you all immediately that we DO have multiple

companies with internal implementations of the

Orchestration-Choreography section in the NCOIC Services WorkGroup and

the diagrams in particular, so I will be required to fight for retaining

it with only the changes we discussed, and keep the diagrams with the

minimal changes discussed. I don't have much concern that there will be

pushback on our additional note that composition, especially in

aggregations of multiple services will likely include orchestrations and

choreographies in various combinations as needed.

I'm writing now because this feedback already convinces me to insist on

keeping the diagrams. I have not yet asked the military audience, which

I was most concerned about with the DoDAF Metamodel work. Our

clarification of Activity as a collection of actions and joint actions

is pretty certain to please that audience.

Sorry if this upsets any applecarts, but dropping the diagrams entirely

would negatively impact the constituencies I work with most. We also

need to specifically address the comment submitted by Dr. Allen Jones

for Boeing.

Cheers,

Rex

soa-rm-ra message