RE: [soa-rm] Sub Comittee and/or User guide

Duane,

Respectfully, I wonder if we’re splitting hairs here. IMHO, many would consider “reference model” and “reference architecture” to be synonymous. If they are not, it begs the question: How many steps does one *really* need to go through to finally implement a service-oriented architecture? Are we really making this easier for the world, or more difficult? I say the latter.

For this reason, I (as only one member of this TC) would not support the notion of a new SC to tackle a more useful and actionable product than the one it seems we are heading toward. I would recommend that we revisit our interpretation of the definition of “reference model”, and – if necessary – call a vote to determine what the TC members really believe the end product of this TC should be. My concern is that this is becoming more of an academic exercise than a practical one (but I would very much like to be proven wrong).

Some (hopefully) supporting points:

- The description of this TC on the OASIS site is: “Developing a core reference model to guide and foster the creation of specific, service-oriented architectures”. According to how abstract our reference model is turning out to be, I don’t believe that we are fulfilling that description (“guide and foster”)

- Here is the definition of Reference Model in our charter, reproduced here for emphasis: “Reference Model - A reference model is an abstract framework for understanding significant relationships among the entities of some environment, and for the development of consistent standards or specifications supporting that environment. A reference model is based on a small number of unifying concepts and may be used as a basis for education and explaining standards to a non-specialist.”

IMHO, I for one do not believe that including the concepts of “service consumer” and “service producer” would violate this definition – however, I do recognize that the definition is subjective.

- I’ve searched the Web for some reference models that exist – here are some, along with my comments :

(1) NASA CCSDS Reference Model for an Open Archival Information System (OAIS): http://www.ccsds.org/documents/650x0b1.pdf

- From p.10: This reference model:

o provides a framework for the understanding and increased awareness of archival concepts needed for Long Term digital information preservation and access;

o provides the concepts needed by non-archival organizations to be effective

participants in the preservation process;

o provides a framework, including terminology and concepts, for describing and comparing architectures and operations of existing and future archives;

o [etc.]

- Note second bullet above “provides the concepts needed”.

- Also, from p.30: See Figure 2-4: IMO, this figure provides enough details to make the reference model useful and actionable.

(2) WfMC Workflow Reference Model:

http://www.wfmc.org/standards/model.htm

- This figure includes components such as: Workflow Client Application, Invoked Applications, Process Definition etc. According to what I’ve been hearing up to now by some, these would be “too concrete”. Please clarify then if you believe this WfMC model is in your opinion truly a reference model. IMHO, it seems to be highly useful and actionable.

(3) Also the OSI Model: I know that we’ve been making repeated references to this model (pardon the pun), but many people do believe that the OSI model failed and TCP/IP triumphed (not sure if that changes our perspective). Also, I don’t believe that the OSI model was intended for the same type of purpose as we are intending here. The OSI model’s mission was quite narrowly scoped, while ours – by virtue of the fact that we are about service-oriented architecture – is much broader. Different treatments for different missions

So in summary, I would very much like us to consider loosening the constraints on our reference model just a bit, to be able to include (or at least consider at length) useful components such as security, orchestration, messaging, etc. Let's make our product (a) useful and (b) actionable.

My $0.02 :)

Joe

All:

I have read through the last batch of email. There are a couple of
things I would like to propose for comments. Please read this entire
email before replying.

1. I will concede that many members of the public will likely have the
same kinds of trouble interpreting a reference model vs. a reference
architecture vs a specific architecture as seems to be pervasive on this
list. If we have the problem in our context, it is likely to be present
outside of this list.

2. We cannot redefine what a "reference model" is or what it includes.
If we tried to change the industry definition of reference model to one
that has concrete items in it or things that are not part of SOA (which
is tricky since it is still undefined), it will not be a true reference
model and hence not accepted by industry.

3. Service provider and service consumer are not part of a reference
model. They are roles visible only in a runtime or infrastructure views
of a specific architecture. To prove this point, please look once again
at the OSI reference model. It is a communications stack RM yet does
not contain notions of a message sender and message receiver.

4. We cannot mix abstract concepts and "things people can chew on"
(implying concrete items) in our work. Such does run adverse to
accepted architectural conventions.

SOLUTION:

One way forward is to probably create a sub committee to work on a
reference architecture for SOA. A Reference Architecture could be
developed in parallel to the reference model and is fair game to
illustrate things like security, consumers, providers, agents etc. It
is within our charter to do such.

After reading through some older emails, I would assert that such a
thing is probably essential along with some sort of white paper or user
guide that explains the relationships between the RM, the RA and other
architecture.

Reference Model
(is a guide for developing a)
[ Reference Architecture || * Architecture ]

There are several people on this list who also have stated specific
needs for what they see in SOA. Perhaps this may be a good Sub
Committee (SC) consideration also.

Government Service Oriented Reference Architecture???
etc.

I can already see there are many of you who could lead such an effort as
a sub committee.

Comments?

Duane

--
***********
Senior Standards Strategist - Adobe Systems, Inc. - http://www.adobe.com
Vice Chair - UN/CEFACT Bureau Plenary - http://www.unece.org/cefact/
Adobe Enterprise Developer Resources - http://www.adobe.com/enterprise/developer/main.html
***********

soa-rm message