Martin –
What would be providing the orchestration,
workflow, etc within the SOA if not another service? If that is the case,
the RM should really abstract out to dealing with a service, which may or may
not orchestrate other services. That orchestration, however, would become
part of the functionality that is implied by the service contract or shaped by
the service policy.
It seems like the real question is whether
an orchestration concept that is independent of the orchestrated services is a
valid piece of the RM.
Rebekah
Rebekah Metz
Associate
Booz Allen Hamilton
Voice: (703) 377-1471
Fax: (703) 902-3457
From: Smith, Martin
[mailto:Martin.Smith@DHS.GOV]
Sent: Tuesday, April 12, 2005 3:24
PM
To: Chiusano Joseph; Duane Nickull
Cc: soa-rm@lists.oasis-open.org
Subject: RE: [soa-rm] RE:
Definition of Reference Model (Was RE: [soa-rm] Definition of "Service
Consumer")
Joe - -
I agree that all implementations
shouldn’t have to include all elements, or, in other words, the RM should
not be limited to elements that are included in all implementations.
I still think that both services
composition (orchestration, workflow, program logic, BPM . . . ) should
be an element of the RM. It’s not going to be a very useful or
interesting environment that is limited to interactions with one service.
I also think security should be a core
element: it may not be necessary to have security on trivial, free
services, but all the interesting cases will have it, and the RM should inform
how that element will interact with others.
Also, I saw someone refer to our task as
defining a RM for “distributed systems” (or a particular style.)
I think that is why we are all here, and using that as the
“it” we are trying to describe would bring some focus, IMHO.
Regards,
Martin
-----Original Message-----
From: Chiusano Joseph
[mailto:chiusano_joseph@bah.com]
Sent: Tuesday, April 12, 2005 1:37
PM
To: Duane Nickull
Cc: soa-rm@lists.oasis-open.org
Subject: [soa-rm] RE: Definition
of Reference Model (Was RE: [soa-rm] Definition of "Service
Consumer")
Thanks
Duane. Just to clarify, I was not questioning the fundamental role and nature of
reference models (i.e. it's understood that the are not implementable directly,
and that they do not go beyond the level of detail than you specified below).
What I was questioning was our *interpretation* of that definition, in that our
interpretation may be considered as too restrictive.
I also
have not seen a univeral definition of reference model (or an example of
one) that states that something has to be absolutely mandatory (even if it is
related, such as security, messaging, etc.) in order to be included in a
reference model. IMHO, a reference model that does not provide - well - value
for reference because it is too restrictive, or too abstract (too far from
being implementable* in the real world) is not a valuable reference model.
As always, I really appreciate your
insight.
Kind Regards,
Joseph Chiusano
Booz Allen Hamilton
*meaning in this case, for creating SOA architectures
- not SOA implementations directly - from it
From: Duane
Nickull [mailto:dnickull@adobe.com]
Sent: Tue 4/12/2005 1:19 PM
To: Chiusano Joseph
Cc: soa-rm@lists.oasis-open.org
Subject: Re: Definition of Reference
Model (Was RE: [soa-rm] Definition of "Service Consumer")
Joseph:
We should not re-define what a reference model is. The term has
industry accepted connotations. A Reference Model is NOT implement able
directly. It does not include the necessary level of detail to build
anything with other than architecture. If it did, then by definition it
would not be a reference model. A reference model is also not
restrictive by nature. You may add more things to the core set as you
require them.
Your email does bring up an interesting and recurring theme however.
Perhaps in addition to a reference model, it would be useful to deliver
a user guide/framework for development. We may have to consider at
least a separate FAQ or similar as we progress. I do suspect however,
that most software architects understand how to use a RM.
There will also be plenty of opportunity to build architecture in this
TC as a Sub TC. I had a request yesterday for a group to start a Sub TC
within this TC to concentrate on a concrete architecture for a very
specific field, using the SOA RM as their guide. I hope that we have at
least 2-3 of these as time goes on.
Other comments inline:
Chiusano Joseph wrote:
>This definition does not state that only "mandatory" concepts are
>included, but rather implies it through its reference to "a small
number
>of unifying concepts". I recommend that at this point, since our
>definition of "reference model" is being used as a filter for
everything
>- and so many things are (perhaps justifiably!) being "kicked
out" of it
>- we gain consensus on what we believe "reference model" means.
>
If something is not a unifying concept for all things SOA, it should not
be in the RM however this does raise the chicken and egg question. How
do we determine if something is SOA before the RM is complete? There is
no quantitative way to determine if something is or isn't SOA.
Accordingly, the types of discussions we are having are very healthy
IMO. I enjoy the fact that we have all types of opinions on this list
to consider.
>I bring this up because I have seen cases where reference models are not
>as restrictive as the one that we are heading toward (but maybe they
>should be?). I am also (frankly) concerned that we might be developing
>something that is - and I don't mean this in a pejorative sense - too
>"academic" and not implementable in real-world settings.
>
A RM by definition is abstract, therefor you may correctly imply
"academic". It is very useful to those who develop
architecture.
>I also wonder if it might be more valuable process-wise to concentrate
>on concrete architectures, and then determine from among those concrete
>architectures what concepts are mandatory, what are optional etc. - that
>is, to build our reference model from outward-in, instead of the
>opposite which I believe is our current process.
>
>
This is what we have been and should continue to do. We are scouring
all things claiming to be or perceived to be SOA and trying to determine
what are the common elements in those, how to describe them in an
abstract manner and what the relationships amongst those things are.
Duane
>Thoughts?
>
>Joe
>
>Joseph Chiusano
>Booz Allen Hamilton
>Visit us online@ http://www.boozallen.com
>
>
>
>
>>-----Original Message-----
>>From: Ken Laskey [mailto:klaskey@mitre.org]
>>Sent: Monday, April 11, 2005 9:02 PM
>>To: Matthew MacKenzie
>>Cc: 'Frank McCabe'; Duane Nickull;
>>soa-rm@lists.oasis-open.org; vikas@sonoasystems.com; Chiusano
>>Joseph; 'Andrew Nash'
>>Subject: Re: [soa-rm] Definition of "Service Consumer"
>>
>>Let's leave this as an open issue, if we may. Except for a
>>very simple, very closed system, I cannot imagine a viable
>>SOA in a real environment without security. I am willing to
>>be educated about situations where security can legitimately
>>be skipped, but I don't think it can be left out of a useful RM.
>>
>>Ken
>>
>>On Apr 11, 2005, at 3:32 PM, Matthew MacKenzie wrote:
>>
>>
>>
>>>I don't believe that all SOAs do or will have security. I
think we
>>>should simply not mention it. This is, after all, an abstract
>>>reference model. We can produce the warmNfuzzy that having
>>>
>>>
>>a security
>>
>>
>>>component adds in our own SOA designs that are identifiable
>>>
>>>
>>with the
>>
>>
>>>SOA-RM.
>>>
>>>-matt
>>>
>>>On 11-Apr-05, at 12:28 PM, Duane Nickull wrote:
>>>
>>>
>>>
>>>>Ken:
>>>>
>>>>I am not 100% sure about this. I would like to research
this on a
>>>>more philosophical basis. Not all SOA's use explicit
security
>>>>protocols (the internet doesn't). The fundamental
philosophical
>>>>question may be " does the explicit statement conveying
>>>>
>>>>
>>the absence
>>
>>
>>>>of any security still imply a security model"?
>>>>
>>>>The danger in saying "yes" is that it opens the door
for more
>>>>"things" to be part of the RM.
>>>>
>>>>I would like to mull this over and do some research. I am
>>>>
>>>>
>>sure Matt
>>
>>
>>>>has a good answer ;-)
>>>>
>>>>Duane
>>>>
>>>>Ken Laskey wrote:
>>>>
>>>>
>>>>
>>>>>Moreover, the question is whether all SOAs SHOULD have
>>>>>
>>>>>
>>security and
>>
>>
>>>>>whether that needs to be captured in the RM. As
noted,
>>>>>
>>>>>
>>secuirty is
>>
>>
>>>>>often just tacked on and that may not be sufficient for
>>>>>
>>>>>
>>*any* SOA to
>>
>>
>>>>>be successful.
>>>>>
>>>>>Ken
>>>>>
>>>>>At 02:27 PM 4/11/2005, Duane Nickull wrote:
>>>>>
>>>>>
>>>>>
>>>>>>The RM does not support security models. A
reference
>>>>>>
>>>>>>
>>model is used
>>
>>
>>>>>>to guide the design of architecture that may include
specific
>>>>>>security protocols or models. Our requirement must be
to ensure
>>>>>>that nothing we place in the RM makes any specific
>>>>>>
>>>>>>
>>security model a
>>
>>
>>>>>>requirement (since not all SOA's have security) and to
>>>>>>
>>>>>>
>>ensure that
>>
>>
>>>>>>we do not preclude a specific type of security model
from being
>>>>>>used.
>>>>>>Duane
>>>>>>
>>>>>>Vikas Deolaliker wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>>I think the question should be how many different
types of
>>>>>>>security models will this RM support?
>>>>>>>Vikas
>>>>>>>
>>>>>>>--
>>>>>>>
>>>>>>>
>>>>>>--
>>>>>>***********
>>>>>>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
>>>>>>***********
>>>>>>
>>>>>>
>>>>>>
>>>>>--
>>>>>
>>>>>
>>>>>
>>--------------------------------------------------------------------
>>
>>
>>>>>-
>>>>>------------
>>>>> / Ken
Laskey
>>>>>
>>>>>
>>
>>
>>
>>>>>
\
>>>>> | MITRE Corporation, M/S
H305 phone: 703-883-7934 |
>>>>> | 7515 Colshire Drive
fax:
>>>>>
>>>>>
>>703-883-1379
>>
>>
>>>>> |
>>>>> \ McLean
VA 22102-7508
>>>>>
>>>>>
>>
>>
>>
>>>>> /
>>>>>
>>>>>
>>>>>
>>>>>
>>--------------------------------------------------------------------
>>
>>
>>>>>-
>>>>>-------------
>>>>>
>>>>>
>>>>>
>>>>>
>>>>--
>>>>***********
>>>>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
>>>>***********
>>>>
>>>>
>>>>
>>>
>>>
>>--------------------------------------------------------------
>>----------
>>------------------
>>Ken Laskey
>>MITRE Corporation, M/S H305 phone:
703-883-7934
>>7515 Colshire Drive
fax: 703-883-1379
>>McLean VA 22102-7508
>>
>>
>>
>>
>
>
>
--
***********
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
***********
|