RE: [soa-rm] Groups - Generic SOA-RM Boot Camp(SOAReferenceModel-Generic-rev1_05-03.ppt) uploaded

["Smith, Martin" <Martin.Smith@DHS.GOV>]

Duane - - 

Thanks for the response.

What I meant by "distributed" is adequately captured by your paraphrased
definition.  This could mean separate computers on a LAN, but the more
interesting case is applications running across the Internet.  Not only
does the Internet involve different systems, but also different
management domains. Dealing with different management domains is an
important part of the core problem SOA is trying to address.  In fact,
the move to the Internet as the typical implementation environment seems
to be the reason that our terminology evolved from "component-based
architecture" to "services-oriented architecture."

Regarding BPM or process management, I'm not that fond of the
Google-sources definition you came up with.  In the framework of
distributed apps, BPM plays the role of business logic connecting large
chunks of encapsulated functionality (the services).  I suppose it's
possible to have a useful single-service app, but surely in the general
case we're connecting multiple services somehow . . .  (And again, I
think we need to scope the RM to cover the common cases, not the minimum
case excluding optional elements.)

Thanks,

Martin


-----Original Message-----
From: Duane Nickull [mailto:dnickull@adobe.com] 
Sent: Tuesday, April 19, 2005 12:37 PM
To: Smith, Martin
Cc: 
Subject: Re: [soa-rm] Groups - Generic SOA-RM Boot
Camp(SOAReferenceModel-Generic-rev1_05-03.ppt) uploaded

Martin:

These are good points we need to understand.  Can I ask for 
clarification on a few points.

(1) What is your definition of "distributed"?  This will probably be 
important to reach consensus on in our TC.  I googled "define: 
distributed system" and found these definitions:
http://www.google.com/search?q=define%3A+distributed+system&btnG=Google+
Search

Paraphrased:  Distributed System:  an aggregation of functionality from 
several computing entities to collaboratively run a computational task 
in a transparent and coherent way, so that they appear as a single, 
centralized system.

Some of the definitions talked specifically about being distributed on a

network (not good for RM).  The physical separation aspect is not 
something I am clear about WRT being distributed.  Can a chip 
hyperthread two processes and aggregate those at a higher level to 
appear as an opaque computational task?  Perhaps distributed means 
logical separation?  Not sure. 

(2)  Even though we have not made a decision WRT BPM in the RM, do you 
think we could use a more abstract term for it such as "Management" or 
"Process Management"?  This is primarily in recognition of the fact that

not all SOA's are built for businesses.  The concept can be added to our

glossary and will need to be defined.  I also used google to find some 
starter suggestions.  I paraphrased several as:

Process Management "..a multi-layered approach to the orchestration of 
lower activities, which produces a result for a given customer at a 
higher level."

Would this work for folks?  Comments?

Duane


Smith, Martin wrote:

>Duane - - 
> 
>This a a good sort of thing to have available - - sort of a graphical
FAQ.  Thanks.
> 
>Couple of substantive comments:
>(1) I think many of us are scoping the work too narrowly: a "services
RM" does not solve what I (at least) need, which is a "services-oriented
distributed application" reference model. I had always assumed that when
people said "SOA" they meant the latter.  This draft seems to define the
former.
>(2)  I don't think the RM should necessarily be limited to elements
that are mandatory in all implementations.  You suggest a formulation
that avoids this for security (i.e, null security policy is a security
policy); Rebecka used a similar device in observing that all mechanisms
in a software environment can be defined as services and therefore we
might not need to include any particular "SOA infrastructure" services
separately in our RM.  I would prefer to be explicit in getting a
maximum set of elements on the table (BPM for example) and then coming
back later to refactor or simplify the model. Otherwise we may end up
with an elegantly simple but trivial product, meaning one that does not
provide any useful guidance to architects.
> 
>Martin
> 
>
>________________________________
>
>From: dnickull@adobe.com [mailto:dnickull@adobe.com]
>Sent: Fri 4/15/2005 4:29 PM
>To: soa-rm@lists.oasis-open.org
>Subject: [soa-rm] Groups - Generic SOA-RM Boot Camp
(SOAReferenceModel-Generic-rev1_05-03.ppt) uploaded
>
>
>
>This is a draft submission to help orient new members as they join.
Please
>review and comment on.  Want to ensure that we are in basic agreement
on
>these items.
>
> -- Duane Nickull
>
>The document revision named Generic SOA-RM Boot Camp
>(SOAReferenceModel-Generic-rev1_05-03.ppt) has been submitted by Duane
>Nickull to the OASIS SOA Reference Model TC document repository.  This
>document is revision #1 of SOAReferenceModel-Generic-rev1_05-03.ppt.
>
>Document Description:
>A draft presentation for orienting new members.
>
>View Document Details:
>http://www.oasis-open.org/apps/org/workgroup/soa-rm/document.php?docume
nt_id=12276
>
>Download Document: 
>http://www.oasis-open.org/apps/org/workgroup/soa-rm/download.php/12276/
SOAReferenceModel-Generic-rev1_05-03.ppt
>
>Revision:
>This document is revision #1 of
SOAReferenceModel-Generic-rev1_05-03.ppt.
>The document details page referenced above will show the complete
revision
>history.
>
>
>PLEASE NOTE:  If the above links do not work for you, your email
application
>may be breaking the link into two pieces.  You may be able to copy and
paste
>the entire link address into the address field of your web browser.
>
>-OASIS Open Administration
>
>
>  
>

-- 
***********
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
***********