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] Some initial reflections and questions

I agree absolutely and I think we addressed this adequately on the call today.





From: mpoulin@usa.com [mailto:mpoulin@usa.com]
Sent: Wednesday, 01 December, 2010 02:37
To: peter-oasis@justbrown.net; jeffrey.a.estefan@jpl.nasa.gov
Cc: soa-rm-ra@lists.oasis-open.org
Subject: Re: [soa-rm-ra] Some initial reflections and questions


Just to share some thought: when I deal with IEEE 1471, I am trying to be very accurate because too many people take the viewpoint method as a form of an architecture definition while it is only an architecture description methodology. (IEEE 1471 has an architecture definition as well and it takes on sentence only) 


As a consequence of misreading of IEEE 1471 (which even directly states that it is a set of Recommendations for description of architecture), the view on or external viewpoint is taken as an element of definition which is wrong (looking outside-in with no knowledge about the 'in' part there is a very high chance to a) consider a piece of outside environment as a part of the architecture; 2) add subjective opinions to the view).


I'd recommend to keep in mind the inside-out approach regarding viewpoints and views. In this way  all viewpoints aim the interests of stakeholders with only internal, truthful information. For example, an expression 'Service Viewpoint' would mean how SOA considered/defined Service for different needs of stakeholders rather than how they might see a service for their needs.

- Michael

-----Original Message-----
From: Peter F Brown (OASIS Individual Member) <peter-oasis@justbrown.net>
To: 'Estefan, Jeff A (3100)' <jeffrey.a.estefan@jpl.nasa.gov>
Cc: soa-rm-ra@lists.oasis-open.org
Sent: Wed, Dec 1, 2010 12:36 am
Subject: RE: [soa-rm-ra] Some initial reflections and questions


Thanks for sharing this. What I take away from these comments is:

-          It remains important to highlight the concept of ‘ecosystem’ and its central importance in SOA (because of the distributed nature of SOA systems, ownership boundaries and interactions, etc);

-          We need perhaps another term/label for the first ‘viewpoint’ (currently labelled as ‘Service Ecosystem’) and the concept it represents;

-          The two concepts are possibly orthogonal and certainly not the same ‘thing’ – which addresses my concern expressed earlier about “an ecosystem can’t be both the sum of a set of IEEE 1471 concepts and a viewpoint”


I’m still struggling with another key question, maybe for discussion tomorrow: is our work product,

-          a ‘reference architecture’?;

-          the ‘(foundational) part of a reference architecture’?;

-          ‘building blocks for (an SOA) reference architecture?

Or something else? What’s the problem that we are addressing? That there are no (or only bad) RA’s ‘out there’ and we need to come up with (a new) one? That any RA for a specific SOA system ought to use some common models, and this provides those? Or that our product provides the foundation for any and all SOA RA’s worth their salt?


Clarity on this would help me immensely and guide my weeding…


Cheers, and talk tomorrow,



From: Estefan, Jeff A (3100) [mailto:jeffrey.a.estefan@jpl.nasa.gov]
Sent: Tuesday, 30 November, 2010 14:30
To: soa-rm@lists.oasis-open.org; soa-rm-ra@lists.oasis-open.org
Subject: Re: [soa-rm-ra] Some initial reflections and questions


Peter and Chris,


A little background …


The choice to use 1471 (now 42010) was originally mine and I stand by that decision.




Because what we were about to embark on 4+ years ago was not producing an architecture per se but rather an architecture description (a work product).  Recall that all systems have an architecture whether described or not.


Our objective, as I recall, was to produce an abstraction of the SOA paradigm (again, in the form of an architecture description) where we would use standards-based visual models and supporting text to describe the key elements and their relationships needed to enable SOA-based systems to be "used," "realized," and "owned."  These three perspectives formed the basis of our viewpoints.


With respect to viewpoint 1, the original name was "Business Viewpoint" with the associated "Business View." The viewpoint was subsequently renamed to “Service Ecosystem Viewpoint” (see why in the next paragraph).  I would prefer we call it the "Using SOAs Viewpoint" and associated "Using SOAs View," which would be consistent with the original scope of this body of work.


Later in our subcommittee progress, it was suggested that the RA go beyond the RM to suggest the paradigm of SOA can be characterized as an ecosystem-based paradigm.  Unfortunately, the term "ecosystem" is not used consistently in the current draft of the spec.  Sometimes we call it "SOA Ecosystem" and sometimes we call it "Service Ecosystem."  We need to be consistent.  This conceptual mapping came later in the development of the spec and is what found its way into renaming viewpoint 1 the "Service Ecosystem Viewpoint."  On reflection, this seems an erroneous decision. Again, I suggest "Using SOAs Viewpoint" as a possible alternative.


These viewpoints, while foundational for our architecture description, were not intended to be representative of a Library of Viewpoints to be used by other concrete or reference architectures.  They were crafted specifically for this RA, which we felt (and others like our Open Group friends) that what we were producing was foundational in nature with specific focus on describing (architecturally) the paradigm of SOA (See The Open Group’s architecture continuum at http://www.opengroup.org/architecture/togaf9-doc/arch/chap39.html#tag_40_04_01).  I would say that the content of the views themselves and in particular the Architectural Implications have been designed to be used by other concrete or reference architectures.  I also agree that "weeding" is in order if we find content that does not meet the spirit of that intended purpose and scope. 


With respect to the 1471/42010 conceptual model, there are a couple of sources in the open literature and even TOGAF V9 has one (also in the open literature), but I believe we may still need to obtain permission for its use in our spec.  You may recall somebody asked us to remove the full 1471 spec from our Kavi site.


Hope this helps “demystify” some of the decisions and how we got to where we are today.




- Jeff




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