RE: [ogsa-wg] GGF-DMTF Work Register

Fred, Just to be clear ... You say "A resource model has semantics and a rendering. The semantics are just, say, a UML model with some textual descriptions. CIM itself only has semantics." This is indeed true - but UML classes (and CIM) have both properties and operations. The operations are behavior - and while their implementation is not specified, their parameters are defined. Andrea
-----Original Message----- From: owner-ogsa-wg@ggf.org [mailto:owner-ogsa-wg@ggf.org] On Behalf Of Fred Maciel Sent: Tuesday, August 02, 2005 11:30 AM To: Donal K. Fellows; Tom Maguire Cc: Ellen Stokes; ogsa-wg@ggf.org; Tom Roney Subject: RE: [ogsa-wg] GGF-DMTF Work Register
Hi Donal,
So my point is: 1) ANY resource model we use MUST have a normative xml expression 2) ANY resource model we use MUST have a normative wsdl expression
For specific, specific disciplines we undoubtedly will use
Tom Maguire wrote: portions of
CIM. We (OGSA) will undoubtedly use portions of other resource models (perhaps IETF for networking). All of the models we use MUST meet the normative low bar above 1 & 2.
I wasn't aware that resource models were inherently services. They're just a way of describing the terms used in some data relating to the description of a resource, surely? Or do you distinguish between purely passive descriptive models and active manipulable models of a resource?
I don't know if you were in the teleconference (we gave some more background on this there), so here is a rather short clarification.
A resource model has semantics and a rendering. The semantics are just, say, a UML model with some textual descriptions. CIM itself only has semantics. If you want to access or exchange information using the model semantics you need a rendering -- say, a mapping of the semantics over XML, plus a binding over a given network protocol, etc. I presume that by "services" and "active manipulable model" you mean the rendering. So, I agree that, the model [semantics] are not services.
What we are proposing to the DMTF is to create a rendering, but not down to the bindings. It's just an XML representation of CIM, plus *part* of the WSDL definition. The most concise way that I can explain this is: "everything that can be made common among multiple OGSA basic profiles".
Hope that helps,
Fred Maciel Hitachi America R&D

Andrea Westerinen (andreaw) wrote:
Fred, Just to be clear ... You say "A resource model has semantics and a rendering. The semantics are just, say, a UML model with some textual descriptions. CIM itself only has semantics."
This is indeed true - but UML classes (and CIM) have both properties and operations. The operations are behavior - and while their implementation is not specified, their parameters are defined.
Given that, a standardized rendering as a webservice would make sense and foster interoperability. (My original query was mainly because I'm more used to other data-models that are operation-less, purely describing an arrangement published through some mechanism unrelated to the model, so I forgot that CIM does more). Donal.
participants (2)
-
Andrea Westerinen (andreaw)
-
Donal K. Fellows