[ogsa-wg] GGF-DMTF Work Register
Tom Maguire
tmaguire at us.ibm.com
Tue Aug 2 09:40:39 CDT 2005
Comments inlined
owner-ogsa-wg at ggf.org wrote on 08/02/2005 09:47:30 AM:
> Thanks Tom,
>
> My comment in line <HK>.
> ----
> Hiro Kishimoto
>
> Tom Maguire wrote:
> > owner-ogsa-wg at ggf.org wrote on 08/02/2005 06:49:20 AM:
> >
> >
> >>Hi Tom,
> >>
> >>Thank you very much for your Milestone proposal.
> >>I will make no objection about this proposal but have some
> >>reservations.
> >>
> >>Even though CIM is the most promising and widely adopted resource model
> >>in the world, OGSA-WG is still discussing our strategic direction of
> >>resource modeling. Reasonable work process is something like:
> >>
> >>(1) Determine the strategic direction including resource model
> >> selection (CIM?).
> >>(2) List up and prioritize requirements on models.
> >>(3) Ask external experts to fulfill each requirement.
> >
> >
> > Question (1) is moot given that we are talking specifically about
> > the GGF-DMTF Work Register.
>
> <HK>
> I don't think so. It is very unlikely but if OGSA-WG chooses i.e. GLUE
> schema as our resource model, we don't need to put our names on Work
> Register document.
>
> Actually, we are going to work with DMTF but our strategic decision
> should come before Work Register.
> </HK>
My point is that we already have a work register in place with DMTF.
That work register is being updated to include xml namespace and schema.
The question of strategic or not is orthogonal and I do not think we
need to serialize behind that decision. The decision of what resource
model to use and where it is standardized is going to be discipline
specific.
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 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.
So given that we are likely to use portions of CIM and given that
we need normative XML and normative wsdl expressions. I am uncertain
what strategic decision needs to be further made that would effect
the work register.
> > Question (2) is irrelevant to the additional milestone since both the
> > OGSA-D milestone and the addtional milestone are 'common' mapping
> > requirements which are orthogonal to specific resource modeling
> > requirements
> > (eg. changes/additions/deletions of properties, operations,
associations or
> > classes).
> > That being said IMO we should be asking external experts (DMTF in this
> > context) to provide us with these high level constructs.
>
> <HK>
> I got your point.
> </HK>
>
> >>Since we will make (3) in parallel with (1) and (2), I see two risks;
> >>
> >>(a) We ask DMTF and they make it on time (Feb. 2006) but we are not
> >>ready to utilize them at that time or end up using only small part of
> >>their work, or
> >>(b) We ask DMTF and they ask us back which is the highest priority
> >>item since our request is broad range. But we cannot answer until
> >>we sort out (1) and (2).
> >
> >
> > On point (a) I suspect anyone using CIM uses only a small part of it so
> > I think that one is ok.
>
> <HK>
> Given you are asking a common mapping to wsdl, it makes sense.
> </HK>
>
> > On point (b) I would suggest that the OGSA-D
> > milestone is more time critical; so I would prioritize that above this
> > additional work. As I pointed out above this is 'common' mapping work
> > so I do not believe it is tied at all to the resource modeling work.
>
> <HK>
> Agreed.
> </HK>
>
> >>So, I propose not to specify deadline request for this milestone from
> >>us. Instead Let DMTF say when they might be ready with this work item.
> >>Then OGSA and DMTF discuss priority together.
> >
> >
> > I think a milestone without a date will be ignored. Further I am
fairly
> > certain that the DMTF TC will request a date.
>
> <HK>
> Yes, milestone should have date. I insist DMTF should pick or lead to
> pick date. It is ok you suggest a date though.
> </HK>
>
> >>And specific question/comment follows;
> >>
> >>
> >>> The DMTF will define and publicize abstract wsdl:portType for
> >>> classes inthe CIM model (February 2006)
> >>
> >>Do you ask all of CIM classes (there are tons of them) or small number
> >>of subset?
> >
> >
> > We are asking for a common mapping to wsdl
> >
> >
> >>> Each 'extrinsic' operation on a CIM class will be modeled
> >>
> >>What is 'extrinsic' operation?
> >
> >
> > Operations that are expressed in the CIM schema on CIM classes
> >
> >
> >>> as a wsdl:operation
> >>> Properties of the CIM class which are rendered in XML
> >>> schema (detailed in previous deliverable) will be
> >>
> >>What is "previous deliverable"? Is it wsdl:operation (above line) or
> >>namespace and XML schema (previous paragraph)?
> >
> >
> > I was assuming that the OGSA-D XML schema would be logically place
> > above/before this milestone/deliverable
> >
> >
> >>> available throughthe wsdl:types section
> >>> It is envisioned that some set of qualifiers will be mapped to xml
> >>> schema and wsdl constructs. The specifics of this mapping are
> >>> viewed as explicitly part of the work to render the model. As such
> >>> GGF has no specific requirements at this time but may have
> >>> requirements as the workproduct develops.
> >>> The intent of this deliverable is to make available a set of
> >>> standard wsdl:portTypes representing the CIM modeled aspect of
> >>> classes (extrinsics and properties only). It is envisioned that
> >>> these portTypes will be mixed in with additional capabilities from
> >>> Grid disciplines to support Basic Execution Services, Data
> >>> Virtualization and other areas as they are developed.
> >>
> >>p.s.
> >>Since WS-CIM call this week is canceled, we have time to think it over.
> >
> >
> > This is not a discussion for WS-CIM directly. It is a discussion with
the
> > DMTF TC and they meet today. Further, any work register requires DMTF
> > board approval and the board meets on 8/11. After that our next board
> > meeting is in September.
> >
> > Tom
> >
> >
> >>Thanks,
> >>-----
> >>Hiro Kishimoto
> >>
> >>Tom Maguire wrote:
> >>
> >>>Folks,
> >>>On the OGSA-WG telecon 8/1 I took an AI to add a milestone/deliverable
> >
> > to
> >
> >>>the GGF-DMTF Work Register. Specifically, the milestone reflects the
> >>>urgent desire of the OGSA-WG to have a normative WSDL for classes
> >
> > expressed
> >
> >>>in the CIM schema. This milestone builds on the current addition to
> >
> > the
> >
> >>>work register from OGSA-D; namely the namespace and xml schema for the
> >>>properties of CIM classes. The addition to the work register being
> >>>proposed is as follows:
> >>>
> >>> The DMTF will define and publicize abstract wsdl:portType for
> >
> > classes in
> >
> >>> the CIM model (February 2006)
> >>> Each 'extrinsic' operation on a CIM class will be modeled
> >
> > as a
> >
> >>> wsdl:operation
> >>> Properties of the CIM class which are rendered in XML
> >
> > schema
> >
> >>> (detailed in previous deliverable) will be available
> >
> > through
> >
> >>> the wsdl:types section
> >>> It is envisioned that some set of qualifiers will be mapped to xml
> >>> schema and wsdl constructs. The specifics of this mapping are
> >
> > viewed as
> >
> >>> explicitly part of the work to render the model. As such GGF has
no
> >>> specific requirements at this time but may have requirements as the
> >
> > work
> >
> >>> product develops.
> >>> The intent of this deliverable is to make available a set of
> >
> > standard
> >
> >>> wsdl:portTypes representing the CIM modeled aspect of classes
> >>> (extrinsics and properties only). It is envisioned that these
> >
> > portTypes
> >
> >>> will be mixed in with additional capabilities from Grid disciplines
> >
> > to
> >
> >>> support Basic Execution Services, Data Virtualization and other
> >
> > areas as
> >
> >>> they are developed.
> >>>
> >>>Note: Abstract wsdl:portType refers to wsdl:portType definitions
> >
> > without
> >
> >>>bindings.
> >>>
> >>>OGSA-WG members please give me your feedback and edits ASAP.
> >>>
> >>>Tom
> >>>
> >>>
> >>>
> >>>Frey’s Law: “Every 5 years the number of architecture components
double
> >
> > and
> >
> >>>the ability to comprehend them halves”
> >>>
> >>>
> >>>Perfection is achieved, not when there is nothing more to add, but
when
> >>>there is nothing left to take away. – Antoine de Saint-Exupery
> >>>
> >>>
> >>>T o m M a g u i r e
> >>>
> >>>
> >>>STSM, On Demand Architecture
> >>>
> >>>
> >>>Poughkeepsie, NY 12601
> >>>
> >>>owner-ogsa-wg at ggf.org wrote on 07/29/2005 07:40:26 PM:
> >>>
> >>>
> >>>
> >>>>Hi all,
> >>>>
> >>>>Tom Roney, GGF forum-level liaison to DMTF, ask us to review
> >>>>updated GGF-DMTF Work Register document. OGSA-WG already has
> >>>>dialogue with DMTF folks and I think this document is consistent
> >>>>with our on going activity.
> >>>>
> >>>>Thus I would like to put this item into next Monday call agenda.
> >>>>
> >>>>Thanks,
> >>>>--
> >>>>Hiro Kishimoto
> >>>>
> >>>>----- Message from Tom Roney <troney at ncsa.uiuc.edu> on Tue, 26 Jul
> >>>>2005 09:23:32 -0500 -----
> >>>>
> >>>>To:
> >>>>
> >>>>mark.linesch at hp.com, hiro.kishimoto at jp.fujitsu.com, tmaguire at us.ibm.
> >>>>com, fred-m at crl.hitachi.co.jp, malaika at us.ibm.com, norm at cs.man.ac.
> >>>>uk, dave.pearson at oracle.com, dejan.milojicic at hp.com, tkojo at mvi.
> >>>>biglobe.ne.jp, eboyd at internet2.edu, rich at a3.ph.man.ac.uk, M.J.
> >>>>Leese at dl.ac.uk, david.wallom at bristol.ac.uk, jputley at earthlink.net,
> >>>>lfm at psc.edu, Shoshani at lbl.gov, pkunszt at mail.cern.ch, Jon.
> >>>>Crowcroft at cl.cam.ac.uk, travos at nortelnetworks.com, mulmo at pdc.kth.se,
> >>>>dane at fnal.gov, asm100 at doc.ic.ac.uk, darrenp at cadence.com, ali at epcc.
> >>>>ed.ac.uk, dder at ecs.soton.ac.uk, carole at cs.man.ac.uk
> >>>>
> >>>>cc:
> >>>>
> >>>>Tom <troney at ncsa.uiuc.edu>
> >>>>
> >>>>Subject:
> >>>>
> >>>>GGF-DMTF Work Register
> >>>>
> >>>>
> >>>>All,
> >>>>
> >>>>Attached is a draft of the GGF-DMTF Work Register, a document
> >>>>describing the collaboration between the two organizations. You
> >>>>are receiving a copy of the draft because your name appears in the
> >>>>document as having some role in the collaboration. Please review
> >>>>the document and let me hear back from you, either accepting the
> >>>>document in its entirety, or requesting additions, subtractions,
> >>>>or modifications to the document. The final draft will then be
> >>>>presented to the GGF Steering Group and the DMTF Board for their
> >>>>comments and approval.
> >>>>
> >>>>Your careful consideration is appreciated. If you are aware of
> >>>>a GGF working group not listed but having a role to play in this
> >>>>collaborative effort, please do make this known.
> >>>>
> >>>>Tom
> >>>>
> >>>> Note: It occurs to me now that I have forgotten the
> >>>> Semantic Grid Research Group. Listed as co-chairs
> >>>> will be David De Roure and Carole Goble.
> >>>>
> >>>>[attachment "GGF-DMTFWorkRegisterDraft.doc" deleted by Tom
> >>>>Maguire/Hawthorne/IBM]
> >>>
> >>>
> >>>
> >
> >
>
More information about the ogsa-wg
mailing list