OGSA Teleconference December 15, 2004 ===================================== * Participants Dave Berry (NeSC) Mike Behrens (R2AD, LLC) Andrew Grimshaw (UVa) Hiro Kishimoto (Fujitsu) Fred Maciel (Hitachi) Tom Maguire (IBM) Mark Morgan (UVa) Takuya Mori (ANL, NEC) Steven Newhouse (OMII) Andreas Savva (Fujitsu) David Snelling (Fujitsu) Latha Srinivasan (HP) Jem Treadwell (HP) Minutes: Andreas Savva * Summary of actions from this meeting - Action: Andrew to check costs with getting a hotel meeting room, wireless access and telephone line; but not proceed further until dates are finalized - Action: Hiro to send an email to Ian/Steve and confirm their availability Jan 26-28 - Actions for Dave: - Consensus on the call is to add a requirement for "widely adopted by industry" and "at least one is widely adopted by community" and that "this is a judgement call." [for profiles] - Should multiple implementations be required or would just a single one be sufficient? - Agreed that a lower bar is appropriate here so revise to "at least one implementation exists for a draft version" - Status is defined as {complete, evolving, draft} - Agreed to change these to "standard, stable and draft" and still need to define these precisely - Action: Dave to add an extensions bullet in the Profile Template - Action: Dave and Tom will initially write up "Profile definition" as an independent document. * No minutes available for approval * Next F2F Logistics - At the last F2F it was agreed that next F2F must be held well before GGF13 to allow time to follow up on actions and get the first draft of the documents out in time. - Candidate dates & places (1) Feb. 1 & 2 @ UK (London) (2) Feb. 21-23 @ US (Chicago or Charlottesville) - At the F2F there was a slight preference for option (2) - Ian and Steve cannot attend anywhere but Chicago. Their preferred choice is also (2) - Andrew has a minor conflict at the beginning of that week; it can be worked out. - Dave has a major conflict with option (2) since he will be in Japan (Naregi symposium) - Preference for London at beginning of February is to co-locate with the UK e-Science OGSA meeting; also allows Dave to participate (maybe) on Feb. 1. - Profile discussion is probably the main topic; key people to lead the basic profile discussion would be Ian, Steve, Tom, Dave, so should schedule for them. - Proposal to co-locate with GlobusWorld, Boston (Feb 7-11) and try to get key Globus people to attend part of the time. - Rejected because it does not look plausible - Proposal for the week of Feb 14, for example - 14-16 near Chicago (e.g., Airport Hotel) - 14-16 in Europe (Rome or London); convenient for people travelling to EGEE meeting that week - Rejected because of members' family committments. - Proposal for Feb 21-23 @Tokyo - Co-locate with Naregi Symposium - Rejected since it is unikely to get the Globus team - Proposal for week of Feb 1 @Raleigh - Co-locate with WSRF, WSN meetings - Rejected as Ian/Steve can't travel - Considered splitting the meeting up in two parts in both space and time (e.g., first part in London, co-locate with GlobusWorld for the last part) - Rejected as difficult to arrange - Tentative agreement for late in the last week of January (Wed-Fri, Jan 26-28) in the Chicago area - Andrew has a program committee meeting in Chicago that week; can make it (part of the time at least). - Dave will have to be in the US the week after so it is not incovenient; also part of the same PC - Works for the UK participants and is conveniently just before the UK e-Science OGSA meeting. - No conflicts raised by anyone on the call - Preferred location at the hotel where the PC will be held - Agreed to pay for our own coffee/food; unless some funding comes through. - Action: Andrew to check costs with getting a hotel meeting room, wireless access and telephone line; but not proceed further until dates are finalized - Action: Hiro to send an email to Ian/Steve and confirm their availability Jan 26-28 * F2F review - No minutes out; review based on presentations produced during the F2F ** Profile proposal http://tinyurl.com/5fxfd/Profile_proposal/en/1 - [p2]: It is clarification of existing process; nothing really new. - [p3-4]: Profiles are new - What about dependencies between profiles and feedback between them? - Not explicitly mentioned here (or discussed at the F2F) - 'Consistency' as an invariant; no definition of a process to achieve it; but it is a two way street between profiles and information documents. - Can profiles change over time and how is that handled? E.g., versioning of profiles? - We must make sure that profiles are "nailed down hard" and that they do not "shift over time"; must choose specs in which there is confidence that they will be adopted or have already been adopted. ** Profile definition (Profile Template) http://tinyurl.com/5fxfd/ProfileTemplate/en/1 - Definition of profile and candidate profile - Using existing GGF process as far as possible - [p1] Profile Definition: "Required for proposed recommendation" - What is 'stable'? - There must be a concrete (nailed down) specification - WSRF was mentioned as an example of a stable specification, but there was immediate disagreement; points to the need to define 'stable' precisely - [p1] Profile Definition: "Required for recommendation" - Need for interoperable implementations - Experience document is needed according to the GGF process. This would typically be from OGSA-WG but GGF allows individuals to submit these. - Would any two implementation do or should OGSA-WG have more stringent criteria (so above rule is "necessary but not sufficient")? - Someone mentioned (by email) to constrain implementations to 'commercial' only? - Difficulty with definining commercial; also it is fairly easy to start a company so this may not necessarily set a high bar. - Academic projects are likely early adopters; why discourage them? - Consensus on the call : No to 'only commercial' - Consensus that OGSA-WG has to 'approve' the implementations in the sense that it can evaluate the experience documents and decide to move forward only if the 'appropriate' ones are submitted; and that this is a judgement call. - Consensus to add a requirement for "widely adopted by industry" and "at least one is widely adopted by community" and that again "this is a judgement call." - [p2] Candidate profile - Clarify that a candidate profile is "as good" as a profile but probably would need more time to reach the same maturity level. - Clarified that candidate profiles are NOT intended as speculative documents. The point is not to produce unstable profiles. - Should multiple implementations be required or would just a single one be sufficient for candidate profiles? - Agreed that a lower bar is appropriate here so revise to "at least one implementation exists for a draft version" - Discussed requirements for 'commitment.' Again this is a judgement call. - [p3] Profile Template - Status is defined as {complete, evolving, draft} - These are not precisely defined. In particular need to define difference between "evolving, draft" - Agreed to change these to "standard, stable and draft" and still need to define these precisely - Only candidate profiles can use draft specifications - (Restrictions) May impose more stringent requirements on parts of the spcifications in the profile (for interoperability) - Interactions: What to do when combining specifications in order to make sure they are interoperable. - Definition is similar (but may have subtle differences) to WS-I definitions, in particular 'Reference' and 'Restriction' look the same. 'Interaction' may be different. WS-I has notion of 'roles' and they may come round to doing something similar. - WS-I also has 'extensions' to standards - Extensions are used very sparingly to fix specific problems with existing specs - e.g., they 'fixed' wsdl problem of no open content on portType. - Action: Dave to add an extensions bullet in the Profile Template - Action: Dave and Tom will initially write up "Profile definition" as an independent document. No decision yet whether this should be part of the OGSA roadmap, a separate OGSA document or a more general GGF document describing a process that other WGs may want to use. - Roadmap should refer to all other OGSA document and may end up being pointer to (bind) all the other OGSA documents ** Next steps presentation http://tinyurl.com/5fxfd/next_step_slides_at_2004dec_f2f/en/1 - p1: To do - Confirmation of schedule and next steps - Extent of first profile is still undecided - GGF14 deadline is for entering the recommendation track not for exiting it - p3: Basic Profile 1.0 - Adding Naming and Security to this profile is still unclear (for different reasons) - Basic profile meant to be used together with other profiles; it is intended to be fairly simple. - p4: EM profile - Concerns that it is prejudicial to include specific specs here; no hard decisions have been made yet. - EMS team should probably meet to discuss the contents of this profile. - This is the first proposal ; read it as an example. * Quick status review of OGSA v1 and Glossary - Revised glossary should be made available by next week - Andreas (and Andrew) have a couple of unresolved actions - OGSA v1 comment review; Andreas and Jem are holding separate teleconfs; expect comment review to finish this week and perhaps an almost complete document next week.