DAIS Telcon Minutes - 11/01/05 ============================== Chair: Dave Pearson Note taker: Mario Antonioletti Present: ------- Dave Pearson, Oracle Mario Antonioletti, EPCC Simon Laws, IBM Susan Malaika, IBM Thomas Soddemann, RZG Allen Luniewski, IBM Dave Berry, NeSC Amy Krause, EPCC Agenda: ------ Review of outstanding actions Review of revised Socialising document, questions and target audience DAIS presence at GGF13 VLDB Workshop Invitation Actions: ------- New Actions ----------- [Dave P] Send an email to the mailing list to see if anyone is interested in attending the OGSA f2f on behalf of DAIS. Outstanding ----------- [Norman & Simon] Identify scenarios for disconnected data sets. Post these on Grid Forge. Review in a future telcon. [All] Discuss Simon's posting to the DAIS list about naming. [Mario] To make a concrete proposal about a top level default data access operation and inform Norman. [Simon & Susan] Discuss outstanding WS-DAIR issues and post to the list. [Susan] To make a pass through WS-DAIR and note any outstanding issues. [Shannon] Interact with Amy to fix inconsistent naming of the ca-tory portTypes. [Shannon] To propose a way forward regarding support for XUpdate and post to the list. [Shannon] To post a proposal on optionality regarding XMLCollectionAccess::AddSchema to the list. [Simon] To add an issue regarding the fact that the WS-DAIX WSDL is not consistent with WS-DAI WSDL and discuss with Amy. [Norman] To arrange a mid-Jan telcon date to discuss outstanding issues on Grid Forge. [Simon] Wrap up the mappings document and submit it (pass on to Norman). To be done by the end of November. [Norman] See if a new telcon number can be sorted out for future DAIS calls. Completed Actions ----------------- [Dave P.] To draft and distribute a new version of the socialising document. Done. [Susan] To email to Dave contacts in XQJ and XQuery groups. Done. ---- o Review of outstanding actions These were processed. Dave: is there anything that we want to be able to do for the documents for GGF13? Mario: It would be useful if the xml spec issues were sorted out. Simon: It would be useful to review if we want to change the basic approach. Lots of things have been happening. Susan: there is also the relation with the OGSA profiles work and whether this affects DAIS. Mario: Dave Berry said he might call in. Dave: there is also the Andrew Grimshaw meeting. Susan: is this an official OGSA meeting? Dave reads out the email and sees no reason why it should not be an official OGSA meeting. Susan: dates coincide with the WSRF face-to-face meeting in the US. Mario: when is the submission date for GGF13? Susan: 18th of February. o Review of revised Socialising document, questions and target audience Dave: have had some feedback from Mario. He said that the introduction might be couched in too high a level ... the overview needs to be read by people who are in higher level management. The really technical stuff is done by more technical people... Allen: I think that is the right target audience. Susan: I do agree with Mario's comments. Mario: gives an overview of the high level comments: - overview too general and long - links to individual documents should be used as opposed to a pointer in Grid Forge - questions too verbose, should be shorter and snappier Dave: I agree with the links; trying to put in context round the questions so that we get back answers ... Simon: like simple questions but there is a timing question - we are still unsure how the DAIS specs fit in with everything else. It is unfair to ask people about things that we ourselves feel unsure about. Dave: at a functional level we have a clear understanding but how it fits in and what framework you you use then that is more difficult... Simon: the functions are relatively simple now but the model is difficult to explain without a specific mapping. Some of this is speculative. Dave: We are putting a list of questions which are difficult if you do not have a mapping. Susan: I kind of agree with that ... the answer might depend on the assumptions and these have not been firmly decided yet. Dave: my concern is that people would come back that there is insufficient information. We could get a negative response. We really want this to be a rubber stamping. Mario: However Malcolm keeps on asking us to submit DAIS so OGSA-DAI can implement it. Also, we are in danger of missing out on being incorporated into an OGSA profile -- we could use such a survey to leverage our way in. Simon: I agree with that ... Susan: need to choose a framework. Dave: I think that this is more important. ... the consensus then appears to be that the survey is only feasible once we have a framework. Then the questionnaire would be much more specific. Mario: what is actually stopping us from adopting a framework? Simon: the fact that there is more than one framework. Susan: the fact that there is more than one, that OGSA is going to be use profiles ... Mario: if OGSA produce a basic profile would that allow us to choose a framework? Susan: need to know whether we are going to follow the OGSA framework. ... Simon: maybe that stuff will be mapped in the basic profile but having an acceptable mapping for those standards is what has caused us so much difficulty .... Dave: but if WS-I and WSRF are going to be in the basic profile then we have the same problems that were outlined in the mappings document Simon: we need to keep the WS-I and the WSRF people happy. The profile would be interesting to see but it may not be sufficiently clear for what we want to represent. Dave Berry joins the call. Dave P: we should hold off on the survey on the socialising. Mario: Need to be able to leverage the fact that DAIS is an important spec to be included in an OGSA profile. Simon: I agree but we need to do more work. Dave: will hold this survey until at least the next meeting. Can now ask Dave about profiles. Dave B.: there was an OGSA f2f in Dec. For that meeting a doc was produced suggesting that OGSA should build profiles suggesting how standards are going to be used. There have been some presentations about this at various meetings. Currently there are 3 profiles being touted: the basic profile is getting the most attention - wsrf + naming (may use ws-renewable references) and ws basic notification. Currently have a draft that extends off ws-i 1.1. Also, need profiles for data and execution management services. In data I want to see DAIS included + some file stuff + GridFTP and some human name to abstract name mapping stuff. A completed profile is only meant for "stable" and "widely accepted specifications". There is a draft template that states what is required for a profile document to be accepted as a recommendation. Also, the constraints that should be placed on these specs should be more strict on what is required by the existing GGF process. There must at least be an implementation and a degree of stability in the specification for it to be accepted. Some want at least 2 inter-operable implementations and wide acceptance to carry a profile document through. DAIS has one implementation - possibly two if you count eldas. Pushing hard to have something like DAIS to be included. There is a chicken and egg process ... GGF should ratify its own work. Simon: need to be careful because OGSA-DAI does not implement the current DAIS spec. It is heading in that direction but it is not there yet. Dave B.: we would need to be able to convince that this is true .. the two are beginning to converge ... Mario: is there a date for the basic profile to come out? Dave B.: intend to submit it by GGF14. There is going to be a uk e-science meeting on the 31st January. Hope to arrange a follow-up data meeting as Andrew Grimshaw will be present. There is going to be an OGSA f2f meeting on Feb 14 which I will not be able to attend where they will discuss the basic profiles. There is awareness that DAIS has issues about WSRF. Dave Snelling was hoping that someone from DAIS/OGSA-DAI might be able to attend. At GGF13 OGSA are going to have 13 WG meetings. First session will be plenary then there are number of profile sessions: data, basic, naming, EMS, etc ... Might want to consider how DAIS engages with this. Should try to get people talking about the basic profile. Dave P.: who is presenting on the data profile? Dave B.: that would be in my space. If someone wishes to take on this notion happy to work with them. Dave P.: assume the data profile would say yes or no regarding the inclusion of DAIS. Dave B.: but you also need to say how it inter-operates with other specs/profiles. Could use GGF to kick this process together. Susan: are these profiles inter-operability? APIs? ... Dave B.: inter-operability like WS-I. Someone who conforms with a profile should be inter-operable ... Susan: a profile could incorporate all the DAIS bits ... Dave P.: will this require WSRF? Dave B.: yes, all services would have to satisfy the basic profile. Susan: and the basic profile has WSN and WSRF? Dave B.: yes. I can't remember all the details but I think it's most of WSRF and only basic notification. Susan: so for DAIS to be an OGSA profile it would have to have a WSRF compatible description? Dave B.: yes and use ws-addressing. Dave P.: does that mean that will be expected of all GGF standards? Dave B.: you could have GGF standards that are not OGSA compliant, at least as things stand at the moment. Susan: it sounds like DAIS wants to discuss where it wants to be part of a profile. Dave P.: that has to be discussed and on whose terms it wants to be a profile. Dave B.: important to engage with what is in a basic profile. The data profile is up to myself and Andrew Grimshaw. You can say no to DAIS being included. Currently trying to defend the space for DAIS to be included. Susan: in the UK, there will be input from OGSA-DAI. Sounds like quite an important meeting. Dave B.: this meeting is trying to decide what uk e-science wants to come out of the standardisation bodies. This is being led by Steve Newhouse with input from Tony Hey. Susan: after that meeting who else will you be getting input for data? Dave B.: trying to get some people to come along and discuss the draft data architecture. Susan: what do you hope for GGF13? Dave B.: would like DAIS to be part of a profile. Need to determine whether there will be one or more data profiles. Dave P.: what about infod? Dave B.: could be. Did not mean to exclude it. Dave P.: what about JDBC? Dave B.: would not expect so as that is Java specific. Susan: what do you expect of DAIS then? Dave B.: there was that posting that you sent about wsrf to the mailing list that is being taken seriously. I gather there is discussion in wsrf about use of ws-addressing. Maybe you could push for different OGSA profiles ... Susan: DAIS is trying to determine on what pieces of infrastructure it depends on or what it could depend on. Dave P.: also want to depend on things that we know are going to be widely adopted. Dave B.: part of the reason for having the profiles is to bolster adoption. You don't need to include everything in the profile but if you include bits in a profile there is no requirement that you use it but if you do use it you must use it as specified in the profile, e.g. UDDI in WS-I. So, things could work in a similar way. Will try to clarify. Dave P.: thanks for that overview. Dave B.: anyone interested in attending the OGSA f2f in Chicago? Susan: that is 3 days from the GGF meeting deadline. o DAIS presence at GGF13 Dave: Norman is not available for GGF13. Want to give an overview of what DAIS is ... the OGSA profile document is going to be very important. Who is available for GGF13? Susan will try to be. Simon does not have a plan to attend. Mario probably yes. Amy yes. Thomas probably not. Allen will be there. Dave B. will be there. Dave P. will probably be there. Dave P.: so we will have a presence. Norman was suggesting that we should have a couple of DAIS sessions. Dave B.:will try to ensure that the data meetings do not clash with DAIS. Dave P.: probably do not want more than 2 sessions ... ... Simon: should comitt to some slides if not the document for GGF13. DONM: 18/01/05. 4pm GMT. To discuss: OGSA profiles. What is going to happen at GGF13. VLDB workshop invitation.