OGSA Teleconference - 18 July 2005 ================================== * Participants Pete Ziu (Northrop Grumman) Jay Unger (IBM) Osamu Tatebe (AIST) Latha Srinivasan (HP) Dave Snelling (Fujitsu) Andreas Savva (Fujitsu) Manuel Pereira (IBM) Steven Newhouse (OMII) Takuya Mori (NEC) Mark Morgan (UVa) Tom Maguire (IBM) Hiro Kishimoto (Fujitsu) Andrew Grimshaw (UVa) Dave Berry (NeSC) Mike Behrens (R2AD, LLC) * Minutes approval Minutes approved with minor changes. Proposed changes described in http://www-unix.gridforum.org/mail_archive/ogsa-wg/2005/07/msg00056.html * August F2F updates - Andrew is leaving Thursday - Proposal to start Naming early and finish at 11am on Thursday; or move the Naming session to Wednesday - Agreed to move Naming session to Wednesday - Hiro and Mark have to leave before Friday noon - Agreed to start Friday early and if possible finish the session by 11am; continue without them, if necessary. - ByteIO moved to Tuesday - Proposal to allocate time during the F2F for resource modelling - Hiro to allocate three hours on Wednesday afternoon in the OGSA session. - WSDM conflict update: Heather Kreger has agreed to move the WSDM meeting to another week. - Hiro has contacted Allen L.: Prefers Data session later in the week. - Data moved to Thursday - Dave B wants to have a discussion on data staging as a separate job rather than as part of a single job. A number of JSDL-WG members will be at the F2F so having a discussion as part of the Data session is a possibility. * WSRF BP 1.0 review - Tom has uploaded v27 on gridforge Action: Tom continues to have the pen and will upload a new draft after the call. ** Canonizalization proposal - There are several (three) canonicalization proposals but many do not define a serialization. The proposed addition (below) does. - l.336-340 modified to use the OASIS specification "Schema Centric XML Canonicalization." - Also softened language (MUST to SHOULD) because it is not clear what the implementation status of this specification is. - This specification is a 'Committee Specification.' In other words, the specification has been approved by the TC as a CD, submitted to public comment (and only minor changes were required to address comments). But it has not gone through a final approval vote. In terms of the Profile Definition classification it is at least "Evolving Institutional." Action: Tom will check to see the implementation status. (Contact the IBM co-author of this proposal.) Action: Takuya will review this addition. He will also check the implementation status (within NEC). Agreed to delete comments against this proposal. ** Security Discussed security issues raised on the list. Action: Tom to arrange a separate teleconf to discuss the security issues below. Invite Marty, Dave S, Mark McKeown, Frank, Takuya. *** Signed endpoint references Issue raised on the list by Mark McKeown http://www-unix.gridforum.org/mail_archive/ogsa-wg/2005/07/msg00049.html - Presumably transport level security would not require this. - For Message level security, need to define what to sign; probably everything including reference properties. Action: Takuya will follow up with Frank on whether it should be mandated. *** Mandating transport or message security The profile as it stand does not allow non-encrypted messages or channels. There are cases when one would not want either, e.g., large data transfers that may cause performance degradation. Also depending on the environment it might be acceptable to not encrypt (e.g., operating within enterprise (behind firewall)). (If corruption is the issue then signatures and not encryption is appropriate.) Consensus on softening the requirement: - Change l.466 'requires' to 'recommends' - And also change transport level compliance statements R0811-14 from MUST to SHOULD. *** Mutual authentication - Discussed and agreed on last week. - The non-normative text is out of sync with the normative text. Agreed to change l.489 to 'recommended.' Action: Takuya to reply to Marty's email. *** User name/password authentication There is no statement about name/password authentication in the WSRF BP. It is allowed by the WS-I Security BP (token authentication). l.491 is problematical, however. The phrasing explicitly allows only x.509. But this is a non-normative statement and no normative statements back this up. Action: Takuya will look into this issue and report back to the group. * Resource modelling discussion [Timeboxed to 15 minutes.] The BES has put resource modelling out of scope. But it looks more and more that OGSA should face this issue sooner rather than later. And avoid or minimize rework later on There was a presentation about CIM/JSDL mapping at GGF14. This is a good starting point and OGSA-WG should discuss whether CIM is the solution for OGSA. The main questions are - what is a reasonable model for OGSA (for Grid); and - is CIM it Are there other candidates besides CIM? - It was noted that WSDM also seems to be turning towards CIM. - GLUE is an option but there is work to provide a CIM mapping. The rendering and semantics distinction/definition is discussed in GFD-I.45. Available at http://www.ggf.org/documents/GFD.45.pdf There is also the WS-CIM work. The group should check how far that work satisfies OGSA requirements. Initially a limited experiment might be in order. For example apply CIM to BES/JSDL. But BES-WG cannot take this work on, and RSS-WG also is moving to put it out of scope. One related issue is how to 'sell' CIM to the wider Grid community. It was proposed to put together a concrete proposal in a document format so as to have something concrete to discuss by the F2F. The RM call agenda (Tuesday) is set to discuss this topic. And continue this discussion on another OGSA call (possibly next Monday). * RNS specification review Draft: https://forge.gridforum.org/projects/gfs-wg/document/RNS-Proposed_Final_Draft-v1.7/en/7 Manuel gave an overview of the specification. The referenced draft is being prepared for submission. - Some grid file system parts (e.g. one of the appendices) will be removed for final submission. - Also to remove sentence in Sec.1.2.1 about manipulation via WSRF. It gives the wrong impression. Andrew had a number of comments and he will post them on the OGSA and GFS lists. There are two ways to proceed: - GFS submits the document to GGF Editor soon and there is followup discussion as part of the public comment period; or - GFS discusses with OGSA, adjusts the RNS specification if necessary and then submits. Action: Manuel will take this proposal back to the GFS group and will let OGSA know in a subsequent email. * Next call (Wednesday) - Next steps discussion - Teleconf schedule - WSRF BP 1.0 review