GGF NM WG Notes 04/05/05 Call began at 11:00 a.m. EDT. Attendees: Mark Leese Martin Swany Jason Zurawski Jim Ferguson Jeff Boote John Estabrook Dan Gunter Susan Evett (scribe) AGENDA GGF13 discussion Work with new schemas Mark gave a brief overview of the GGF13 sessions – he felt they went quite well; 28 attendees and participants expressed interest in the work. Richard is going to follow-up with specific people about getting them more involved in the work. Session 1: review of the schema work with a demo (JRA4 work from EEGE) that allowed us to show a single piece of software using version 1 schema getting data from several points on different networks. Second demo was the pipes data pulling data in from various networks. Followed this up with an overview of the second version of the schema. Richard tried to clarify the reason for the changes to schema 2. Session 2 was a couple of presenters from Asia – a Japanese presenter was talking about the different boxes they have in various locations in Asia and the U.S. They were concerned that the data was too complicated for the schema. Richard and Mark Second speaker was Minki Noh (KISTI) – he was very keen on using the schemas but was not very interested in participating in its development. Richard finished the session with thoughts on NMWG future work – vision for the next few years. They also talked about the new web site, explaining that we are relocating to a new site hosted by Internet2. Mark gave a brief overview of the changes in the organization of GGF – new area directors, etc. NMWG is moving to a new division (Infrastructure, which has 4 active groups and at least 2 that are dormant) – losing John ? and Jeffery Fox – but he doesn’t see this making a lot of day to day difference. Richard and Mark met with the existing area directors so we could move on (finish the schema v1.0 work to bed). Need to stabilize it now for the people who are early adopters. Approach suggested by Area Directors was to document the current schema as ‘an experimental document’; pull together the information wehave already, business logic, brief feedback from early adopters, and create a document that becomes a publishable version of the early schemas. Richard offered to be the editor on this so it should leave the folks working on version 2. John felt this was a good resolution to the problem identified in SLC re: how to get the v.1.0 work published and available while not ‘recommending’ it wholeheartedly. Jim asked (re: Minki Noh from KISTI) about the correct spelling of the presenter’s name; Jim is going to get in touch with people he knows at KISTI to try to get other KISTI folks involved. Jeff Fox offered to have someone from his group give a tutorial on web services at the GGF14 (Chicago) since the NMWG is planning a whole-day get together. Martin noted that they hoped to have the second version of the new schema ‘finished’ by then (that is, captures the easy, common tools used by the community and establishes an extensible base). Mark thought the Chicago F2F meeting would be: changes made to the schemas, spend much of the day going over the new schemas. Richard and Mark thought the web services tutorial, if it happened, should be on some other day because 1) not everyone wants to attend that tutorial and 2) the NMWG needs a full day, uninterrupted, to go over the new schemas. Dan felt that it would be useful to get several groups into that tutorial (not just NMWG) because they are struggling, as is the NMWG, with this process. Jeff asked if this would include the WSN-notification stuff – Mark said he thought so but, if folks have specific things they’d like to understand, please let Mark know (preferably with a ‘justification’ for why it is important). Jeff noted that the JRA1 folks are also interested in this – Jeff may convince several of them to come to the GGF meeting to attend this. Martin gave a quick review – he asked if anyone had taken a look at the implementation that is out there (no response). He feels the base schema is stable, after making these last changes. Extended by reducing the element name (expand the name space, if needed) to increase the speed of processing. Mark asked for an example; in XML schemas, there is no ‘proper’ inheritance. Jeff said he still was struggling with it and asked if there were any examples to present. Martin said, right now, no but examples will be created soon. They are creating code to provide the examples; Jeff asked if they could create a tar ball and send it out to the list. Martin said that they would definitely be willing to send it out (to the list? Or just to the 5 people who requested it?) but he wanted the receivers to be aware that this is under changes. Mark brought up the discussion of messages – in response to particular events; Martin noted that the original design was to streamline the request & response schemas. Have a single element – say, buffer-size – and the same element definition works for both the request and response schema. Ideal to have the same element in a request pushed out for notification. Dan: does the top level element say ‘request’ or ‘response’? Would that be a way to specify the language? Have we decided that yet? Martin: we’d like the XML that expresses the metadata to be flexible; publish a document describing its state (maybe not be able to use XML data) that can be pushed out (i.e. ‘I have some Iperf data from machine A to machine B’). Dan asked if this was a level above what the group is doing. Mark asked for everyone’s initial reactions – group felt it would be helpful to see this written down, alongside the code, put a short blurb alongside the changes. Mark wanted to know if the group should be waiting for the next version to come out before making comments or should folks be looking at it right now? Martin said he’d like comments coming now. Jeff asked for clarification on wild carding. (had to step away) Mark asked the group if they had any further topics – no response. He suggested that the next call be setup for 2 weeks after Martin and Dan sent out the new version of the schemas. ACTION ITEMS: 1. Martin & Dan to document the examples/code to describe the fast forward (ideologically). 2. Eric/Mark to setup a confcall 2 weeks after the new version of the schemas are released. Call ended at 12:06 EDT.