Hi folks,
Here's a review of GGF13 from an NM-WG perspective. It's not necessary to read it all, but it's there if you want it :) As a quick summary...
* The proposed re-org of GGF was covered in detail during the plenary sessions, and while the ideas should improve the GGF "process" there will be little change on a day-to-day basis for NM-WG
* We held two sessions.
+ At the first session Mark reviewed the V1 schemas and gave demos of EGEE JRA4 and Internet2 software using the schemas. Many thanks to JRA4 and Warren Matthews for their help on these! Through the combination of the two demos we were able to demonstrate obtaining historic data from backbone and end-to-end monitoring infrastructures (i.e. multi-domain) and requesting on-demand tests. Richard also introduced the V2 work. Thanks to Dan, Martin and Jason for the support on this.
+ During the second session we had two interesting talks from more local speakers (Yuetsu Kodama from AIST in Japan and Minki Noh from KISTI in Korea). Richard then presented some thoughts on future work NM-WG.
* There was some interest in our work from the Asia attendees, which Richard will follow up
* We discussed how to finalise our V1 schema work with our current and new (as a result of the re-org) Area Directors. Documenting what we have as a GGF "Experimental document" seems the best way forward, as it will allow us to release a definitive version of the schemas with some supporting documentation WITHOUT it detracting from the push for V2. Richard has offered to be the editor, and will make a start by pulling together what we have already (some requirements, business logic and the schemas themselves). Mark will continue to finalise the schemas. The only input we hope to require from everyone else are some comments/observations in response to using the schemas (i.e. minimal effort required). Richard and I feel this is a win-win situation: we get something documented for the early adopters, it shows that NM-WG and GGF are producing things, and it doesn't take effort from V2.
* Also as a result of the discussion with the ADs, Geoffrey Fox will organise a Web Services tutorial for us (and any other interested groups) to be run at GGF14 (Chicago, 26-29th June). The exact format is yet to be decided.
As always, feel free to ask questions. The slides should be posted on the NM-WG website over the next few days.
As a reminder, we have an NM-WG call today (Tuesday 4th April). Susan has already sent out the call in details.
Cheers,
Mark.
NM-WG at GGF13 (Seoul, March 13-16th)
-------------------------------------
NM-WG sessions
--------------
We held two 1.5 hr sessions on the last day of the conference, with 28 attendees, which went well. Richard introduced the V2 work, while Mark presented V1 and gave two successful demonstrations, showing the schemas being used to request data from backbone and end-to-end monitoring infrastructures, and request on-demand tests. We also had two interesting talks from more local people - thanks again to Yuetsu Kodama and Minki Noh. All the presentations
See the attached file for the notes (thanks to Tom Sugden).
Other Activity
--------------
There was much talk during the plenary sessions of the proposed GGF re-organisation. The slides from theses sessions should appear on the GGF website soon. The changes aim to improve the GGF "process" but on a day-to-day basis the only real changes for NM-WG are that:
* We move into a new "infrastructure" group along with the IPv6, GHPN (Grid High Performance Networking) and NMA (Network Measurements for Applications) groups, and the dormant OGSI and DT (Data Transport) groups
* We lose John Tollefsrud and Geoffrey Fox as Area Directors, and gain Cees de Laat
We met with our existing (John Tollefsrud and Geoffrey Fox) and new (Cees de Laat) ADs to discuss documenting the V1 schema work. We said:
* Most people involved with the group want to move to V2. Some current users may take longer but we think they'll move eventually.
* We need to stabilise V1 as there are people using or waiting to use it (who cannot wait for V2) and we're not sure how long V2 will take, especially since we will unfortunately be losing some effort (Dan Gunter and Brian Tierney).
We had a long discussion, and the comments that summarise it best are:
GF: "You need to make V1 as viable as possible, for now, and in case V2 is delayed"
JT "You could use an experimental document to record what you have for posterity. This is described as 'Results of Grid related experiments, implementations, or other operational experience.'"
MJL: "Okay, we can try this based on the belief it will be a win-win situation: we get something documented for the early adopters, it shows the group and GGF are producing things, and it doesn't take effort away from the new V2 work."
JT "Experimental is best - shows you tried something; that it was put into use."
GF "You dont need to deal with security a great deal either. It's beyond your scope. You could say 'We are following security group <xyz>. We will adopt Web Services security as it matures'"
We also asked Geoffrey about help from his group with Web Services:
GT "It's not your job to do example implementations. There are lots of ways to do things with web services, and you can't cater for them all, e.g. the way Axis does things is different to other software, and I don't think the way it works is particulary good. I could organise some kind of tutorial for you for the Chicago GGF (26th-29th June). I'll ask around to see if it would benefit any other groups."
JT "Good idea. WS are certainly non-trivial"
GT "Yes, it would stop WGs wasting time on WS leaving them to concentrate on their real work. I thought in Brussels that you were spending too much time on Web Services instead of the schemas themselves"
Richard and I were able to do some (human) networking, which Richard will follow up and elaborate on in a future call. There was interest from:
* Yuexuan Wang, a researcher from Tsinghua University in Beijing
* Yuetsu Kodama and Tomohiro Kudoh from AIST (Japan's National Institute of Advanced Industrial Science and Technology) who during the second session introduced us to their gigabit network analyser. They were interested in using the schemas to publish data from the analyser.
* Minki Noh from KISTI (Korea Institute of Science and Technology Information) who also spoke at our second session, although he couldn't committ effort to active participation
* one other....hey, I can't remember everything ;-)
There were two interesting BOFs which people may want to find out more about: Grid VPNs and firewall issues. Richard also attended a session in the Scheduling and Resource Management (SRM) area where the group showed a useful looking approach involving producing abstract and implementational/operational versions of essentially the same schema. He will follow this up.
