InterOpMay2005 Astro-RG (IVOA/GGF interface)




Participants

NicholasWalton, GuyRixon, ReaganMoore, FrancoiseGenova, PeterQuinn, Pepi Fabbiano, RayPlante, RiccardoSmareglia, NoelWinstanley, KevinBenson, plus others.

Timetable

Time TopicSorted ascending Leader
15.00 Data Preservation ReaganMoore
16.00 Discussion  
14.20 GGF Overview ReaganMoore
14.40 GGF/IVOA overlap from the GWS group GuyRixon
14.00 Introduction NicholasWalton
15.20 Joint IVOA/GGF workshops: possible topics NicholasWalton
17.00 Wrap up and action points NicholasWalton

Materials from the work group and plenary sessions

Notes of the session

(prepared by KevinBenson)

1. Introduction

NW welcomed all (16 in attendance).

2. GGF - Global Grid Forum

RM gave a review of the GGF - noting the web site at http://www.ggf.org
  • Started as a standard access forum to grid.
  • Building upon globus tools being able to access applications.
  • Changed to a service arictecture built upon WSDL.
  • Switched to WSRF recently, those groups released recently a alpha/beta version for WSRF version.
  • Grid Forum needs feedback from IVOA for levels of robustness to run our services.
  • So GGF is going through changes (currently).
  • Interactions with commercial entity to produce grid standards, but still academic research initiative who started the grid forum ideas.
  • GGF 14 - June 27-30 2005 in Chicago, USA.
  • Working groups at http://www.ggf.org/ggf_areasgrps_overview.htm

Astro-RG We are listed under Community Areas/Groups (Astro-RG that is).

Other significant standards are:

Help by making standard by using the services. Commercials now competing, sort of like a faculty meeting.

Why did some groups get turned off? They failed or They finished their documents or some evolved to another working group and throw away the old one.

One difficulty about grid services no way of guarantee consistency.

3. Universal Worker Service + security notes

GTR discussed topical issues from the IVOA GWS WG and how these interface to te GGF.

Any idea you put in ggf could be viewed as competing and not going anywhere fast on a consistent standard. An idea is to do something parallel to it, add a few tid-bits and have ggf take over it if they like it.

IVOA approach to grid add a facade layer you get 2 kinds of services. Simple synchronous which are SIAP, SSAP, ConeSearch, VOStore. Then asynchronous job- managers - UWS, CEA.

CEA is a batch mgmt system. UWS is like CEA2, refactors some services of CEA.

Job-control interfaces CEC is astronomy free could be usefull for ggf to have.

UWS job control v2 - createJob(jobdescription), execute(), resumeJob()

Security: use ggf standards, or satisfy our own and feed back to ggf. Sounds like we might "satisfy our own", since ggf I think has got a little stuck with there security standard model. Might help ggf, need to get a concensus soon hopefully here in Kyoto in our security.

VOStore: similiar to 6 groups in ggf. Our concept of storing a file or a table with the same interface is quite nice maybe we can ask ggf for it. Cross Community to accept Votable might be a big struggle to get into ggf and also specefic to astronomy in areas.

PF noted that Harvard have a group working on 'Streaming over network via grids'. Margaret from Harvard may get in contact (I think contact with Guy Rixon).

Security, come up with a solution ourselves, but be prepared in 2 years that this might need some revision and get close to what ggf has.

3.1 Use of 'grid' services in VO projects

RM noted that a goal of this interest group was to map ivoa community to ggf services.

In this context he asked how many projects currently use or have tried prototypes with 'grid' s/w:

  • GSI: NVOx2, Astrogrid, JVO, China-VO (Grid Security Infrastructure
  • GridFTP: as above
  • WSRF standards: NVOx2

Use for encapsulating as a web service. plumbing work such as libraries - not going via GGF.

In UK core e-science will have the compute nodes to use the machines.

4. Preservation:

RM gave a presentation - see slides.

Authenticity - need an assertion to make sure the data is there .

Three things to do for preservation:

  1. Standard vocabulary.
  2. Standard endocing format.
  3. Standard services for manipulationg the encoding format.

Registries to work with METS profiles (METS-Metadata encoding and Transmission Standard).

Production Preservation environments are currently being used.

SRB is being used by a lot and now supports table.

In discussion RM noted that SRB was a storage solution, and it was important that it had an interface to VOStores.

Being used by many - single shared, stanford linear - BaBar (Stanford), Kek - High energy physics.

Minimize the number of messages sent over the WAN. So grab a section of files and then store. -srb: http://www.sdsc.edu/srb

5. IVOA/GGF workshops

Tag it on the GGF meetings? This is the first time for activity on the ivoa side.

Useful topics for workshop? - What services can ivoa put in the ggf

It was agreed that a June meeting at the GGF would probably be best, focusing on IVOA expereience with implementations.

6. Key points from the session

NW summarised the key action points:
  • issues under security not wait around for ggf. Go with our own.
  • asynchronous job execution
  • VOStore and SRB seems to be very similiar, may need to go over what is different besides the interface methods.
  • Standard break, see if something is generic or is it astronomy specefic.
  • Build up a index of which grid standards projects are used in there implementations. Contact project reps for this, but sounds like Nic and Reagan will do most of it.


Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r8 - 2005-05-31 - MarcoLeoni
 
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright © 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback