[Nml-wg] Minutes of the OGF 20 meeting

[Jeroen van der Ham]

Network Markup Language Working Group
=====================================

Attendants:
Paola Grosso, cochair, Universiteit van Amsterdam
Martin Swany, cochair, University of Delaware, Internet2 Fellow, cochair
of NM-WG
Cees de Laat, Universiteit van Amsterdam, (retiring) area director of
infrastructure
David De Roure, University of Southampton, cochair of Semantic Grid
Jason Zurawski, Internet2
Anand Patil, Dante
Mark Leese, Daresbury Laboratory, UK e-Science
?, University of Pisa, Coregrid
Pascale Primet, INRIA
Ralph Niederberger, Research Center Juelich
Tomohiro Kudoh, GTRC, AIST
Richard Hughes-Jones, The University of Manchester
Freek Dijkstra, Universiteit van Amsterdam
Jeroen van der Ham, Universiteit van Amsterdam

Paola and Martin opened the meeting, setting the agenda and began a
round of introductions.
Most people present were working on provisioning tools and were aware of
the importance of topology description.

Jeroen gave an introduction to NDL, the Network Description Langauge, as
developed by the Universiteit van Amsterdam. He described the history
and gave a rough outline of the current status. More information can be
found at http://www.science.uva.nl/research/sne/ndl/
    Richard asked about the Services object, which is for pointing to
policy information
    Ralph started a discussion on the level of detail of a description.
This was postponed to later in the session, and due to time constraints
to the next session.

Martin presented the schema of the Network Measurements WG.
    Pascalle noted that there was no Location object in the schema.
Martin replied that the Node object could be recursive and in that way
fulfil the Locations role.
    Martin warned that layer 4 information also became a requirement for
describing their measurements (difference between TCP, UDP, etc.).
    The new schema (version 3) is moving towards technology specific
namespaces instead of layer specific namespaces.
    Publishing the new schema will now become a responsibility of the
NML-WG.

    Jason has made a script to translate single layer NDL descriptions
to NMWG descriptions.
    Easy translation between the two schemas should be a first step.
    It is agreed that both schemas are equally expressive, at least for
the topology part, which is our current concern.

Freek gives a presentation about Model vs Syntax and invites everyone to
review his document on XML vs RDF
(<https://forge.gridforum.org/sf/go/doc14257?nav=1>).
He proposes to use G.805 terminology for networks to avoid confusion.

Richard stresses that the abstraction level must be an explicit
assumption when starting the modelling.

Paola invites everyone to provide requirements for the schemas.
Pascale asks what kind of networks we want to model, i.e. do we also
want to be able to model wireless networks?
The group agrees that we do not wish to exclude wireless networks
(wireless sensor networks are seen as important application).
Martin responds that in the NM-WG it has proved effective to be as
extensible as possible and not to limit the schema. We invite everyone
to extend the schema where they see fit.

Cees presents his vision of RDF descriptions of grid resources, e.g.
databases, files, network, cpus, visualisation, etc. He invites David to
respond to this. David refers to Freeks presentation, saying that the
ontology should come first, RDF is a good technology, but not a
necessity for this vision.

Time flies and Richard reminds the chairs about setting action points:
- NDL and the (nameless) NMWG schema should further align. Both groups
will work on the translation between the two formats.
- The chairs will work on the scope document and present this to the
mailinglist
- All participants are requested to review the XML vs RDF document.
- Pascale asks about use cases and volunteers to write one for asset
management
- Cees proposes to create an example network to describe with both
schemas, proposals are the new Internet2 network, and/or the joint SC07
demonstration.
- Freek proposes to discuss his buckets proposal on the mailinglist
- The chairs will announce a phone conference on the mailinglist