May 2007 - nml-wg - lists.ogf.org

Buckets
by Freek Dijkstra 20 Jul '07

20 Jul '07

Hi, At the OGF meeting next Wednesday, I would like to have a discussion on "buckets" (or, since the meeting is in England, perhaps "baskets" would be a more proper wording). Buckets are meant to split the broad range of topics that NML is supposed to cover in smaller subtopics, were perhaps different subgroups can decide which properties should go in which schema. In the way I think about it, each bucket will be a separate schema or ontology. Ideally, there should only be a few dependencies between the different schemas/buckets. Below is a list of buckets I propose. This list is a bit broad. I hope that the NML-WG at the OGF in Manchester can decide which buckets are on-topic, and which are off-topic for the NML-WG, or if there is a better division in buckets. Note that because the charter specifically noted not to be technology specific, there are no technology specific buckets (e.g. there is no "Ethernet" bucket). Between brackets is an example list of the objects that could go in that bucket. That doesn't mean it should go in that bucket: what goes in is decided later, now is only the time to define the buckets. In fact, it would be great if the "bucket filling team" determines there is an existing bucket, erm, schema that can be re-used. - topology (single layer) e.g. (virtual) devices, (virtual) interfaces, links, network connections Re-use existing terminology? e.g. ITU-T G.805, Graph theory, GMPLS... - physical equipment e.g. physical devices, locations, geographic location - organizational boundaries e.g. domains, owner domain, administrative domain, organization/people - device/domain capabilities can switch (at which layer), interworking - device configuration (current configuration, available labels) - usage (path setup request, reservation ID, specific path) - technology characteristics Thus meta-data to describe a technology, not technologies itself. e.g. layer, technology, adaptation - circuit switched properties e.g. labels, channels, capacity - packet switched properties e.g. source & destination labels, buffers, packet size, bandwidth - Layer 1 specifics e.g. TDM, WDM, Fiber, UTP, wireless, Fiber bundles (ducts), OTN, ... - Layer 2 specifics e.g. Ethernet (MAC + VLAN), ATM, PPP, L2TP, Myrinet/Infiniband, ... - Layer 3+ specifics e.g. IP, TCP, ... Regards, Freek Dijkstra

2 1

Minutes of the OGF 20 meeting
by Jeroen van der Ham 16 May '07

16 May '07

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

1 1

Use cases
by Freek Dijkstra 11 May '07

11 May '07

In the meeting last Wednesday (surely, the designated volunteers will post their meeting notes shortly), there was a comment that we could use use cases. I recommend to have use cases in at least these four areas. Not because we want to cover them all per se, but at least it can make use aware what is required for different kind of applications: 1. Path finding 2. Visualization 3. Asset management (inventory of network devices) 4. Network measurements Pascale Primet offered to create a use case for #3 (Grid 5000 use case). I'm fine to create a short use case for #1 (if no-one else likes to do so). Regards, Freek

1 0

subscribe
by Jason Zurawski 09 May '07

09 May '07

1 0