Hi Scott, Thanks a lot for the comments - please see inline.
On 10/30/13 5:40 PM, "Scott Whyte" <swh...@google.com> wrote: >On Tue, Oct 22, 2013 at 05:03:39PM +0000, Jan Medved (jmedved) wrote: >> We updated the Network Topology drafts - use cases >> (http://datatracker.ietf.org/doc/draft-amante-i2rs-topology-use-cases/) >> and the topology information model >> (http://datatracker.ietf.org/doc/draft-medved-i2rs-topology-im/). >> Please have a look - comments & feedback would be greatly >> appreciated. >> >> Also, as agreed at the last WG meeting in Berlin, we'd like to open >> a discussion whether network topology (as defined in the above use >> cases and the information model) is in the WG charter. The >> statement "The ability to extract information about topology from >> the network" in the WG charter says that topology is in charter, but >> does not exactly say how. > >Jan, > >WRT to the use-case draft: >If section 2 would explicitly define Topology that would help. Ok. >The draft abstract seems to define topology as "routing, forwarding and >policy information". Section 1 seems to treat Topology as graph >information only, separate from inventory and stastics. Section 3 >seems to flip that and go back to the abstract's definition. Good point. We'll align better Section 1 with the the rest of the draft. > >WRT the IM draft: >There seems to be a differentiation made between topology and topology >model. If I read it correctly, a topology model consists of graph >information (what I think of as topology), and other baggage. If I >got that right, this seems similar to my mental model of ISIS: the >topology consists of information describing how the nodes are >connected, and IP information is essentially baggage that makes it >possible to route IP packets. Correct. Also, the graph seems to be common for all kinds of topologies (L2, L3, service, VPN, etc), while the baggage is specific to a given typology type. > >Personally I think defining a topology model that uses different >graphs and corresponding baggage to provide useful abstractions is in >scope. I'd like it if you could tighten up the language though. Ok. > >-Scott Thanks, Jan >> > >> _______________________________________________ >> i2rs mailing list >> i2rs@ietf.org >> https://www.ietf.org/mailman/listinfo/i2rs > _______________________________________________ i2rs mailing list i2rs@ietf.org https://www.ietf.org/mailman/listinfo/i2rs
