Minutes IETF101: pim
minutes-101-pim-00

PIM 101 minutes

    Note well presented:
    Agenda bashing: 20 mins for first draft and 10 for each of the
others if you all can do it faster that would be great as we are
running behind.
    Status of documents: See slides
        Stig: I would like to ask for WGLC for drlb?  For? 1 hand.
Against? 1 hand
        Jake: I don't think its ready yet because the security section
is pretty thin and I don't think it will pass.
        Alvero: regarding security for yang there is a template for
that that the ADs put together so you might want to look at that.
        Justin Dean: Suggestion, in MANET WG for a document with
little feedback, we asked for an area review and got good feedback.
        Mike: Should we WGLC? 2 hands are for lass call
pim-ipv4-prefix-over-ipv6-nh

    Multicast from the MANET working group presentation
    Brian Adamson (presenter)
    Justin Dean: chair MANET WG, brought presentation here since
chartered to do this in our WG, may affect more than MANET. Looking
for people that are interested.
    Mike: which list to discuss. Justin, useful to do it on both lists.
    Brian: seconds doing it both places.
    Charlie: SMF was developed for the need of efficient flooding.
Tough to deal with mobile relay nodes. How effiicent if the member
ship is a small number of devices.
    Brian: elastic multicast helps. Right now an experimental
implementation. There is a draft given some overview. Should perhaps
write a more specific draft.
    Charlie, why is it called elastic
    Brian, it's a bit dynamic with the relay mesh exapnding and
contracting to support the group membership as needed ...
    Jake Holland: is the question to see how to do this in pim etc.
    Mike: idea is to get more participation in MANET
    Justin: MANET is currently chartered to do this work.

    IGMP & MLD Snooping model
    Zhang

    meng-pim-sm-enhancement Zhe Chen
    Rui Meng (Presenter)
    We need more review and comments
    Jake: is there any advanage of this over using a GRE tunnel)
    Rui: This would be done automaticly
    Do you have this unicast for ssm join or only for RP?
    Rui: (note taker didn't understand reply)
    Discussion about details of proposed solution...no clear answers
take discussion to the list

    Use of p2mp BFD in PIMpSM over shared-media segment
    Cerec Minsky (sp?) (Presenter)
    Andrew ?: ther eis a lot of state it may look nice on a slide but
if you have multiple routers and you have establish links too fast you
are going to get instablity.
    Presenter: So the BDR already has a state but it doesn't act on
it.  If we look at the secenrio that we only have 2 routers why use
p2p BDB Dr is not really interested in BDR so that would be overhead.
Another is that we can imagine we have more than 2 PSM nodes on a
segment and p2mp BFD works there.
    Andrew Docolw?: If we are going to try to make a fiat into a
farrai the fiat is going to fall apart.
    Presenter: this is a proposal it's open for discussion.  If it
seems subsecond discovery is too fast nothing precludes it from being
slower.
    Andrew: if we want it faster then we need to define the scope
because i've seen them fall apart in the real work at larger scales.
    Stig: There are several impimentations of BFD to pimpsm but they
are normally full mesh and this would be more scalable than those
solutions
    Propose to use optional TLV in HELLO packet to allow DR to
bootstrap their session; Pt-to-MultiPoint doesn't have 3-way handshake
so there needs to be out of band mechanism to advertise the
discriminator being assigned to session.  Should use the same source
IP as in PIM hello message.


   PIM Reserved Bits presentation
    Stig Venaas (presenter)

    Charlie Perkins: If you want more sub-types, make types 13-15 all
use the subtype field.
    Stig - no reason not to do that if acceptable to the group
    Please read the drafts and provide comments on the list.

    PIM Bier Extension
    Jingrong (presenter)

    ?? - Is it a BIER or PIM domain?
    Jinrong - it is a single BIER area

    XXX - Are you trying to say both PIM and BIER are signaled at same
time?  Is this PIM over BIER or two separate trees?

    Jingrong - it's a combination of BIER and PIM.  One in control
plane and one in the data plane

    XXX - So bascially it's PIM over BIER?

    YYY - instead of pulling apart the draft, please read the use
cases for BIER and then come to conclusion about your draft.

    Mike McBride - will be presented in BIER, too.  He is here to
propose extensions to PIM to build BIER forwarding

    PIM NULL Register Packing
    Mankamana (presenter)

    No comments or questions.

    Chair - please read draft and provide comments on the list


    Reliable PIM Registers
    Stig Venaas (presenter)

    No comments.
    Meeting adjourned.