Minutes IETF110: pim

Meeting Minutes Protocols for IP Multicast (pim) WG
Title Minutes IETF110: pim
State Active
Other versions plain text
Last updated 2021-03-19

Meeting Minutes

WG Overview - Chairs
Two yang models in RFC editors queue
Two drafts publication requested
pim-dr-improvement returned to the wg, needs more work.
null-register-packing also returned to the wg for more work.
pim-igmp-md-proxy-yang nothing new since October.
pim-assert-packing new revision discussed today.
pim-sr-p2mp-policy new revision discussed today
progressing igmpv3/mldv2 to internet standard. survey done. looking for
volunteers to work on revising IGMPv3/MLDv2 RFC's.

draft-ietf-pim-dr-improvement - Mankamana
Next step is more wg review before progressing.

draft-mankamana-pim-bdr - Mankamana
Lenny - this concept of priority and preemption is not unique to pim: vrrp,
rsvp with backup paths, etc. can we leverage from those? Was it protocol level
stuff or vendor implementations, those could be good examples. leave it up to
implementations? Alvaro - what has me confused is talking about two solutions
that are basically the same thing. A good argument has been made on how the
previous draft isn't needed. It would be nice if all the solutions was
considered in one draft. We seem to be circuling around implementations, first
resolve if we want single or multiple solutions. And then understand how they
interact. Stig - I agree. We initially only had one sticky DR in other draft,
now we have two proposals. Do we actually need two solutions? Are there
different use cases where one is better then the other? Alvaro - I'm not
advocating for one or two, the wg to decide. maybe we define multiple use
cases. Needs more coordination. Stig - if the wg decides we only need one
solution that covers all the use cases we probably only want to publish one of
them. Mike - some may want to have a hello option and others may not. And right
now we only have one wg document. Let's say we do adopt this draft, should we
hold off on progressing both documents until they are both progressed together?
Alvaro - That would be nice. they are not dependent on each other. they don't
have to progress together. progressing close would be nice. Stig - we shouldn't
progress any document until we carefully decide what solution is best or if we
want both solutions. Lets compare both options. Mike - let's poll for adoption.
Stig - just because we adopt both documents doesn't mean we publish both
documents. Poll - 9 in favor and 2 against. Will take to the list.

draft-vgovindan-pim-jp-extensions-lisp - Prasad Govindan
Stig - better to have a new attribute instead of overloading the existing one.
Acee - whats the relationship with the lisp wg? Where should this be
progressed? It's sort of like mvpn and bess owns that. This should be owned by
lisp. Prasad - yes should be presented in lisp next wg. Should be progressed in
lisp or pim? The type he is proposing is similar to the unicast type. Mike -
the lisp charter calls out multicast and pim so they may want to work on this,
Definitely need to coordinate with them. Alvaro - The difference between this
and mvpn is the details. This is a very short draft to request the new type.
Coordinate with lisp, will need to add more details. If they want it and we can
give them a type then its ok to progress it in pim if that's all that the draft
will say. Stig - exactly what happened last time, main work done in pim but
also discussed in lisp to make sure they thought it was a good idea. Sandy - is
this for *,g and/or s,g. Prasad - both. Mike - We will reach out to lisp
chairs. Alvaro - lisp didn't meet this week but will have an interim so please
present there.

draft-ietf-pim-sr-p2mp-policy - Hooman Bidgoli
Should we move hb-spring-sr-p2mp-policy-yang to pim?
Mike - We discussed this with spring chairs and they want to keep it in spring
unless it was broke up into two documents one for replication segment and one
for p2mp policy. One done here and one done there. Hooman - will discuss with
authors. Mike - only addition is SRv6 in the appendix, will it stay there?
Hooman - likely yes.

draft-ietf-pim-assert-packing - Yisong Liu
Stig - This looks good. If you have assert timers that expire at the same time,
don't want to send asserts too late, earlier is ok. If you have many asserts,
you could send all of them now and they would then be in sync. Stig - should we
consider wglc? Yisong - not ready for last call. Stig - read through it again
and it's in good shape.

draft-ietf-pim-null-register-packing - Ananya Gopal
Stig - there were comments about the downgrade scenario. We could do that
during wglc. Mike - when the new rev is submitted it sounds like we need to
issue another wglc? Stig - probably should at least have those who have
reviewed it previously to review. Mike - we will ping reviewers and then issue
wglc after next rev.

draft-chen-pim-srv6-p2mp-path - Huaimo Chen
Stig - what the relationship should be with spring.
Huaimo - sent a request to spring and waiting for comments.
Mike - spring is so busy and higher priorities and mcast is not a high priority
which is why we are working on this topic in pim. Stig - would be good to hear
whether this update addressed concerns. People were in favor and some against.
Would be good to see if people have changed their minds. Mike - (individual).
we already have a good solution in pim do we need another solution? My view we
should be open to multiple solutions if they are technically sound. we do need
to get buy in from spring regardless. Stig - Either way its important to get
comments and keep improving document. Will ask on the mailing list for input
and reach out to spring again.