Notes on SIP Session 1 at IETF 68
reported by Bruce Lowekamp

note well slide displayed

GRUU
jdr

topic: do not muck with uri parameters
adam roach: there might be some parameters you need to muck with.
(loose route is an example motivation)
jdr: will make change to loosen requirement and review existing
uri-params

topic: ekr's stateless token draft
conclusion: a useful reference, but need some AD work to determine if
this needs to be normative or not and to determine where to send EKR's
draft

issue: notify immediately after reg-event has a race condition.
proposed solution is to resolve by stating when call-id changes, flush
all temp-gruus.  Discussion of change to require minimum time between
reg-event and a new notify.

poll: all in favor of this technique of mass expiring
vs all in favor of adding timing requirement:
approx a tie

query: is it critical to resolve now?
not many people

after clarifying that the poll involves backing out the change that
allows mass invalidation:
1) back out changes with no way to mass expire and require timer after updates

2) if you implement gruu-reg-event, you can invalidate with mass
expire

with options reversed, sparks now assumes the polling

1) like idea of bulk invalidate even if it means extra work

2) want to remove bulk invalidation and add timers

small number for both options, but significantly in favor of (1)
document change will remain as is


ICE-option-tag
jdr

critical changes to 3261 process
sparks

jdr and sparks: if you can do this as an extension, it doesn't go in
doc

philip: why not repeatedly do 3489bis
fluffy as ad: would cause a review of entire doc, including security section
and other changes

numerous speakers arguing that BNF is not critical, but clarifications
are probably not

fluffy as ad: will do as many things as extensions as possible.
Length will also be a factor; if this gets too large, we will have to
obsolete 3261.