Re: [AVT] draft-ietf-avt-app-rtp-keepalive

[Xavier Marjou <[e-mail]>]

Extracted from http://tools.ietf.org/wg/avt/minutes?item=minutes77.html ,
here is the text I was referring to:


<http://tools.ietf.org/wg/avt/minutes?item=minutes77.html>

          Comment 2
          ---------
          The proposed resolution is to remove the recommendation for a
          fallback solution. Robert Sparks accepted this proposal.

          ...

          Keith Drage summed up: the proposal to remove the recommendation
          while retaining documentation of the alternative and associated
          drawbacks is accepted. The author should send a text proposal to the
          list and we would go on from there.



On Fri, Sep 24, 2010 at 3:16 PM, Cullen Jennings <[e-mail]> wrote:

>
> Can you point me to the minutes from the IETF 77 discussion where this
> consensus was reached ?
>
>
> On Sep 24, 2010, at 6:50 AM, Xavier Marjou wrote:
>
> > So I meant:
> > "The discussion during the next IETF (IETF-77) resulted in recommending
> only RFC5761 and "only" documenting other mechanisms."
> >
> > On Fri, Sep 24, 2010 at 2:49 PM, Xavier Marjou <
> [e-mail]> wrote:
> > I have just posted an updated version of the draft with a TCP timeout of
> 7200 seconds and no indication about DCCP timeout.
> > This is inline with you first two comment below, as well as with the
> feedback from the BEHAVE WG.
> >
> > For RTCP, the original idea was indeed to make recommendations when RTCP
> is not used. This raised a lot of concerns during the IESG review. The
> discussion during the newt IETF (IETF-77) resulted in recommending only
> RFC5761 and "only" recommending other mechanisms. This was reflected in
> version -08 and was not changed in today's -09 version.
> >
> > Cheers,
> > Xavier
> >
> >
> >
> > On Fri, Sep 17, 2010 at 10:07 PM, Cullen Jennings <[e-mail]>
> wrote:
> >
> > With regrades to the a few of the discuss on the 08 version of this draft
> >
> > For the TCP timeout, I recommend 7200 seconds. This matches the TCP keep
> alive used by default by many OS and will work well with the case I am aware
> of.
> >
> > For the DCCP timeout, I would suggest we have pretty much zero experience
> with firewalls and NATs with DCCP support so it would be premature to
> recommend anything here and we should make no recommendation for DCCP.
> >
> > For the topic or RTCP being out of scope. I would point out RTCP has
> existing ways of regularly sending packets covered by the RTCP spec and this
> specification makes no changes to theses.  I'm not disagreeing with Magnus,
> just saying that the RTCP is not a problem, it's the RTP we have problems
> with.
> >
> > I note that the WG had recommendation for what to do when the RTCP could
> not be used but the IESG has removed theses. This seems critical given the
> fact that a larger percentage of the deployed equipment does support RTCP. I
> think the specification needs to add back in what to do when there is not
> RTCP support or it needs to be brought back to the WG for discussion.
> >
> >
> > _______________________________________________
> > Audio/Video Transport Working Group
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/avt
> >
> >
>
>
> Cullen Jennings
> For corporate legal information go to:
> http://www.cisco.com/web/about/doing_business/legal/cri/index.html
>
>
>