Re: [MMUSIC] Handling of unverified data and media

Bernard, you are right that this is possible with ORTC, even though I think
it's impossible with WebRTC.  But that's clearly out of scope for the
MMUSIC WG.  Perhaps the right forum to discuss Cullen's 1-800-fedex use
case is in the context of ORTC (or WebRTC NV).

On Thu, Mar 30, 2017 at 3:50 PM Martin Thomson <martin.thomson@gmail.com>
wrote:

> These both sound like legitimate cases where ICE can precede
> signaling.  I would recommend that the W3C think carefully about
> whether they might like to accept (potentially) invalid data before we
> spend a whole lot more time on the issue.
>
> On 30 March 2017 at 17:11, Bernard Aboba <bernard.aboba@gmail.com> wrote:
> > The unverified media scenarios seem to depend on ICE connectivity being
> > bi-directionally enabled so as to permit the DTLS negotiation to proceed
> in
> > advance of remote fingerprint arrival. If ICE candidates are signaled
> > separately from the DTLS fingerprint exchange it might be feasible, such
> as
> > in ORTC signaling where the ICE parameters are exchanged before the
> > DtlsParameters.
> >
> > At the last WebRTC interim a scenario involving PRANSWER and Trickle ICE
> was
> > presented. In the scenario, the PRANSWER included a fingerprint, but
> > possibly one which did not match the certificate provided in DTLS unlike
> the
> > final answer. I do not see how this could work but perhaps I am missing
> > something.
> >
> > On Mar 30, 2017, at 14:14, Peter Thatcher <pthatcher@google.com> wrote:
> >
> > We have a mailing list discussion (here), a bug
> > (https://github.com/w3c/webrtc-pc/issues/849) and a PR
> > (https://github.com/w3c/webrtc-pc/pull/1026#issuecomment-279238215)
> about
> > this.  I've copied the following comments to the latter two, so I'm
> adding
> > them here as well.
> >
> > TL;DR: I don't think unverified media is compatible with ICE+DTLS.  Here
> is
> > why (you can go see the bug, too):
> >
> > You can receive DTLS from the remote side before receiving the remote
> > description (and thus fingerprint). This happens if the remote side
> sends an
> > ICE connectivity check and the local side sends a response and then the
> > remote side sends a DTLS packet.
> >
> > You cannot send DTLS from the local side before receiving the remote
> > description (and thus fingerprint). This is because you can't send an ICE
> > connectivity check until you have the remote ICE ufrag and pwd, and thus
> > can't get an ICE connectivity check response, and thus can't send DTLS.
> This
> > is because you can't send anything other than ICE until you get an ICE
> > connectivity check response.
> >
> > Since you can't send DTLS, you can't complete the handshake, and thus
> can't
> > extract the SRTP key.
> >
> >
> > Maybe I'm missing something, but I think this is impossible.
> >
> > On Sat, Mar 25, 2017 at 1:12 PM Cullen Jennings <fluffy@iii.ca> wrote:
> >>
> >>
> >> On Mar 13, 2017, at 3:44 PM, Christer Holmberg
> >> <christer.holmberg@ericsson.com> wrote:
> >>
> >> My question is: is this something that’s causing problems in real
> >> deployments, and requires a change in the standard?
> >>
> >>
> >> 1-800 go fedex. See webrtc requirements documents from many years ago.
> >> _______________________________________________
> >> mmusic mailing list
> >> mmusic@ietf.org
> >> https://www.ietf.org/mailman/listinfo/mmusic
> >
> > _______________________________________________
> > mmusic mailing list
> > mmusic@ietf.org
> > https://www.ietf.org/mailman/listinfo/mmusic
> >
> >
> > _______________________________________________
> > mmusic mailing list
> > mmusic@ietf.org
> > https://www.ietf.org/mailman/listinfo/mmusic
> >
>