On Fri, Dec 13, 2013 at 10:10 AM, <[email protected]> wrote: > What items do you believe are in the way of good group conferencing?
- Multiple streams of audio and video on the same PC. - Integrated AEC. More generally, I consider TURN-TLS and HTTP CONNECT to be more important and that is going to be the same people who would do ICE-TCP :) > Regarding contributing a patch, that is a fair request, though I'm not sure > that I have the resources either. Still, I will give it a shot if time > permits. That would be great. Feel free to reach out to me privately if you want to discuss. -Ekr > > > On Thu, Dec 12, 2013 at 3:54 PM, Eric Rescorla <[email protected]> wrote: >> >> I understand this point, but at present there are a number of other >> things standing >> in the way of doing good group conferencing, so we need to focus our >> resources >> on them first. >> >> However, If you would like to contribute a patch that did ICE-TCP, we >> would be willing to dedicate some resources to helping you and/or >> reviewing it... >> >> Best, >> -Ekr >> >> >> On Fri, Dec 13, 2013 at 7:09 AM, <[email protected]> wrote: >> > Thanks for your response. Here's my plea for ICE-TCP ;) >> > >> > I know that TURN/TCP has landed in Fx28 and that is one way to establish >> > a >> > TCP connection when UDP is blocked. However, UDP blockage is not the >> > only >> > reason to require TCP. Currently, I think that there is no way to force >> > the >> > client to use TURN/TCP when UDP is *not* blocked. >> > >> > The reason this matters to me is for a group conference with a large fan >> > out. For example, 1 person is speaking, and his audio/video stream is >> > going >> > out to 500 listeners. (a common scenario in my product). There is an >> > enormous advantage to be gained if you can make sure no packets go >> > missing >> > between the sender and the MCU. >> > >> > When it is possible to craft an offer containing a single ICE TCP >> > candidate, >> > you can be certain that the sender is connected by TCP. However, if you >> > have a UDP candidate only, along with a TURN server, the sender might >> > use >> > TURN/TCP, or he might not, depending on the connection check results. >> > >> > NACK exists for video packets, but audio is a different story, and some >> > people's connections have high loss. When communicating 1:1 low latency >> > is >> > more important than packet loss. When communicating 1:N, I think that >> > packet loss on the sender-server leg becomes very important for large N. >> > >> > >> > >> > >> > >> > On Thu, Dec 12, 2013 at 2:33 PM, Eric Rescorla <[email protected]> wrote: >> >> >> >> ICE-TCP is not currently scheduled. >> >> >> >> We are currently working on BUNDLE but don't have a schedule. >> >> I did know that. >> >> >> >> -Ekr >> >> >> >> >> >> On Fri, Dec 13, 2013 at 4:29 AM, <[email protected]> wrote: >> >> > Can you provide an update on when these features are expected to >> >> > land? >> >> > _______________________________________________ >> >> > dev-media mailing list >> >> > [email protected] >> >> > https://lists.mozilla.org/listinfo/dev-media >> > >> > > > _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
