Re: [rtcweb] Common ways of handling video conferences? (Re: Why requiring pre-announcement)

[Emil Ivov <[e-mail]>]

On 16.05.13, 23:28, Harald Alvestrand wrote:
> On 05/16/2013 09:22 PM, Emil Ivov wrote:
>>> Then it is another
>>>> question if it happens via SDP or via RTP header extensions or some
>>>> other means.
>>>>
>>>> There was a discussion at the Stockholm rtcweb interim on what
>>>> topologies that would be supported, but I fail to remember if this case
>>>> was included or excluded.
>> I couldn't find the discussion (in what WG did it happen?) but
>> topologies based on RTP translators of one sort or another seem to be
>> the default way of handling video conferencing these days so I don't see
>> how we could possibly rule them out.
>>
> 
> This is an interesting statement. I've also heard the claim that "RTP 
> translators hardly exist outside the lab" - I think that quip was 
> referring to RFC 5117 section 3.3 Topo-Trn-Translators, the story may be 
> different for Topo-Media-Translators - but I thought the most common 
> form was the section 3.4 Topo-Mixer?

Until recently, yes certainly, mixers were prevalent (which is arguably
part of the reason why very few people had access to video conferencing).

I think we'd all agree that this doesn't seem to be the case any longer
and that for various reasons (including cost, quality and flexibility),
most of the conferencing solutions today prefer shifting packets rather
than mixing content.

My statement above was made in that sense and not in the sense of
"Topo-Trn-Translators" beat all the other topos described in 5117.

> Do we have numbers or names attached to this?
> 
> I'm afraid all I can contribute real data on is Google+ hangouts, which 
> is a section 3.6 Topo-RTCP-terminating-MCU.

I can't know much about Google+ hangouts but I am quite convinced that
the problems I described earlier in this thread fully apply there. (They
certainly do for 3.6 Topo-RTCP-terminating-MCU) The reason you are not
experiencing them has much more to do with the fact that hangouts run in
a controlled environment where all clients and use cases are rather well
known. The right assumptions can therefore be made about use of repair
flows, simulcasting, presence or absence of cascaded translators (or
MCUs if you prefer) and so on.

Cheers,
Emil

-- 
https://jitsi.org