Re: [rtcweb] Endpoints that don't support RTCP

[Emil Ivov <[e-mail]>]

Hey Colin,

On 23.04.15 14:01, Colin Perkins wrote:
> Emil,
>
> On 23 Apr 2015, at 10:38, Emil Ivov <[e-mail]> wrote:
>> On Thu, Apr 23, 2015 at 10:46 AM, Colin Perkins <[e-mail]>
>> wrote:
>>> Emil,
>>>
>>> I have repeatedly presented evidence,
>>
>> You inadvertently forgot to paste the link to these presentations.
>> I am particularly interested in the "real-world networks" ones.
>
> They're in the IETF proceedings, for example AVTCORE at IETF 87 and
> 88:
> http://www.ietf.org/proceedings/87/slides/slides-87-avtcore-2.pdf
> http://www.ietf.org/proceedings/88/slides/slides-88-avtcore-4.pdf
>
> I had assumed you were at the meetings and saw the presentations, but
> perhaps not.

Indeed not. Thanks very much for forwarding!

Unfortunately, as far as Circuit Breakers are concerned I fail to find 
the "measurements on real-world networks" that you mentioned. Maybe I 
missed something?

The results that Varun and you presented are based on a simulation 
derived from a semi-synthetic data set. Zahed's is exclusively based on 
simulations.

All of this is obviously great work and very rewarding to read (some of 
it was new to me), so thanks for forwarding. Still, none of it however 
represents a practical evaluation of CB behaviour.

Ideally what we need in order to validate CBs is:

A) For a given real-world deployment, how much congestion do we see with 
and without CBs
B) For a given real-world deployment, how does user satisfaction vary 
with and without CBs

What I am worried about is that A) might not be particularly important 
because users, server-side apps and JS apps themselves would 
automatically resume calls.

As for B) I suspect that real-world scenarios might cause more cases 
where CBs trigger prematurely, which would have a negative impact on 
user satisfaction.

So with all that in mind I find that at this stage we are likely overly 
confident in CBs to MUST them in cases with no congestion control. Still 
given how people simply ignore the parts they don't like when 
implementing IETF RFCs I don't think we are in the face of a disaster or 
anything ...

Emil


-- 
https://jitsi.org