[R-C] A simple comment about draft-perkins-avtcore-rtp-circuit-breakers
Michael Welzl
michawe at ifi.uio.no
Wed Mar 28 01:26:54 CEST 2012
So this is supposed to be for severe situations only, i.e. we want to
avoid a false positive...
the draft says:
The congestion circuit breaker is therefore: when RTCP RR packets are
received, estimate the TCP throughput using the above equation and the
measured R, p (approximated by the loss fraction), and s. Compare this
with the actual sending rate. If the actual sending rate has been more
than an order of magnitude greater than the throughput equation
estimate for two or more RTCP reporting intervals, stop transmitting.
(where "the above equation" is the TCP throughput equation)
but in the presentation today, a statement about reacting when you
have reached 10 * that value was made. This means that the stream
would stop sending when it's as aggressive as 10 TCP flows... on a
host where there can easily be many more flows than that active in
parallel with the RTP stream, I wouldn't judge that as a "circuit
breaker" condition. My simple suggestion is to pick a much higher
value, say 100?
Cheers,
Michael
More information about the Rtp-congestion
mailing list