[R-C] Follow up to my presentation at RMCAT

Matt Mathis mattmathis at google.com
Sun Aug 5 06:52:04 CEST 2012


I went back and reviewed my slides.  As Van pointed out, I slightly
oversold QoS.  The really important point was is that it does not work
for real time protocols to share queues with TCP (or other throughput
maximizing protocol).   QoS is just one way to manage segregated
queues.  There are several simpler alternatives such as Stochastic
Fair Queuing and fq_codel(), which can be deployed without any global
consensus on DSCP markings.

If you missed the presentation, the slides are here:
http://www.ietf.org/proceedings/84/slides/slides-84-rmcat-3.pdf

Thanks,
--MM--
The best way to predict the future is to create it.  - Alan Kay


More information about the Rtp-congestion mailing list