Skip to Content.
Sympa Menu

wg-multicast - Re: QoS requirements of multicast applications

Subject: All things related to multicast

List archive

Re: QoS requirements of multicast applications


Chronological Thread 
  • From: Bill Nickless <>
  • To: David Richardson <>
  • Cc: Hugh LaMaster <>, Multicast WG Internet2 <>
  • Subject: Re: QoS requirements of multicast applications
  • Date: Tue, 05 Mar 2002 11:17:58 -0600

At 12:26 PM 3/4/2002 -0800, David Richardson wrote:
- Less than best effort QoS might be helpful. Unfortunately, there are
social challenges to deploying QoS mechanisms that attempt to provide
better than best effort -- who gets to have their packets preferred?
Less than best effort allows an application to choose how it should
be degraded, and avoids some of the social problems. An MPEG app would
mark P and B frames less than best effort, hoping that at least audio
and I frames make it through whole.

Fundamentally, the idea of QoS is to move congestion control out of the forwarding plane into the control plane (or above!). That's obviously the right thing to do for those folks focused on the forwarding plane, but not so obvious if you're worried about the control/management/political/religious plane.

There are some situations where an application can tell you what packets are less valuable than others. I agree with David: our experience is that people are willing to tolerate a certain level of video stream loss (lower framerate, etc), but cannot tolerate any loss in the audio. The loss of a single word in the middle of a sentence causes human-speech-time requests for retransmission. A preferred queuing strategy for audio packets can help avoid this situation in a constrained-bandwidth context.

Perhaps QoS efforts would find more success focusing on intra-application problems rather than inter-application or inter-user "fairness" or "prioritization"?

Maybe the question we should ask about QoS approaches is "what packets can we throw away?" rather than "what packets should be protected?" Audio/video conferencing applications can answer the "throw away" question rationally. Also consider non-TCP bulk-data transfer protocols; perhaps they should be allowed to blast UDP traffic without TCP back-off behavior if they're queued at a lower priority than TCP packets?
===

Bill Nickless http://www.mcs.anl.gov/people/nickless +1 630 252 7390
PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74 F7





Archive powered by MHonArc 2.6.16.

Top of Page