QoS for video

Unanswered Question
Aug 12th, 2009
User Badges:

I am looking at implementing qos on our network for some ip surveilence cameras and also for streaming video from a quicktime server on mostly 2960 and 3560 switches. Can I use autoqos for this or am I better off creating MQC policies? If I use autoqos what would be the best configuration.


Thanks for your help.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
jjnagel102 Wed, 08/12/2009 - 18:34
User Badges:

Thanks. From what I had read AutoQos seemed more for voice traffic.

I recently had to design a global QoS solution for VideoConf installation on large dual-MPLS WAN network. Helps with DSCP tag recommendations and queuing.


HTH,

Andy


Here is a really great whitepaper on this exact topic: http://hive2.hive.packetizer.com/users/h323forum/papers/QoSforIPVideoconferencing.pdf

Details:

AF41 is therefore considered the most suitable for video traffic. As there is no advantage in treating voice packets better than the video packets in an IP videoconferencing application, AF41 should be used as

the DSCP value for both voice and video media in a videoconference. The EF marking is not a good choice for use with video conferencing for these reasons:


· Video packets are much larger than voice packets, usually as large as the maximum link


MTU size. If video packets are marked as EF then they will be allowed into the same


priority queue as voice. If a small VOIP packet enters the queue behind a large video packet


(or worse, a whole bunch of such packets), then the delay in the VOIP packet may


increase depending upon the traffic engineering in the network. Such delay will adversely


affect the performance of VOIP applications.


· Given that most EF queues are very small, using them for video traffic may lead to dropped


packets.


· Video coders tend to have much longer coding delays than voice coders, and hence giving


the audio stream(s) of a videoconference absolute priority only causes them to arrive early


and be held in order to achieve lip sync. Therefore, it does not help to put voice packets


associated with a videoconference in a queue with better service than that given to the video


packets.



Actions

This Discussion