Setting up QOS in for video conferencing.

Discussion in 'Cisco' started by rsjimmy, Mar 6, 2006.

  1. rsjimmy

    rsjimmy Guest

    I work in a large enterprise using video conferencing daily. I have
    read much about QOS, studied for it, but never really seen it in
    action. I just wanted to ask if there are techs here who have
    implemented it in their production networks and what sort of pitfalls
    to look out for with it. At this point I'm mostly interested in using
    it for our video and voice locations, but wanted to know if when doing
    so( I assume) it must be implemented all the way through the chain.
    ie... remote site - to - core site - to - other remote. Maybe see some
    stripped down interface configs to see it in action. Also some areas
    cross routers on slower wan circuits and other locations are routed but
    are connected gigabit; so I'm a bit interested as to how things are
    treated differently over layer 2 connections and layer 3 connections. I
    have read all about the ip precedence, the DIffServ, and QOS in
    general, but am just really curious to see how it has been done by
    someone in the real world out there. I hope I can get som egood input
    here.. Thanks
    rsjimmy, Mar 6, 2006
    1. Advertisements

  2. rsjimmy

    Charlie Root Guest

    The most important thing is to understand your traffic profiles (what kind
    of traffic, how much, from where, to where and how does it go), link
    utilisation, queueing capabilities of the devices (how many queues
    supported, scheduling of transit and self-originated traffic), how to treat
    traffic on L2/L3 devices (i.e. do routers need to care about 802.1p settings
    or not). Bear in mind that QoS configurations kick-in only when there is
    congestion (that is when there is no space left in TX-ring buffer, which is
    often much before your link is 90% loaded), all other time it's FIFO. Also,
    enabling QoS does increase load on the devices, and more so if configuration
    is suboptimal.

    While you don't actually have to enable QoS all the way through the chain,
    you do need to understand behaviour of the equipment when they don't have
    QoS configuration explicitly applied. Obviously the overloaded 2Mbps links
    are the first one where you should consider deployment of QoS policies,
    while barely loaded gigabit links will hardly show any difference with or
    without QoS applied.

    If you have more specific questions - feel free to ask.

    As for general reading you could start here,
    or get a book like "End-to-End QoS Network Design" (ISBN: 1-58705-176-1).

    Kind regards,
    Charlie Root, Mar 6, 2006
    1. Advertisements

  3. rsjimmy

    J.Cottingim Guest

    It's always good practice to implement QoS clear through the network
    path that the video/voice traffic would take. That would include Layer
    2 in your core network even if it's gigabit attached.
    That being said, one thing to remember that QoS only affects traffic on
    an interface when there's contention for the available bandwidth. (ie.
    Your link is)
    Here's some hints...
    * Bandwidth monitoring will not necessarily show the need for QoS as
    it's normally monitored on an average over 5 minutes.
    * If you're using ATM or Frame-Relay, do not oversubscribe the
    physical link. Cisco's QoS works on a per-interface (or sub-interface)
    basis. If you oversubscribe the link, all bets are off with regards to
    * If your using a managed MPLS WAN, you'll need to work with your

    There are probably a thousand different ways to implement QoS. Here's a
    L3 EXAMPLE config for you as requested: (CBWFQ)

    Ploycom<--->LAN<--->router-A <--->WAN<---> router-B<--->LAN<--->Polycom

    Both routers have similar configs so only one is listed here.
    class-map match-all AUDIO-CLASS
    match access-group name NEW-AUDIO-ACL
    class-map match-all VIDEO-CLASS
    match access-group name NEW-VIDEO-ACL
    policy-map VID-CONF-POLICY
    class AUDIO-CLASS
    priority 96
    class VIDEO-CLASS
    bandwidth 320
    class class-default
    interface ATM0/0.11 point-to-point
    description PVC to Remote Site
    ip address
    pvc 1/65
    vbr-rt 768 768
    oam-pvc manage
    service-policy output VID-CONF-POLICY
    interface FastEthernet0/0
    ip address
    ip access-list extended NEW-AUDIO-ACL
    permit ip host any dscp af41
    ip access-list extended NEW-VIDEO-ACL
    permit ip any host dscp af42

    Here's the break down:
    The ACL's and class-map's are used to classify the traffic.
    The Policy-map assigns priority and bandwidth according to needs.
    Then tie it all together on the interface the traffic would exit
    from. with the service policy.

    Remember that any QoS config is specific to the individual needs. -
    taking in account for bandwidth needs, WAN speeds, etc. The example
    above does not take in to account anything for routing protocol for

    J.Cottingim, Mar 6, 2006
  4. rsjimmy

    rsjimmy Guest

    I wanted to say thanks to all that replied. It has been a good help for
    me to take the next step.

    Thanks again
    rsjimmy, Mar 7, 2006
  5. rsjimmy


    Sep 25, 2019
    Likes Received:
    Dear respected team,
    we have site A and Site B connected with 30Mbps p2p link... Now we need to configure QOS for Video conference between 2 sites over P2P link. I am confuse about QOS configuration.
    site A site B site A vc , site B vc

    Please help me to understand how to configure QOS for VC. Thanks in Advanced
    bisdas19, Sep 25, 2019
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.