wb 2 task 8.1 QOS Legacy Frame Relay Traffic shaping

Hi all,

 

it's not clear how bc it is fixed to 2560 without any given tc.

I mean ,if the cir is 256000 how I can calculate 2 variables like tc and bc?

Is perhaps missing tc=10ms in the task or is how we have to set like default with Voip packets?

Comments

  • The last point does say to ensure that all the shaping intervals are the smallest possible, which is 10ms.

     

  • Thanks so much Edmund for your fast help,now I found the doc!

     

    http://www.cisco.com/en/US/tech/tk652/tk698/technologies_tech_note09186a00800d6788.shtml

     

     


    Frame Relay Traffic Shaping For Voice

    When configuring FRTS for voice, data performance may suffer at the
    expense of good voice quality. Here are some guidelines to enhance voice
    quality when configuring FRTS for voice:

    • Do not exceed the CIR of the PVC

      Most users have difficulty following this recommendation because the
      result is the router will no longer be able to burst to port speed.
      Because voice quality cannot tolerate much delay, any queueing of voice
      packets within the Frame Relay cloud must be minimized. When CIR is
      exceeded (PVC CIR, not the router configured CIR), depending on the
      provider and how congested the rest of the Frame Relay network is,
      packets may begin queue in the Frame Relay network. By the time the
      Frame Relay switch queues have backed up enough to trigger BECNs, the
      voice quality is already diminished. Because customers have many
      different Frame Relay providers and differing amounts of congestion
      across their sites, it is difficult to forecast what configuration
      works. Maintaining values at (or below) CIR on the PVCs that transport
      voice has proven to work consistently.

      Some providers sell a Frame Relay service of 0 CIR. Obviously, not
      exceeding CIR in this case would prevent any voice from being sent
      across the frame link. A service of 0 CIR may be used for voice but
      there needs to be a Service Level Agreement (SLA) with the provider to
      guarantee minimal delay and jitter for a certain bandwidth across the 0
      CIR PVC.

    • Do not use frame relay adaptive shaping

      If the configured CIR within the frame relay map class is the same as
      the true CIR of the PVC, there is no need to throttle down traffic due
      to BECNs. If CIR is not exceeded, BECNs are not generated.

    • Make Bc small so that Tc (shaping interval) is small (Tc = Bc/CIR)

      The minimum Tc value is 10 ms, which is ideal for voice. With a small
      Tc value, there is no risk of large packets using all the shaping
      credits. Large Tc values can lead to large gaps between packets sent
      because the traffic shaper waits an entire Tc period to build up
      additional credits to send the next frame. Making Bc = 1000 bits is
      usually a low enough value to force the router to use the minimum Tc of
      10ms. This setting should not affect data throughput.

    • Set Be = zero

      To ensure the CIR value is not exceeded, Be is set to zero so there is no excess burst within the first shaping interval.

    Note:  A good solution employed by some customers is to use
    separate PVCs for data and voice. This solution enables the customer to
    transmit up to port speeds in the data only PVC while maintaining a load
    at or below CIR on the voice PVC. Some frame providers may not find the
    solution appropriate depending on the frame switch and its queuing
    structure. If possible, have the Frame Relay provider prioritize the
    voice PVC over the data one so that there is not any queuing delay
    because of the data packets.

  • Hello All,

     

    My question is about the use of fragment size. From my understanding, the IOS calculates this for you automatically. If so, is this required as part of the answer. If I leave it out from the soultion, will I get the marks for it?

  • Just in case you need assistance with the calculation 

     

    How to obtain the Fragment Size:

     

    Serialization Delay : Fragment Size(bit) /link  bw (bps)

    So in this case we are given the Serialization Delay which is 10 ms and the link bw (256kbp)

    So the formula will be :

    Fragment Size = Serialization Delay * Link BW

    X= 0,01 * 256000

    X= 2560

    That's the fragment Size that we can send over that link in 10 ms.

    Now, on the Frame-Relay configuration we MUST set it in bytes so we must turn the 2560 bits into Bytes:

    2560/8 = 320.

     

    That's how you get the value

     

    Regards,

    Jcarvaja

    Follow me on http://laguiadelnetworking.com

Sign In or Register to comment.