Re: question regarding queueing on Catalyst

From: Bob Sinclair (bob@bobsinclair.net)
Date: Tue Oct 28 2008 - 13:01:07 ARST


 Hi Reza,

Looks right to me! Does this differ from my previous message?

-Bob

Reza Toghraee wrote:

  Bob,
  
  Thanks for your reply.
  I got a little confused, can you please kindly correct me?
  
  1) default settings for a 3560's interface is :
   Shaped queue weights (absolute) : 25 0 0 0
   Shared queue weights : 25 25 25 25
  
  2) The shape weight values are "inverse absolute" (e.g shape 25 means 1/25
  of interface bandwidth which is 4Mbps for a 100Mbps interface)
  3) Queues 2,3,4 will share the remaining bandwidth of (100-4==96) 96Mbps.
  They share it equally (96/3=32) 32Mbps each queue.
  
  Thanks
  Reza Toghraee

  -----Original Message-----
  From: nobody@groupstudy.com [ mailto:nobody@groupstudy.com ] On Behalf Of Bob
  Sinclair
  Sent: Tuesday, October 28, 2008 6:24 PM
  To: Roger RPF
  Cc: 'GS CCIE-Lab'
  Subject: Re: question regarding queueing on Catalyst
  
  Hi Roger,
  
  In my experience you cannot induce congestion on this switch by changing
  the software reference bandwidth on the interface. Instead, try
  changing the port speed, or do an "srr-queue bandwidth limit" command.
  
  Remember that the configured values for the srr-queue bandwidth share
  command are not literally percentages, but relative weights. They
  appear as percentages only when they happen to add to 100. You might
  want to check the command reference on this one. Also keep in mind that
  shaped weights override shared weights. By default queue 1 has a shaped
  weight of 25, so the "40" weight in your example would not be used
  unless you zero-out the default shaped weight for queue 1. With the
  default shaping weights, queues 2, 3, and 4 would share remaining
  bandwidth in the ratio 10/60, 20/60 and 30/60, respectively.
  
  HTH,
  
  -Bob Sinclair CCIE 10427 CCSI 30427 www.netmasterclass.net
  
  Roger RPF wrote:

    Hi Group,
    
    I have a question regarding queueing on the switches, for example on a

  3560.

    If I have a FastEthernet port and I enable srr-queueing as below:
    
    mls qos
    Int Fa0/15
      srr-queue bandwidth share 40 10 20 30
    
    The configured values, reflect the percentage which the queue in case of
    congestion is allocated. Which bandwidth value does the switch take to
    calculate the real values? The Interface bandwidth, in our case 100Mb?

  If

    I configure for example bandwidth 10, are the calculations based on 10Mb?

  So

    the switch will start already start queueing (think that he is concested)
    when traffic reaches 10Mb?
    
    many thanks in advance
    
    Roger

    Blogs and organic groups at http://www.ccie.net
    _______________________________________________________________________
    Subscription information may be found at: http://www.groupstudy.com/list/CCIELab.html

  Blogs and organic groups at http://www.ccie.net
  _______________________________________________________________________
  Subscription information may be found at: http://www.groupstudy.com/list/CCIELab.html
  
  Blogs and organic groups at http://www.ccie.net
  _______________________________________________________________________
  Subscription information may be found at: http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net



This archive was generated by hypermail 2.1.4 : Sat Nov 01 2008 - 15:35:23 ARST