Re: hierarchical shaping versus shaping in conjunction to cbwfq

From: Petr Lapukhov (petrsoft@gmail.com)
Date: Sat Jun 03 2006 - 14:29:29 ART


Pierre,

I may put the things a bit unclear, sorry :) The main thing I want
to point at, is that "bandwith" and "shape" do quite different things.

You are correct, one may use "bandwith" to allocate resourses in case
of "oversubscription". It's just useless if you set "shape" and "bandwidth"
to equal values :)

Say, if we have 512K physical port speed, and we need to divide it equally
between two classes of traffic in case of congestion.

To achieve that, we issue "bandwidth 256" under each class.

But next, we want let every class use more of available bandwith, in case
if it's actually available, but not too much :)

So we shape every class to 384k:

policy-map policy
class class1
  shape average 384000
  bandwidth 256
class class2
  shape average 384000
  bandwidth 256

HTH
Petr



This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:31 ART