From: Anas Tarsha (ra3i@yahoo.com)
Date: Mon Dec 09 2002 - 06:49:02 GMT-3
Well in your scenario there is no real difference
between the two techniques. Both methods will
accomplish the same goal. But remember that CBWFQ was
designed to give you more flexibility of combining
multiple queuing techniques inside the same class. For
example if you would use the command "priority"
command instead of the "bandwidth" command, you would
be creating a priority queue which will be served
first. That's how Real Time traffic is treated in the
real world; it is always filtered to priority queue.
Another advantage of using CBWFQ is the option of
configuring WRED instead of tail dropping. Also with
CBWFQ you have the option of configuring WFQ for your
sub-class so packets within the same class will be
fairly treated. These options are not available with
CQ.
--- Nathan Chessin <nchessin@cisco.com> wrote:
> Hi All,
>
> A few questions regarding queuing theory.
>
> If we are to queue traffic based on letting a
> certain amount of bandwidth go
> through, this can be accomplished with customer
> queuing or cb-wfq right??
>
> If we use cb-wfq, it will look something like this:
>
> access-list 1 permit ip any
> class-map nate
> match access-group 1
>
> policy-map cisco
> class nate
> bandwidth percent 50 (also, is this 50% of
> max-reserved bandwidth -
> which defaults to 75%)
>
> And if we use custom-queuing, we have:
>
> access-list 1 permit ip any
> queue-list 1 protocol ip 1 list 1
> queue-list 1 default 2
>
> Because both of these default to 1500 byte count.
> Isn't this the same
> thing? Or am I missing the subtle differences?
>
> Thanks,
> Nate
> .
>
This archive was generated by hypermail 2.1.4 : Fri Jan 17 2003 - 17:21:41 GMT-3