From: Michael E. Flannagan (mflannag@xxxxxxxxx)
Date: Thu Jan 25 2001 - 18:38:44 GMT-3
That explanations isn't correct anyhow. Defining byte count of 2.5Mb
isn't going to rate limit anything...that's just how much is drained from
the queue per cycle. As shown below, the router would drain 2.5Mb from
queue 1's bucket on each pass. If there's no congestion there's nothing
in the bucket -- hence, no effect. Even if there was "stuff" in the
bucket, unless there were three other queues defined as 2.5Mb each, you'd
not get the desired result. Furthermore, the "rule of thumb" is:
Packet Size * (2, 3 or 4 - depending on what you read) = byte count of
lowest priority queue. Increase byte count from there to allow higher
priority queues a greater amount of bandwidth (during congestion only, of
course)
------------------------------------------------------------
C i s c o S y s t e m s Michael E. Flannagan
| | Network Consulting Engineer
||| ||| Research Triangle Park, NC
||||||| ||||||| (919) 392-4550
.:|||||||||||:.:|||||||||||:. mflannag@cisco.com
------------------------------------------------------------
On Mon, 22 Jan 2001, Thierry MARTIN wrote:
> Hi,
>
> not right..
>
> default byte-count is 1500
>
> Queue 2 are byte-count 1500
>
> => WWW are 83,3 % of bandwith of LAN 10MB/s.
>
> Don't think that 2 500 000 is 1/4 of 10 000 000
>
> 2 500 000 for queue 1 and 1 500 for queue 2 = 100% of bandwith.
> It's not the same thing.
>
> Best Regard
> thierry
>
>
> >>> Lab Guy <lablist@yahoo.com> 22/01/01 15h58 >>>
> Would this limit www traffic to 2.5 meg on my ethernet
> interface?
>
> queue-list 1 protocol ip 1 list 100
> queue-list 1 queue 1 byte-count 2500000
> queue-list 1 default 2
>
> int e0
> custom-queue-list 1
>
> access-list 100 permit tcp any any eq www
>
>
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:27:44 GMT-3