From: Leigh Harrison (ccileigh@gmail.com)
Date: Sun Sep 18 2005 - 12:25:03 GMT-3
Simon,
Sorry to have to do this, but I totally disagree with you!
I was under the impression that MQC is a centrallised way of writing QoS.
The "police" statement is mqc's version of rate-limiting - that's just
it's name for it.
In the police statements, the Bc is the "chunks" of the CIR. i.e, you
can send x bits per second (cir), but only in chunks of y (Bc), so that
you can set it to send a few big packets, or lots of little packets (for
voice, etc.)
(To me) the statement reads, send 128000kbs, in chunks of no more than
2000 Bytes and you can burst an extra 2000 Bytes on top of that too. (I
guess that the access-rate is 256k ?)
In the mqc police sub section, the "conform-action" would be anything
between 0kbs and 128kbs (within the normal traffic parameters) and the
"exceed-action" is the 128kbs - 256kbs (or whatever the burst size (Be)
is set to). The "violate-action" is anything on top of that. But on a
policeing statement, I think you'd do well to exceed the burst.
The numbers for the rate-limit and police are the same - CIR, Bc, Be.
The only one that you need to have in there is the CIR, the rest "tweak"
the value.
As for the section matching the access-list (101) - that would be
defined by a class-map, shoved into the policy-map and policed.
Apologies if I'm teaching you how to suck eggs!!
LH
--simon hart wrote:
>Hi All, > >I would welcome comments on this task within the IE workbook. I think that >the question and answer may be incorrect, > >The task itself is to change various CAR statements into MQC. The problem I >see here is that CAR and MQC have different policing algorithms and thus to >mimic each statement is extremely diffilcult. > >One is asked to achieve the following > >rate-limit input access-group 101 128000 2000 2000 conform-action transmit >exceed-action set-prec-transmit 0 > >The answer given for the MQC statement is as follows > >police cir 128000 bc 2000 be 2000 >conform-action transmit >exceed-action set-prec-transmit 0 > >Now I believe this answer to be wrong for the following reasons. When the >bc and be are equal within CAR there is no burst, thus no exceed action. >Although the CLI for CAR will accept the 'exceed-action set-prec-transmit 0' >there is in fact no burst and thus no exceed action. >However when entering the command within MQC we have a problem if we mimic >the CAR statements. Because MQC uses a single rate three colour marker it >has two discrete token buckets. By setting bc and be we are setting the >values for each token bucket and hence enabling an exceed action. Therefore >we are allowing a burst action that will set the precedence of 0, however >the CAR statement does not allow that. > >I believe the correct answer would be > >police cir 128000 bc 2000 >conform-action transmit >exceed-action drop > > >One is also asked to perform the following > >rate-limit input access-group 102 256000 4000 8000 conform-action transmit >exceed-action set-prec-transmit 0 > >The answer given for this in MQC is: > >police cir 256000 bc 4000 be 8000 >conform-action transmit >exceed-action set-prec-transmit 0 > >For similar reasons that I gave above I believe that this answer is also >incorrect. I believe that the correct answer should be > >police cir 256000 bc 4000 be 4000 >conform-action transmit >exceed-action set-prec-transmit 0 >violate-action drop > >The CAR statement implies a burst of 4000 (the difference between 4000 and >8000), therefore within MQC bc and be should be 4000 a piece. In addition >one has to create a violate-action otherwise the circuit will never be >limited to an average of 256000 as the exceed action will be used on all out >of contract traffic. > >Other views most welcome > >Simon Hart > > > > >-- >No virus found in this outgoing message. >Checked by AVG Anti-Virus. >Version: 7.0.344 / Virus Database: 267.11.1/104 - Release Date: 16/09/2005 > >_______________________________________________________________________ >Subscription information may be found at: >http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sun Oct 02 2005 - 14:40:15 GMT-3