From: Shane (cscoshane@sympatico.ca)
Date: Sat Feb 01 2003 - 13:19:28 GMT-3
CBWFQ could guarantee bandwidth and then police it to the same value
even when there isn't congestion.
Eg.
Policy-map blah
class iptraffic
bandwidth xxx
police xxx
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
cannonr
Sent: Saturday, February 01, 2003 10:49
To: Umair Hoodbhoy; ccielab@groupstudy.com
Subject: Re: How to approach QoS questions in the lab
Umair,
If they wanted you to use CBWFQ for you first question, they would have
worded the question more like this.....
"During times of congestion, IP traffic gets 60% of available bandwidth
on a link, DLSw gets 20%, etc."
For the second question, it is obvious that they wan't you to limit the
bandwidth that PING uses to 5% and drop anything above the limit -
probably to prevent DOS attacks. PQ, CQ, and CBWFQ would not accomplish
this.
Royce
----- Original Message -----
From: "Umair Hoodbhoy" <umair@cisco.com>
To: <ccielab@groupstudy.com>
Sent: Sunday, January 26, 2003 10:11 AM
Subject: How to approach QoS questions in the lab
> Hi all,
>
> I have a question about what the approach should be when tackling QoS
> questions in the lab. I was doing a few questions from a practice lab
> and they tend to ask for the Big Picture. But I don't think that they
> can be implemented in more than one way. Which brings me to my
> question: How can you identify what they are asking for?
>
> For example, in the practice lab, one question asked to guarantee that
> IP traffic gets 60% of the available bandwidth on a link, DLSw gets
> 20%, and all other traffic gets the remaining 20%. I thought it could
> be done by CBWFQ using the classmap-policymap-servicepolicy
> combination. However the solution configs used custom queuing. Could
> it have been done using CBWFQ and how would you recognize that?
>
> Another question asked to limit a single router (no mention of which
> interfaces) to allow up to 5% of its bandwidth for Ping traffic. Here
> the solutions used CAR and the rate-limit command in conjunction with
> an access-list. But what would the thought process be into coming to
> that conclusion?
>
> Thanks in advance for your help,
>
> -- Umair
> .
.
This archive was generated by hypermail 2.1.4 : Sat Mar 01 2003 - 11:06:00 GMT-3