From: Chris Lewis \(chrlewis\) (chrlewis@cisco.com)
Date: Sat Aug 13 2005 - 06:35:21 GMT-3
Hi Rajib,
Gladston raised a similar question earlier this week. I have been
communicating with a few folks within Cisco in the QoS field and here is
the summary I have obtained so far.
For the purposes of the lab exam, you need to know that under an
interface, random-detet is mutually exclusive to fair-queue or in fact
any other kind of queueing or service-policy configured on the
interface. In addition, you need to know that when using MQC, you can
configure both a queuing strategy and RED within the same class.
Logically the MQC way makes more sense, as random detect is a congestion
avoidance scheme (alternatively it is a different drop strategy to tail
drop) and queueing strategies are congestion management schemes and
therefore they do not need to be mutually exclusive.
The question remains though as to why, at the interface level, RED is
considered a queueing strategy (as displayed by the show interface
output), when it is in fact a congestion management strategy. The only
answer I have so far is that RED support under an interface is there for
legacy support and customers should use MQC to manage QoS configuration
for production environments.
Chris
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Rajib Khan
Sent: Friday, August 12, 2005 5:57 AM
To: ccielab@groupstudy.com
Subject: WRED
Hi Guys,
When configuring random-detect under serial interface or under class
class-deafult does fair-queue must be enable
THnaks in advance
raj
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:01:19 GMT-3