From: Michael Snyder (msnyder@revolutioncomputer.com)
Date: Thu Feb 05 2004 - 10:25:01 GMT-3
After posting my message, I came to the conclusion not to follow the
configuration note posted on the cco page. Mainly because it would just
add another layer of complexity to what I already know.
But here's the rub, at what level am I not responsible of knowing
special cases? Am I supposed to know the some subjects to the nth
degree but idealize others because it messes up my numbers?
BTW, that config note is at the bottom of the example. So, it's ok to
read and apply the first half of the example, but not the second half?
http://www.cisco.com/warp/public/105/cbwfq_frpvs.html
I'm not expecting an answer, just pointing out how frustrating it is to
be a ccie candidate.
-----Original Message-----
From: K. Tahsin Hersan [mailto:tahsin.hersan@equant.com.tr]
Sent: Thursday, February 05, 2004 1:43 AM
To: ccielab@groupstudy.com
Subject: RE: CBWFQ & FR Question
I think the example provided in CCO was clear enough. I understand that
you
have lots of experience on this topic.
But my opinion is, you dont have to dig in to too much detail.
Thank you Saleem and Michael.
Cheers,
Tahsin
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Michael Snyder
Sent: Wednesday, February 04, 2004 7:11 PM
To: ccielab@groupstudy.com
Cc: 'Saleem Rahman (salrahma)'
Subject: RE: CBWFQ & FR Question
Yo, did you read the configuration notes?
Avoid setting the CIR or minCIR at the access rate. Otherwise, you may
see output queues building up and causing big delays in CBWFQ classes.
The reason is that the shape rate does not take into account the
overhead bytes of the flag and Cyclic Redundancy Check (CRC) fields, so
shaping at line rate is actually oversubscribing and will cause
interface congestion. There really is no reason to shape at the access
rate. You should always traffic shape at 95 percent of the access rate
or, more generally, the aggregate shaped rate should always be 95
percent below the access rate.
So if I'm taking the CCIE lab and get this CBWFQ on frame-relay, and
follow this config note. ???
About five times a year, I see issues that we really need a ccie proctor
to issue a ruling on. Cisco I wish you would setup a mechanism clear up
confusing and conflicting cco information. How about advanced
techniques like using TOS fields to set DSCP values? A yea or nay would
be very desirable for the acceptability in the lab of such advanced
techniques.
I'm not asking for trade secrets, but rather clear technical policy on
vague issues that could come up in the lab.
I'm assuming our messages are monitored by cisco. I believe the config
note above is a perfect example needing someone in the CCIE program to
chime in.
Michael
This archive was generated by hypermail 2.1.4 : Fri Mar 05 2004 - 07:13:46 GMT-3