From: Dennis J. Hartmann (dennisjhartmann@hotmail.com)
Date: Wed Mar 02 2005 - 00:50:16 GMT-3
The smallest configurable Bc (normal burst) is 8,000 bytes. I've
never seen a smaller burst size in any platform. If you don't select a
burst size, one will be configured dynamically on your behalf. I've never
seen a smaller burst size, but I have seen a larger burst size (when I
modified an interface for Ethernet JUMBO frame support, the minimum
configurable burst size went up to 9000 bytes).
-Dennis
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Jongsoo.Kim@Intelsat.com
Sent: Tuesday, March 01, 2005 10:40 PM
To: dh8@pobox.com; marvin@ccbootcamp.com; ccielab@groupstudy.com
Subject: RE: CAT CAR question [bcc][faked-from][bayes]
As a satellite telecom engineer, 64Kbit and 64Kbps( bit per second) are
totally different to me.
If it is a typo by writer, then it would be a big typo in engineering
aspect.
1) Let's say it is not typo and time interval is 128ms(1/8 sec).
Then the correct burst size in byte = 64000 *(1/8 sec))*(1/8 byte/bits) =
1000 bytes So CAR would be "police 2000000 1000 exceed-action drop"
Dennis I would appreciate if you can confirm 125ms for CAT 3550.( Thanks in
advance)
2) If it is not typo, then so it was 64kbit.
Then I saw the answer from Marvin's email.
I hope protor knows the difference between bits and bps.
Thanks for your kind help.
Regards
Jongsoo
-----Original Message-----
From: Dennis J. Hartmann [mailto:dennisjhartmann@hotmail.com]
Sent: Tuesday, March 01, 2005 9:35 PM
To: 'marvin greenlee'; dh8@pobox.com; Kim, Jongsoo; ccielab@groupstudy.com
Subject: RE: CAT CAR question [bcc][faked-from][bayes]
I agree. There's the testing (lab) world and then there's the real
world. Sorry about the ms/us mistake. I stand corrected on that one. :-)
-Dennis
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
marvin greenlee
Sent: Tuesday, March 01, 2005 9:30 PM
To: 'dh8@pobox.com'; jongsoo.kim@intelsat.com; ccielab@groupstudy.com
Subject: RE: CAT CAR question [bcc][faked-from][bayes]
For the scope of a practice lab/scenario, if the values specified match the
syntax of the command, I would suspect that the writer of the practice
lab/scenario is really just looking for a basic answer.
'ms' is milliseconds.
125ms is 1/8th of a second.
125 microseconds would be 1/8000th of a second.
As far as I know, Cisco is not expecting CCIE candidates to have knowledge
of details hidden in internal documents that few people have access to.
How the router / switch actually interprets the command is another story.
Marvin Greenlee, CCIE#12237, CCSI# 30483 Network Learning Inc
marvin@ccbootcamp.com www.ccbootcamp.com (Cisco Training)
-----Original Message-----
From: Dennis J. Hartmann [mailto:dennisjhartmann@hotmail.com]
Sent: Tuesday, March 01, 2005 6:15 PM
To: marvin greenlee; jongsoo.kim@intelsat.com; ccielab@groupstudy.com
Subject: RE: CAT CAR question [bcc][faked-from][bayes]
I don't believe it's that cut and dry Marvin. The Bc is based on a
Tc. The timing interval (Tc) for bursting on all switch platforms is tied
to hardware.
The only reason I know this at this level of granularity is because
I teach the Cisco QoS class. It's VERY HARD to find this documentation. I
found it on an internal document.
Although I can't share the doc with you, the rate limiting interval
for the 2950 is 1/8000th of a second or 125ms. The interval is the same for
the 3550, and 4000 (Sup3 or better). The CAT 6500 supports a rate limitting
rate of 1/4000th of a second or 250ms. I'm going to go out on a limb here
and state that this is for the SUPI and SUPII, while the SUP720 supports a
rate-limitting rate of 125ms. I thought I read that in this document as
well, but I can't find it. Maybe it's in the SUP720 product sheet. Not
100% on the 6500 comments.
Sincerely,
Dennis Hartmann
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
marvin greenlee
Sent: Tuesday, March 01, 2005 8:53 PM
To: 'jongsoo.kim@intelsat.com'; ccielab@groupstudy.com
Subject: RE: CAT CAR question [bcc][faked-from][bayes]
On the 3550, you can use the police command to rate limit traffic. The
syntax is 'police rate-bps burst-byte [exceed-action {drop |
policed-dscp-transmit}] ' under policy map configuration.
So, to police to 2Mb with 64Kb you would configure 'police 2000000 8000
exceed-action drop'. (Note that the normal burst value is in bytes, where
the rate is in bits per second.
cat3550-1(config-pmap-c)#police ?
<8000-1000000000> Bits per second
aggregate Choose aggregate policer for current class
cat3550-1(config-pmap-c)#police 2000000 ?
<8000-2000000> Normal burst bytes
cat3550-1(config-pmap-c)#police 2000000 8000 ?
exceed-action action when rate is exceeded
<cr>
cat3550-1(config-pmap-c)#police 2000000 8000 exceed-action ?
drop drop packet
policed-dscp-transmit change dscp per policed-dscp map and send it
cat3550-1(config-pmap-c)#police 2000000 8000 exceed-action drop
3550 Command Reference - Police -
http://www.cisco.com/univercd/cc/td/doc/product/lan/c3550/12225se/3550cr/cli
1.htm#wp1864073
Marvin Greenlee, CCIE#12237, CCSI# 30483 Network Learning Inc
marvin@ccbootcamp.com www.ccbootcamp.com (Cisco Training)
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
jongsoo.kim@intelsat.com
Sent: Tuesday, March 01, 2005 5:37 PM
To: ccielab@groupstudy.com
Subject: CAT CAR question [bcc][faked-from][bayes]
Importance: Low
Question is
Rate limit a traffic on VLAN 20 upto 2M wiht 64Kbps of normal burst rate.
I can't figure out what will be normal burst byte size based on 64Kbps.
I hope someone help this explaining relation between normal burst rate and
byte size.
In case of Frame-relay, since the time interval = 1/8 sec, I can figure out
the size of Be if I know burst rate.
But in case of CAT, I am a little bit clueless.
Thanks
Jongsoo
This archive was generated by hypermail 2.1.4 : Sun Apr 03 2005 - 17:56:38 GMT-3