[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ns] a question on Nortel DiffServ!
Try this and let me know if it works:
For the code point which carries EF traffic its better to use the token
bucket policer since u only want to know if traffic is In or Out of
profile. (The policer entry would require one downgrade CP). Add the PHB
entry say some thing like (10 -> in and 11 -> out)
$que addPHBEntry 10 0 0
$que addPHBEntry 11 0 1
and corresponding config Q could be
$que configQ 0 0 <appropriate min> <app max, could be same as min> <low>
$que configQ 0 1 0 0 1 <= this should drop nearly all packets w/ CP 11
For other code points add whatever policer, scheduling mechanism u want.
Well even I had tried the DROP mode which works but dun really emulate EF
requirement.
HIH.
Regards,
Sudhin.
On Thu, 12 Apr 2001, Sowmya Manjanatha wrote:
> Hi all,
>
> I have been trying to simulate a strict policing scheme (EF) out of the
> Nortel DiffServ module. I have a topology with 12 sources/dest, 6 edge
> routers and 2 cores. The topology is dumb-bell shaped with sources and
> destinations on either side of 3 edge routers connected internally to
> 2 cores something as below :
>
> s1 d1
> . \ / .
> . \ / .
> s4---- e1 e4---d4
> \ /
> s5 \ / d5
> . \ \ / /
> . \ \ / /
> s8---- e2---- c1 --- c2 ----e5---d8
> / \
> s9 / \ d9
> . \ / \ /
> . \ / \ /
> s12---- e3 e6 ---d12
>
>
> I have been using the PRI mode for the scheduler in the core. I have 3
> queues on the core and I set the QueueRate on each of the queues to a
> value of x, y an z. What I want to see is that the routers drop all
> the packets beyond the max. rate of x. I tried setting the drop
> probability on the queue using "$c1q configQ 0 0 10 20 1" to 1 with the
> intention of achieving the above. I also have a TS3CWM policer on the
> edges and I have the PIR and CIR values set to a really low value. But,
> the average rate I have been seeing is much above the rate of x. The
> linkrate between c1 and c2 is set such that the total rate coming out
> of e1, e2 and e3 is higher than the linkrate.
>
> I even tried setting MREDMode to DROP but in vain.
>
> Am I missing something here? Is there any way I can achieve the strict
> policing that I want (any specific policer, scheduler mode combination??).
>
> Any clues, suggestions??
>
> Thanks,
> Sowmya.
>
>
>
>
> *******************************************************************************
>
> Sowmya Manjanatha : University of New Hampshire
> Major : Computer Science
> ***********************************************
>
> email : [email protected]
>
> *******************************************************************************
>
>
****************** Sudhindra Suresh Bengeri ********************
Graduate Research Assistant | Home: |
to Dr. George N. Rouskas | 2110, Gorman St., |
EGRC 454H, Computer Sci. dept.| Raleigh, NC - 27606 |
NCSU, Raleigh, NC 27695 | Ph. (919) 852-1961 |
(919)515-3655(P) / 515-7925(F)
My web projection: http://www4.ncsu.edu/~ssbenger
* Always glad to share my ignorance - I've got plenty. *