All of lore.kernel.org
 help / color / mirror / Atom feed
* [LARTC] Bandwidth throttling using "bounded" classe
@ 2001-05-09 14:22 Chee-Wei Tan
  0 siblings, 0 replies; only message in thread
From: Chee-Wei Tan @ 2001-05-09 14:22 UTC (permalink / raw)
  To: lartc

Dear all,

    I tested bandwidth throttling using TC CBQ using the simplest configuration with "bounded" on Pentium 3 , 648 Computers and it is puzzling to me that the bandwidth throttling mechanism does not work very nicely at very high  throughput.
 I attached only a class 10:1 to the root class and configured the "bounded" desired rate on this class. I chose to use the default fifo queue which is generated automatically when class 10:1 is created to reduce complication.
 I injected a constant flow of 80Mbps of UDP 1250 Bytes packets for 3 cases. 
    Case 1: The rate of 10:1 was "bounded" to 10Mbit, I received an outgoing
    flow of 11.6Mbps.
    Case 2: The rate of 10:1 was "bounded" to 20Mbit, I received an outgoing
    flow of 25.6Mbps.
    Case 3: The rate of 10:1 was "bounded" to 40Mbit, I received an outgoing
    flow of 63.6Mbps.

    Below is my configuration for case 3:

    $TC qdisc add dev $DEV root handle 10: cbq bandwidth 100Mbit avpkt 1000
    cell 8 mpu 64 ewma 8

    $TC class add dev $DEV parent 10:0 classid 10:1 cbq bandwidth 100Mbit rate
    40Mbit allot 1514 weight 1Mbit prio 8 maxburst 63 avpkt 1000 ewma 8 bounded


    $TC filter add dev $DEV parent 10:0 protocol ip prio 100 u32 match ip dst
    192.168.13.0/24 flowid 10:1

I must admit that the results were not very enticing. Why is it that additional 20Mbps of traffic can burst through the "bounded" class 10:1 in case 3?
 I changed my injected traffic from 80Mbps to 50Mbps in case 3 and I got an outgoing traffic of 50Mbps (it was not bounded at 40Mbps !!).
It can be seen that the bounding mechanism does not depend much on the incoming traffic. I tried changing the avpkt value from 1000 to 1250 but  nothing changes. Did I do anything wrong in my config or anyone can offer  me any explanation.Has it anything to do with the not-so-fined grain of  the estimator? I would appreciate a lot for any pointers.

    Thanks and kind regards,
    cheewei

______________________________________________
FREE Personalized Email at Mail.com
Sign up at http://www.mail.com/?sr=signup

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2001-05-09 14:22 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-05-09 14:22 [LARTC] Bandwidth throttling using "bounded" classe Chee-Wei Tan

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.