I''d like to implement something like SFQ, but with a cap of a rate per flow. Essentially, I''d like to be able to limit citrix connections to 30KBps without having an overall traffic flow cap for all citrix connections. Slightly preferable to this would be something that would actually work per flow instead of the stochastic algorithm. Is this possible? Thanks, Bill -- "It''s like being married to my best friend, and he lets me feel his boobs." -- Homer Simpson _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Le mer 09/06/2004 à 22:42, Bill Denney a écrit :> I''d like to implement something like SFQ, but with a cap of a rate per > flow. Essentially, I''d like to be able to limit citrix connections to > 30KBps without having an overall traffic flow cap for all citrix > connections. Slightly preferable to this would be something that would > actually work per flow instead of the stochastic algorithm. > > Is this possible? > > Thanks, > > BillThis could be a case for the HTB qdisc if you have a way to identify the citrix connections (ip, port, or any other field you know), here is an example of a config that limits 1 tcp connection of port source 22D0: tc qdisc add dev eth0 root handle 1: htb tc filter add dev eth0 parent 1:0 protocol all u32 match u32 0x0 0x0 at 0 flowid 1:1 tc class add dev eth0 parent 1:0 classid 1:1 htb rate 5000Kbit tc class add dev eth0 parent 1:1 classid 1:8 htb rate 240Kbit tc filter add dev eth0 parent 1:1 protocol all u32 match u32 0x00000800 0x0000FFFF at -4 match u32 0x00060000 0x00FF0000 at 8 match u32 0x22D00000 0xFFFF0000 at 20 flowid 1:8 The u32 filter permits to match any field of the frame, in the config above, 800 is for IP 06 is for TCP and 22D0 is the source port, this flow won''t go above 240Kbit. The problem is that may be you cannot know in advance the ports or ip numbers of your connections, in that case I don''t know what could help you. t t -- vincent-perrier <vincent-perrier@club-internet.fr> _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
On Wed, 9 Jun 2004, vincent-perrier wrote:> Le mer 09/06/2004 à 22:42, Bill Denney a écrit : > > I''d like to implement something like SFQ, but with a cap of a rate per > > flow. Essentially, I''d like to be able to limit citrix connections to > > 30KBps without having an overall traffic flow cap for all citrix > > connections. > > This could be a case for the HTB qdisc if you have a way to identify the > citrix connections (ip, port, or any other field you know), here is an > example of a config that limits 1 tcp connection of port source 22D0: > > tc qdisc add dev eth0 root handle 1: htb > tc filter add dev eth0 parent 1:0 protocol all u32 match u32 0x0 0x0 at > 0 flowid 1:1 > tc class add dev eth0 parent 1:0 classid 1:1 htb rate 5000Kbit > tc class add dev eth0 parent 1:1 classid 1:8 htb rate 240Kbit > tc filter add dev eth0 parent 1:1 protocol all u32 match u32 0x00000800 > 0x0000FFFF at -4 match u32 0x00060000 0x00FF0000 at 8 match u32 > 0x22D00000 0xFFFF0000 at 20 flowid 1:8 > > The u32 filter permits to match any field of the frame, in the config > above, 800 is for IP 06 is for TCP and 22D0 is the source port, this > flow won''t go above 240Kbit.I may be reading this incorrectly, but as I read it, this seems to limit all connections where 22D0 is the source port to the 240Kbit. (I hope that I am reading it incorrectly.)> The problem is that may be you cannot know in advance the ports or ip > numbers of your connections, in that case I don''t know what could help > you.Perhaps I can explain it a bit better: I have client machines (an unknown number of them) connecting to a certain ip address (i.e. 10.1.1.20) at a known port (1494). I''d like for each machine that connects to get 30Kbps. If there are 5 machines connecting, and 3 are idle (no bandwidth being used), I''d like the two active machines to still only have 30Kbps each. Thanks, Bill -- "The great tragedy of science -- the slaying of a beautiful hypothesis by an ugly fact." -- Thomas Huxley _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/