traffic shaping w/ router as a transparent bridge

Nov 15, 2007 5 Replies

Ive got some issues with traffic shaping, I assume Im just missing something stupid but cant seem to find the issue. Im hoping some kind person can point out the obvious


Ive reviewed: Configuring Generic Traffic Shaping



formatting link
as well as ...
formatting link


Heres my network layout:



h1 (10.0.0.3 / 0017.0830.cb00) ---| | h2 (10.0.0.4) -------------------------|------ (f0/0) r1 (e1/0)



-------- (10.0.0.1) r2 ------- internet | h2 (10.0.0.5) -------------------------|


i need to shape in and out bound traffic for various hosts independantly (ie h1, h2, h3). i also need to do this with a bridge since i dont want to create any more confusion than needed by fooling with more subnets, consuming extra addresses, or nat. i have a spare router (r1) which is:



bw_throttle>sh ver Cisco Internetwork Operating System Software IOS (tm) 3600 Software (C3620-IS-M), Version 12.3(17), RELEASE SOFTWARE (fc2)



and has the following relevant config:



no ip routing ! interface FastEthernet0/0 no ip address traffic-shape group 1 500000 18000 18000 1000 traffic-shape group 700 500000 18000 18000 1000 bridge-group 1 ! interface Ethernet1/0 no ip address traffic-shape group 1 500000 18000 18000 1000 traffic-shape group 700 500000 18000 18000 1000 bridge-group 1 ! access-list 1 permit 10.0.0.3 access-list 700 permit 0017.0830.cb00 ! bridge 1 protocol ieee


this doesn't shape traffic to the expected 500kbps for h1; the only way i can seem to shape any traffic is with something like the following on both interfaces



traffic-shape rate 500000 18000 18000 1000



unfortunately this is not what i need; i instead need to shape traffic for specific hosts, preferably specifiying those hosts by ip address, not mac address, but at this point ill take whatever i can get


my network layout got a bit messed up, hopefully this is better Heres my network layout:

h1 (10.0.0.3 ) (0017.0830.cb00) --| | h2 (10.0.0.4) --------|-- (f0/0) r1 (e1/0) -- (10.0.0.1) r2 --- internet | h2 (10.0.0.5) --------|

*sigh*... sorry, still wrapping and goofing up heres one more diagram attempt

hh1 (10.0.0.3 ) (0017.0830.cb00) | | h2(10.0.0.4)------|- (f0/0) r1 (e1/0) -- (10.0.0.1) r2 - inet | h2 (10.0.0.5) -----|

perhaps i just cant do what im trying to pull off? since im trying to make a router act as a layer 2 device and shape traffic based on layer

3 addresses perhaps thats not a possibility?

Ive not gotten any help with doing this with cisco equipment... though ive moved on to doing it with BSD

Building a transparent traffic-shaping bridge

formatting link
There was a bit more to building the openbsd box, but with some help from the following I figured it out
formatting link
Things not in the guide were I had to add pftop... # pkg_add ftp://ftp.openbsd.org/pub/OpenBSD/4.2/packages/i386/pftop-0.6.tgz and to see queue info once you run pftop you have to use the left and right arrow keys

I also had to find info on how to change and reapply rules

Configure rules for pf: write your rules and save them in pf.test To test your rules type: # pfctl -nf /etc/pf.test When you are confident that you want to apply the rules type: # cp pf.conf pf.old && cp pf.test pf.conf To load your rules type: # pfctl -f /etc/pf.conf

Hope that helpls someone . If anyone figures out how to do it with something better or with cisco equipment id love to know how.

Depends what you mean by 'better'. If Linux is your thing, you can do it with ebtables.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required