whatever happened to RED? (was: Improving Robustness...)
Didn't we have a solution to this problem? And the UDP flood from Real? And a host of other problems with saturated links? My operational issue is: what is the status of RED? Being small and only having cisco and *nix routers, I'd expect it was widely deployed. -- William Allen Simpson Key fingerprint = 17 40 5E 67 15 6F 31 26 DD 0D B9 9B 6A 15 2C 32
William Allen Simpson wrote:
My operational issue is: what is the status of RED? Being small and only having cisco and *nix routers, I'd expect it was widely deployed.
Have some additional deployment data here that some may find interesting: http://condor.depaul.edu/~jkristof/red/ Note the decrease in packet drops during this experiment when RED was enabled. If I ever get around to it, I have more data from later experiments that I'll add to that page. John
"was" == William Allen Simpson <wsimpson@greendragon.com> writes: Didn't we have a solution to this problem? And the UDP flood from Real? And a host of other problems with saturated links?
My operational issue is: what is the status of RED? Being small and only having cisco and *nix routers, I'd expect it was widely deployed.
RED alone doesn't help much against flooding attacks. You'd need something like "RED with penalty box" or (flow-based) WFQ, and those are either in research status or have other fundamental problems. Don't get me wrong, RED is extremely cool - buy only routers that support this at line rate, and configure it on any interface where you expect even a slight chance of congestion. -- Simon.
participants (3)
-
John Kristoff -
Simon Leinen -
William Allen Simpson