The Internet's Immune System
David A. Ulevitch
davidu at everydns.net
Wed Nov 12 18:15:06 UTC 2003
Christopher X. Candreva wrote:
>So in the above example, if I receive the report for 192.168.1.1 being an
>open proxy, I might have my system configured, because that is a residential
>DSL IP, to automaticly do a full port scan on it to look for open proxies,
>and if I confirm that it is open shut the line down, or just kick out a
>ticket for someone to call the customer. Or, start a netflow analysis on it
>to look for virus/worm traffic. Or not do anything until a certain number of
>reports are received, weighted based on the ranking of PGP sigs.
>
>
>
That's a start, but think about this. Worms are fast now. [1]
Lets say you have 30 seconds to stop a worm from the time it hits the
internet to until the time it's fully propagated to the point of serious
network disruption.
Automated techniques are the only thing that will stop it but is your
idea "fast enough?" I don't think so. Relying on user reports is good
for compromises and spambots but it won't do anything to stop CodeRed or
Nimda.
>Paul's use of the word immune system hit it on the head. An immune system
>kicks in automaticly to fight infection, and right now there isn't one on
>the net.
>
>
>
It has to automatically fight it, it has to be accurate and it has to be
fast.
I don't think anything comes close to that today.
-davidu
[1]: http://www.cs.berkeley.edu/~nweaver/cdc.web/
----------------------------------------------------
David A. Ulevitch - Founder, EveryDNS.Net
Washington University in St. Louis
http://david.ulevitch.com -- http://everydns.net
----------------------------------------------------
More information about the NANOG
mailing list