[Snort-users] VERY simple 'virtual' honeypot

Ryan Russell ryan at ...35...
Sat Mar 9 11:44:03 EST 2002


On Sat, 9 Mar 2002, Ofir Arkin wrote:

> You get to pull the attack of the wire only if they complete it...

True.

> If they will not get the right response no attack will be performed.

Only if a response is required.  You certainly won't get everything, but
you can get a lot by simply completing a TCP handshake.  Most of the
exploits and malicious code out there are not very careful.  Few coders
bother to have their code check for appropriate responses.  Code Blue is
one of the few worms I've seen that bothers to check what brand web server
it has found, for example.

> If the aim is to generate responses than you need to have a real
> intelligence engine to produce them in a way the engine itself will not
> get fingerprinted.

The more you can simulate, the more you will catch.  Some of the people I
work with developed a product called CyberCop Sting a while ago.  It does
things like emulate different OSes, applications, etc..  supposedly, it
would even emulate the IP stack well enough that the OS fingerprinting
would work.  (At least for nmap, we'd have to have you try ICMP, eh?)

Sadly, NAI seems to have dropped it.  Maybe they will free it some day.
Anwyay, that's the kind of tool I'd ultimately like to see for this
purpose.

> Also, it is more interesting, in my opinion, to simulate real world
> production environment style to Honeynets rather than a virtual one with
> less functionality.

Again, I think this speaks to the multiple purposes for honeypots.  For
some people, profiling baseline attack traffic may be more important, and
actually making a hacker busybox may be counterproductive.  Not that I
don't agree that it would probably be more interesting, it's just not
always the ultimate goal.

					Ryan





More information about the Snort-users mailing list