Honeypots mailing list archives

Re: Honeypots and cluster computing


From: Valdis.Kletnieks () vt edu
Date: Tue, 02 Mar 2004 13:25:46 -0500

On Tue, 02 Mar 2004 09:50:37 CST, DeVaris P Brown <dpbrown () uiuc edu>  said:
Hi all. My name is DeVaris and I'm a student at the University of Illinois.
I've been doing some basic honeypot research and someone asked me yesterday if
honeypots could be applied to cluster computing without significantly slowing
down processing speed? Has anyone done any investigation into this topic? If so
I would love to collaborate on an upcoming project. Thank you in advance

Well... you could probably do it without slowing things down, because usually a
cluster has a high-speed internal network for its nodes, and then only 2 or 3
"front-end" nodes are attached to the outside world.  So (for instance) you can
park a port-80 http honeypot on all 500+ internal nodes, and never actually see
any packets.

A more important question is "Why do you suspect that a honeypot on a cluster
would *possibly* tell you anything interesting?"  I probably *could* install a
honeypot on the light pole on the sidewalk outside my office - but what would
that prove other than I've got waaay too much time on my hands?  On the other
hand, if I first theorize that this campus has a high incidence of wardrivers,
then a wireless beacon on that lamp post to capture them starts making more
sense....

Attachment: _bin
Description:


Current thread: