IDS mailing list archives

Re: Intrusion Detection Evaluation Datasets


From: Jamie Riden <jamie.riden () gmail com>
Date: Mon, 9 Mar 2009 20:09:58 +0000

2009/3/4 snort user <snort.user () gmail com>:
Greetings to everyone.

I have some questions regarding Intrusion detection evaluation datasets -

Apart from the Darpa datasets and KDD datasets, are there other
publicly available
datasets?
Are these datasets useful for evaluating a new IDS system or even a
new detection
technique?

Not the KDD '99 data set that I've played with - was categorised by
various things, but had no actual payloads if I remember correctly.

IMHO, the only way to evaluate an IDS is to plug it into your network
- no one else is going to share sensitive traffic of that kind, even
if they do it'll be different and the sheer volume of a continuous
100Mbs+ data feed is going to make such an exercise impractical. Also,
I have a degree in machine learning and I know how hard it is to
ensure that one data set (training data) is representative of your
actual problem (test data).

Find a friendly sysadmin and offer to trade: test your IDS in exchange
for supplying them with any useful information you might discover.

Sorry to be difficult :)

cheers,
 Jamie

PS: Not being anti-IDS. At my first security gig, I plugged snort into
my 100Mbs core switch - very enlightening, and I would not be without
an IDS sensor in any security role.
-- 
Jamie Riden / jamesr () europe com / jamie () honeynet org uk
http://www.ukhoneynet.org/members/jamie/



Current thread: