Chris.Keladis at ...2783...
Wed Oct 3 08:28:14 EDT 2001
Erek Adams wrote:
> On Wed, 3 Oct 2001, Roger Bou Aoun wrote:
> > Ca we stop spamming using snort??? If yes how can it be done, I know
> > that commercial Intrusion Detection Systems, are able to do it, can it
> > be done with the open Source software, or limit the number of sessions
> > that each IP can use on a certain port
Roger, how do the commercial IDSs determine a "SPAM" mail? (keyword,
> Some points in no real order:
> 1) How do you determine spam? You must look into the headers for some info.
> That's ALL you should do. If you go into the 'envlope' you are now 'filtering
> based on content'. That's a Bad Thing(tm) in the mailadmin world.
Well i dont think parsing the envelope headers would be as much of a sin
as parsing the letter headers. (After all, most every MTA needs to parse
the envelope headers to deliver the mail).
Even if you match on the envelope headers, SPAM could still get past
since it could have correct envelope headers (say from a forward or a
redirect), but be a spam internally in the letter headers, and i kind of
agree with you, parsing the content (letter headers) is rather lame,
especialy since letter headers are simply strings of the senders
> Just my .02 worth... I was a mailadmin in a previous life, so I'm still
> touchy about these kinds of isssues. :-)
Hehehe.. I hear you there :)
If this feature was seriously needed then i'd say you would need a
dedicated pre-processor, and even then you would have a hell of a time
parsing out the Received: lines since i don't think they need to conform
to any standard, apart from begin with Received: for each mail-hop.
I really think this is a job more suited to a host-based-ids, to plough
through the logs and raise alerts when the MTA (or front-end) sees SPAM.
Perhaps this is what Roger meant??
On the topic of HIDS - Marty, any plans, or is this a FAQ? :)
More information about the Snort-users