I thought the idea was to spread the "tarpits" around, so 1. the odds of the spammers getting a direct connection to a tarpitted server would increase 2. the sysadmins down the stream will notice that their servers are used for spam and (hopefully) do something about it.
Tarpits might not be perfect, but it's the best idea so far. When you fight anything, the rule is to go for the money source. Hopefully tarpits would change the economics of spam, as Marty Lamb envisioned in the original paper. I am working on a PhD in spam identification using AI, and as it is a streaming filter, it would fit nicely with the idea of tarpits. Teo ----- Original Message ----- From: "Andrew Rose" <[EMAIL PROTECTED]> To: "'Samuel Liddicott'" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Wednesday, March 05, 2003 9:23 AM Subject: [good_email] RE: tarproxy fails to reduce spam > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Sam, > > > > What a tarpit does do is use up another resource - > > > elapsed time. However, this is not a critical > > > resource for a spammer. A spammer can easily open > > > several simultaneous connections to the same SMTP > > > server. > > > > I don't see why the tarpit SMTP proxy shuld be unable > > to limit all connections from the same IP and even > > give "500 get lost for the time being" responses to > > further connections from those IP so that it will not > > accept too many connections from the same IP - or at > > list if it looks like spam. > > I covered this objection in my original email. Given that spammers > send out thousands of mails to different people on different servers > all over the world, they will easily be able use up all their bandwidth > without having more than 1 simultaneous connection to the same server. > > Andrew > > -----BEGIN PGP SIGNATURE----- > Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com> > > iQA/AwUBPmXCIgXqoqbqowOrEQJ0wwCeIn6pMyqmMkfxnytWASWqUpn3784AnjtX > rgb8eA+z+mTHzfjxrRFMGB2F > =02Yq > -----END PGP SIGNATURE-----
