xiven.com stating the blatantly obvious since 2002

The Scourge

For the past... 197 days, I've been running an anti-spam program on my e-mail server. It's called the Anti-Spam SMTP Proxy, and the reason I wanted to try it is because it blocks the spam during the SMTP transaction itself.

The idea behind this is that the mail is tested (through various means) for being spam while the sender is communicating with my SMTP server and if spam, it immediately returns a 5xx SMTP error code with appropriate description of why the mail was blocked. This means that in the unlikely event of a genuine e-mail being rejected as spam, the sender will at least receive a notification that the e-mail was not received rather than it (a)just disappearing into a black hole or spam folder that no-one bothers to check or (b)sending a reply to the sender address (which with actual spam is almost certainly forged and thus a very bad idea).

Well in those past 197 days I have had a total of 48,319 e-mails go through my server:

• 10,725 of these have been blocked due to having an invalid or forged HELO
• 630 have been blocked due to failing SPF verification
• 27,270 have been blocked by checking Relay BlackLists
• 185 blocked due to having no recipient address
• 6 with viruses detected (I only enabled virus scanning yesterday)
• 2 from domains I have specifically blacklisted

leaving 9,501 e-mails. The survivors then go through a Bayesian filter (though since I don't fully trust the results of this and I don't want to block non-spam, I use it in the more traditional way of a spam folder - this may change eventually) which blocks another 5,224 mails, leaving me with a (mostly) spam-free 4,277 e-mails.

I'm not using all of the features (by a long way) of this thing yet. In particular I'm not using greylisting - I'm still undecided about the cost/benefit trade-off of it.

Ultimately though, I'm very satisfied with it so far. So much so that my e-mail address can once again be found on this website.

Posted: 2007-05-01 15:00:16 UTC by Xiven | Cross-references (0) | Comments (1)