--On Wednesday, March 16, 2005 15:10 -0600 Chris Myers <[EMAIL PROTECTED]> wrote:
As posted to the (moribund) IETF Anti-Spam Research Group mailing list.
See http://www.titankey.com.
The web pages are disappointingly full of marketing hype. According to this, no other product provides for rejecting mail during the smtp conversation, and spammers usually remove addresses when they get a 550 response. I say disappointingly because it does seem to have some interesting features.
Whitelists per user are probably prior art even in 2000. This however filters not only per user but per full address using something akin to plus addressing (the example given uses "dot addressing"). The whitelist itself is built not only from direct user input but also from watching outgoing mail and possibly other methods (alluded to but not detailed). For non-whitelisted addresses the response can be either plain reject or challenge-response.
The host running the database has to see all mail in and out. I didn't notice anything about replication and scaling.
How's that again: "we've got > 8000 end-users on it with > 2M emails/day passing through our software" ... that's an average 250/day per user, which is about ten times what we have per user.
Joseph Brennan Academic Technologies Group, Academic Information Systems (AcIS) Columbia University in the City of New York
_______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [email protected] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
