I would suggest first looking into MailScanner (http://www.mailscanner.info). Reason being, this interfaces with SpamAssassin as well as virus scanners. It also provides you with lots of other possibilities of scanning inbound mail, all of which can be customized by you. It's a little involved, but I can speak highly of it and of the support available by Julian (the author).

MailScanner can optionally consult the RBLs out there - and allows you to define your own. It's written in perl, and I've never had any problems with it.

From there, I would consider adding to it SpamAssassin, and then virus scanners.

SpamAssassin uses a Bayes Algorithm (genetic) - it basically "learns" from the mail that goes into the gateway - over time, it gets a better idea of what's really spam and whats ham (non spam). It also can "learn" by your sending to it messages that were marked as spam that you want to accept, etc.

It's a little involved to get this all running, but it's well-worth the effort. I've been running mail gateways for several years, and this is the sign-of-the-times where we need to create more involved configurations to deal with the evolving spam/virus issues :)



Forrest


At 05:07 PM 7/3/2003, Chris Ochap wrote:
Yeah, It's been suggested to me by one of my co-workers that SpamAssassin
may be the way to go.  We run qmail/vpopmail/mysql and even though RBL's are
site-wide, we would like to keep TMDA & SpamAssassin set up on a per user
basis.  Would you be confident in saying that SpamAssassin is accurate
enough to trust with automatically reporting to an RBL such as SpamCop?
Thanks.

-Chris

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Behalf Of Forrest Aldrich
Sent: Thursday, July 03, 2003 12:52 PM
To: [EMAIL PROTECTED]
Subject: RE: SpamCop reporting


You could also try using SpamAssassin. After being run for a while, you'll find that SA is extremely accurate at identifying and tagging spam. It's the best product of this type I've ever seen, but can be a little involved to set up properly.

Along with that, and after some time of the SA Bayes engine learning, you
might be able to utilize spamassassin-milter (presuming you use Sendmail)
to act upon SPAM at that level.

I use SpamAssassin in conjunction with MailScanner
(http://www.mailscanner.info) which offers quite a bit of flexibility (and
it's free).



Forrest

_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users

_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users

_____________________________________________ tmda-users mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-users

Reply via email to