From: Aaron Bennett (abennettclarku.edu)
Date: Tue Jul 25 2006 - 11:54:18 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Joey wrote:
> Hello everyone,
>
> OK frustration has hit an all time high in respect to battling spam
> and not succeeding.
> On average using the RBL's ( agressivly ) home made block lists for
> header and body checks, and spam assassin we are blocking between
> 80-87% of all email coming to the server.
> The problem is that the remaining 13-20% email is inundated with spam.
>
I'd consider maia mailguard or a similar solution. The problem with
your system is there's no learning. Maia Mailguard does per-user
quarantining and per-user bayes learning so SA is provided with a
reliable and steady source of spam/ham email to learn from. We just
deployed it and it is stunningly effective; also we've had a total of
THREE false positives in a week with over 40,000 spams caught. It's
based on amavisd-new. The downside is it takes a reasonably beefy
machine because all email gets stored in a database (mysql or
postgress), but it just plain works.

--
Aaron Bennett
Sr. Unix Systems Administrator
Clark University ITS
abennettclarku.edu | 508.781.7315

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]