On Wed, 2003-09-17 at 16:34, Parker Morse wrote: > On Friday, Sep 12, 2003, at 19:18 US/Eastern, Mike McMullen wrote: > > Now that the dust has settled, I'd like to know what combination of > > lists > > is working well with people. > > There's a nice comparison online at > <http://www.sdsc.edu/~jeff/spam/Blacklists_Compared.html>.
That depends upon your environment. ALL blocklists produce false
positives. White listing is a key element. We are encouraging our
clients not to use any blocklists.
Our server averages about 6,000 incoming emails per week and we get
virtually no spam delivered (bouncing back about 30% on average).
You might consider the experimental release of Postfix (which we have
been using for more than a month without a hiccup). There are new
REDIRECT and FILTER capabilities that allow you to be very aggressive
without false positives. We're using three instances of Postfix. Because
of its structure, this creates minimal overhead.
Instance One has all the whitelist data and does all the envelope checks
(HELO, Client, Sender, etc.). Anything that is neither whited nor
bounced goes to Instance two for RegEx header and body checks. Instance
three is used strictly for outbound (unnecessary but it seems logical).
You just have to watch tails of one and two to know how well
choreographed all of this is.
--
----------------------------------------------------
Hart's PGP Key: 0x7BFF655E - http://TQMcube.com/hart_pgp.txt
----------------------------------------------------
Total Quality Management - A Commitment to Excellence
Email acceptance policy: http://www.TQMcube.com/email_policy.html
signature.asc
Description: This is a digitally signed message part
