Sherwood Botsford wrote:
>
> What are people's thoughts on TMDA (www.tmda.net)
> In principle it looks workable. However in a high spam
> location it will triple your email traffic. (If 90% of
> messages are spam with unique return addresses, then
> each incomming message generates 1 additional outgoing
> message, with that possibly bouncing, or sitting on the
> queue.
>
> I'm looking for stories, good an bad.
> --
> Sherwood Botsford
> St. John's School of Alberta
>
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##
>
>
>
Have used it with Qmail, not (yet) with Exim, but FWIW:
A) IF (haven't tried this, but should be do-able with Exim) one does
decent (80-90%?) spam blocking *before* hitting TMDA, extra traffic
should be minimal.. That said, some probes will themselves be bounced,
etc...
B) Even in the 'perfect' (zero spam) environment, lots of folks won't
(want to) understand the TMDA message, hence won't respond. Reduce this
factor by manually building an initial per-user whitelist. PITA.
IMHO, brute-force and Bayesian tools do a 'good enough' job to leave
relatively little gain from TMDA - taking the resistance of legit
correspondents to respond to the one-time challenge into account...
HTH,
Bill Hacker
