Low-load spam filtering (was Re: ticket tracking software)
andrew.jorgensen at gmail.com
Wed Apr 19 09:16:38 MDT 2006
On 4/19/06, Gregory Hill <Gregory_Hill at tni.com> wrote:
> > I have an analogy for this. Better to have an immune system than live
> > your life in a clean room.
> > With a simple bayesian filter I recieve something like 400 spam on a
> > low-volume day and see 1 or maybe 2 a day make it through the filter.
> > haven't had to worry about where my email address shows up in years.
> I'd rather save the extraneous server load and bandwidth. To each his
> own, but I'll take the clean room approach. What happens with the HIV
> spam start hitting and destroy your immune system?
My spam filter consists primarily of some simple restrictions on the
way HELO is used and an RBL. It seems like eventually spammers will
fix their servers, and even their worms, so that they conform to the
SMTP spec but this cuts out almost all spam (I get maybe three per
week now). Both of these methods are very low-load, low-bandwidth
(your server disconnects them as soon as they fail the HELO tests or
the RBL shows them listed).
I also use greylisting, but almost nothing gets that far anyway and it
delays email from people who haven't emailed you before I'm
considering removing that.
More information about the PLUG