[PLUG] Smart greylisting...

someone plug_1 at robinson-west.com
Thu May 31 11:27:43 UTC 2007


Instead of blanket greylisting, I want to restrict when I graylist
by looking at who on my list of recipients a sender wants to reach.

I'm interested in smarter ways to use graylisting than the old
blanket approach.

It want to restrict graylisting to when something appears to be
wrong with the source.  Maybe it's a dynamically assigned ip
address, one that isn't registered in DNS databases at all, or
maybe the helo/ehlo is something suspicious.

I think what I'm talking about is having the outcome of
one smtp filtering policy impact or prevent the application
of another policy.

Say I have a relay recipients list like so:

foo at example.com           graylist
abuse at example.com         x
mail-list at example.com     x

bar at example2.com          x
abuse at example2.com        x
mail-list at example2.com    x

Looking at the above list, understand that the x means let it
through without graylisting it.  Understand also that mail to
example2.com should not be graylisted because that is a
friend's domain I am a backup relay for.  Slowing down his
legit customers on accident is unacceptable.  This example
is not real life, any relation to real life is purely
coincidental.

Does anyone maintain a decent list of which sources cannot be
graylisted?  Are there any secrets to knowing when graylisting
should not be used?

I am using postfix, but that only matters when I try to
implement something.

    --  Michael

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.




More information about the PLUG mailing list