greylisting

Sat Feb 7 03:33:21 UTC 2004

well, it might not scale only at the global scale :)
(while its current 'deployment' is far from being global
as far as i can see)
--
dima.

> -----Original Message-----
> From: owner-nanog at merit.edu [mailto:owner-nanog at merit.edu]On Behalf Of
> Suresh Ramasubramanian
> Sent: Friday, February 06, 2004 10:19 PM
> To: Ken Leland
> Cc: nanog at merit.edu
> Subject: Re: greylisting
> 
> 
> 
> Ken Leland  [2/7/2004 2:11 AM] :
> 
> > On Fri, Feb 06, 2004 at 01:40:14PM -0500, Dmitri Krioukov wrote:
> > 
> >>can someone provide any comments on greylisting?
> >>how effective is it, etc?
> > 
> > 
> > we have 20 regex expressions that block mail directly from a 
> > reverse dns that "looks" like a consumer broadband connection. 
> > Then we have to maintain a whitelist on a case by case basis.
> 
> 
> Er, I think you and Dmitri are talking about different things.
> 
>  From a mail operations standpoint, I am not a big fan of graylisting, 
> because even legitimate senders get 4xx'd for a while, the first time 
> they send mail.
> 
> When any such strategy means that someone else's mail queues are filled 
> with timed out emails waiting for retransmission, it d not be rocket 
> science to see why this just doesn't scale too well.
> 
> -- 
> srs (postmaster|suresh)@outblaze.com // gpg : EDEDEFB9
> manager, outblaze.com security and antispam operations