Qmail
[Top] [All Lists]

Re: [Christmas Gift for free] 50% Spam reduction !

To: qmail@list.cr.yp.to
Subject: Re: [Christmas Gift for free] 50% Spam reduction !
From: John Simpson <jms1@jms1.net>
Date: Sun, 24 Dec 2006 16:09:57 -0500
Delivered-to: sp-com-lists@consult.net
Delivered-to: gmail-qmail@securepoint.com
Delivered-to: sp.com.list@gmail.com
Delivered-to: mailing list qmail@list.cr.yp.to
In-reply-to: <3.0.6.32.20061222190856.00b84580@192.168.192.1>
Mailing-list: contact qmail-help@list.cr.yp.to; run by ezmlm
References: <3.0.6.32.20061222190856.00b84580@192.168.192.1>
On 2006-12-22, at 1308, Erwin Hoffmann wrote:

introducing a delay in qmail-smtpd EHLO greeting by about 30 secs will
reduce the amount of spam emails by typically 50%.

In case you believe it or even if don't believe it, use my 5 line patch for
qmail-smtpd.c available here:

http://www.fehcom.de/qmail/qmail.html

you can take the same thing one step further... many spam programs send their first command without waiting for the server to send the initial banner. this is a violation of RFC 2821 section 3.2, and RFC 821 section 4.3, both of which say that the client is supposed to wait for the server's greeting before sending anything.

http://qmail.jms1.net/patches/greetdelay2.patch adds logic to hang up on clients who "jump the gun" line this. basically, instead of calling sleep(), it calls timeoutread() and barfs if it gets any result other than "timeout".

i'd like to hear others' thoughts about this before i add it to my own combined patch- is it a good idea? are there any "legitimate" clients (MTAs or MUAs) out there which ignore this part of the RFC, and if so are there enough to prevent this idea from being widely used?

--------------------------------------------------
| John M. Simpson - KG4ZOW - Programmer At Large |
| http://www.jms1.net/           <jms1@jms1.net> |
--------------------------------------------------
| Mac OS X proves that it's easier to make UNIX  |
| pretty than it is to make Windows secure.      |
--------------------------------------------------


Attachment: PGP.sig
Description: This is a digitally signed message part

<Prev in Thread] Current Thread [Next in Thread>