Re: Melting proposals together

Richard Rognlie <[email protected]> Fri, 3 Oct 2003 09:49:10 -0400
Newsgroups gmane.ietf.asrg.rmx
Message-ID <[email protected]>
> What do we focus on? 
> 
> - EHLO/HELO name
> - envelope sender
> - header sender (From:)

I am VERY concerned about looking at the header sender (From:).
this will break all known mailing list software.  I send a msg to
a mail msg to [email protected].  It gets processed and
sent to the list at large.  Yes... the env-from can be munged to 
something RMX compatible.   but the From: should remain "me".
(or the list, depending on the list properties).

> 
> I actually have doubts about the EHLO/HELO name. There is 
> no need for a MTA to announce the sender domain names in the 
> HELO name. The HELO name should be the name of the machine, not
> related to the following emails. E.g. the machine delivering 
> this particular email for me is sklave3.rackland.de, but it 
> delivers from [email protected].

This is a true statement.  DRIP does not try to do ANY sender domain
enforcement. It is only attempting to address the issue os malware MTAs
connecting to random hosts and always claiming to be "machine" or
"machine.hotmail.com".  It allows the domain owner to say, "if you're
getting a connection from a machine *claiming* to be machine.hotmail.com
it had better be coming from IP x.y.z.t  (by virtue of checking the
DNS of x_y_z_t.ipv4.relays._email_.machine.hotmail.com and getting back
x.y.z.t) 
an IP other than x.y.z.t tells you that there's a disconnect.  
an IP returned of x.y.z.t tells you they MIGHT have DRIP set up, you
need to check a "bogus" lookup as well to determine if x.y.z.t is a 
wildcard return or not..

an NXDOMAIN doesn't tell you much... just that they don't have DRIP set up
for that IP, AND they don't have a wildcard 0.0.0.0 set.

> How can we assume that this name can be used for sender
> authorization if it is not sufficiently related to the sender?
> 
> 
> A second problem is that this could break existing software.
> Many MTAs do not give a proper name at HELO/EHLO, or do not
> issue this command at all. 

And clearly this is a violation of RFCs then.   RFC2821 spells out 
what is expected.



-- 
 /  \__  | Richard Rognlie / Oracle Prophet / Gamerz.NET Lackey
 \__/  \ | http://www.gamerz.net/rrognlie/    <[email protected]>
 /  \__/ | It is dangerous to be right, when the government is wrong.
 \__/    |                                              -- Voltaire