Hector Santos wrote:
----- Original Message -----
From: "Philip Miller" <millenix(_at_)zemos(_dot_)net>
To: "Hector Santos" <hsantos(_at_)santronics(_dot_)com>
Cc: <ietf-mxcomp(_at_)imc(_dot_)org>
Sent: Wednesday, March 10, 2004 1:52 PM
Subject: Re: LMAP Validation Analysis
The following IP/Domain association assertions are made:
LMAP = IP :: HELO domain
LMAP = IP :: MAIL FROM domain
Most proposals make it possible to assert one, the other, or both, while
this text makes it appear that both must be asserted.
Ok. I'll clear that up.
To make life more interesting, MS's Caller ID is based on IP/"From"
header relationship, with the IP address taken from the "Received
headers". It is different from the MAIL FROM domain which is stored in
the message as a "Return-Path" header.
If we consider MTA MARK as part of LMAP, then there is also an IP/rDNS
tree relationship at connect time before the HELO stage.
A side thought: all of these proposals seek to connect the forward DNS
tree of a domain with the IP (or in MTA MARK, rDNS with IP). There are
two parts to this - publishing the data in DNS, and the second part
involves actually applying the data (MAIL FROM, HELO, etc.). This
distinction might or might be important to the future.
Yakov