----- Original Message -----
From: "wayne" <wayne(_at_)midwestcs(_dot_)com>
To: "Hector Santos" <hsantos(_at_)santronics(_dot_)com>
Sent: Tuesday, June 22, 2004 1:40 PM
Subject: Re: FW: Drive Towards Consensus
Caller-ID is dead. It looks like XML will be dead by the end of the
week.
I think it is important for the MARID group see what you think.
Also, I haven't been paying attention to our statistics over the last few
months, but what I see now, June 2004 is the first month that the number of
MCEP sites looked up is nearly equal to SPF sites.
See http://www.winserver.com/antispam
It should be noted the LMAP remains to be a extremely low percentage of the
total anti-spam schema.
Which brings up another important point related to what you like to bring up
much: Total SPF domains currently published vs. MCEP.
It is not the total domains in DNS but the "Personal Mail Community and
Association" of a particular domain that defines the effectiveness of LMAP.
If your domain has a lot of mail from a particular group, and this has a
majority support for SPF, your rate will be high. But in general, this
will not be the case.
On the other hand, if the top ISPs domains exploited by spammers published
LMAP records, then overall, everyone will benefit. In addition, if a
community of related servers and her customers published records, they will
all benefit among themselves as been the case with our network of customers.
This all boils down to the #1 benefit of LMAP - protecting your own domain
and as we have learned with our product development while implementing the
LMAP methods, new SMTP designs should inherently include support for what I
called Local DIPs: Local Domain/IP associations.
Our statistics shows that atleast 10-15% of spoofers will use your own
domains. That is an instant, highly low overhead optimal rejection at SMTP.
DNS is not required. And it makes sense. The SMTP receiver should be
looking at the 2821 HELO, MAIL FROM domains to perform a straight forward
local DIP check. This should be a SMTP BCP or written into the RFC. Just
imagine in terms of network topology, if every node did a local DIP check,
it would result in major benefit both locally and as a network with
absolutely NO DNS overhead, locally or network wide.
Anyway, I hope this isn't a cut and dry issue as you stated "Caller ID is
Dead", etc. I certainly hope not. I have voiced my opposition only to ONE
CONCEPT that Caller ID promotes - 2822 validation I think this is the wrong
direction for both local and network wide operations. This will be
especially the case if a company like Microsoft does go ahead with Caller ID
in her products. This will force many competitors and the industry to
support it too thus instantly creating a new undesirable higher than normal
PAYLOAD mode of operation. Yet, I am also on record that if we altered the
SMTP transaction model to better interact with 2822 validation concepts,
then I am for it.
--
Hector Santos, Santronics Software, Inc.
http://www.santronics.com