Resent-Date: Wed, 3 Mar 1999 18:16:47 GMT
Resent-Message-Id: <199903031816.SAA51788@out5.ibm.net>
Message-Id: <199903031816.SAA51788@out5.ibm.net>
From: "R. Kelley Cook" <kcook@ibm.net>
Resent-From: "R. Kelley Cook" <kcook@ibm.net>
To: "drums@cs.utk.edu" <drums@cs.utk.edu>
Date: Tue, 02 Mar 1999 17:54:17 -0500 (EST)
Reply-To: "R. Kelley Cook" <kcook@ibm.net>
Priority: Normal
In-Reply-To: <4.2.0.25.19990223115607.00d18e60@shell2.bayarea.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Subject: Re: To MX or not to MX?

On Tue, 23 Feb 1999 11:59:30 +0800, Dave Crocker wrote:

>I concur with them both.  IETF standards pertain to the global Internet and 
>what is required for operation there.
>
>There is nothing wrong with products that permit deviations, for use in 
>private circumstances, but it is essential that the IETF documents make 
>clear what is required for adequate interoperability over the public 
>Internet.  For example, absence of MX support renders an SMTP client 
>useless over the public Internet.

No it doesn't.  

A MUAs communication request goes out over the internet via port 25. 
It talks to the one SMTP server it knows about.  Conversely, that one
server should know that the client is a trusted source.  Most
importantly, it works quite well for over 100 million people everyday.

Unless I was horrendously mistaken, this document is not supposed to be
just for SMTP server to server communications.  By requiring MX lookups
in all instances, Mr. Crocker, Mr. Elz, and Mr. Klensen have changed it
to be so.

 -- Kelley Cook