[Top] [All Lists]

Re: [Another long] commnets on draft-crocker-email-arch-01

2005-02-09 06:07:49

Dave Crocker wrote:

   | -<+
   |   |
   V   | SMTP error (500+)
  MTA -+
The chart is already quite busy, so my feeling is that it is
better to leave protocol-level feedback out of it.

It's still an interesting detail:  If MTA A forwards mail to
MTA B, and B "bounces" (rejects) it with an SMTP error, then
MTA A creates the corresponding "bounce message" (error
notification).  When users talk about "bounces" they generally
refer to the 2nd meaning of "bounce" = "bounce message".

It's also relevant for protocols.  During the design of SPF
some users wanted "I18N" for error messages.  But of course
that won't work, because MTA A only "copies" the SMTP error
of B into its bounce message.  And MTA B has no reliable way
to say that his SMTP errors are reported as Latin-1, UTF-8,
or whatever else.

  rMTA an MTA that is named in the receivers MX list
1. What are the required, functional differences between
these MTAs?

An MX talks with (most) "unknown strangers" (inbound).  And
a "mailout" talks to "unknown stragers" (outbound).  Other
MTAs (MSA, MDA, normal relays) normally only talk with their
buddies.  They all have their own acronym, only the "mailout"
is missing.  You probably can't say RMX, but it's the missing
piece in this puzzle.

2. It bothers me that I have left the DNS entirely out of the
discussion.  I suspect this should change.

Something about MX should help.

You are using POP3 as a mail Relay protocol.  This is done
periodically, but is very far from mainstream use, as far as
I know.  Again, the architecture document is not trying to
list every possible scenario.

You mention ODMR (2645) in the references, apparently they use
POP3 to emulate ODMR, and in that sense you have it.  If their emulation
is "wrong", then it's still a "gateway".

                           Bye, Frank

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