On 11 Jan 2012, at 12:02, Andreas Berton wrote:
Is it possible to use a global user address to manage the delivery to
final destination. So delivery looks something like
u...@myhost.tld glo...@myhost.tld u...@destination.tld
There is absolutely nothing clear about that question. Really.
The phrase "global user address" lacks whatever specific formal meaning
you may think it has. It is not a phrase that appears in the Postfix (or
Sendmail) documentation and in 20 years of running mail systems I don't
recall seeing it used in any formal way. But I confess to getting old
and forgetting things...
In a sense, any Internet email address is "global" in that it is made up
of a domain part and a local part. The domain part is resolvable to mail
exchanger names which in turn resolve to IP addresses via DNS, and the
systems answering port 25 connections on those IP's are expected to be
able to validate local parts and handle their delivery. For a "global"
scope that shares a functionally unified view of DNS (i.e. the same root
authorities and honest resolvers) any address with a resolvable domain
part is the same address globally. There is no "global" directory or
registry of Internet email addresses, and no federation mechanism that
can mimic such a system, unless you count the process of trying to
deliver email.
Shorter: X.400 is mostly mythical (especially outside of obsolete
versions of MS Exchange)
If this is possible, could such scenario create any holes or overides
the normal control of realy processing. And would it be transparent,
meaning without rewriting or messing with the header.
Some examples is appreciated, and where I can read more about this.
The protocol for how email is delivered between unaffiliated Internet
domains is the "Simple Mail Transport Protocol" and it truly is simple
when compared to other schemes that people have used and proposed. The
critical reference for that is RFC5321, which you can find with links to
its references and predecessors at <http://tools.ietf.org/html/rfc5321>
I'm largely guessing, but you may be asking about a mechanism similar to
the "Source Routing" that is used in protocols like UUCP (and X.400 if
my memory is correct) and which was specified in early versions of SMTP
as well. Technically such addressing is still valid for SMTP but you can
pretty much count on it being shunned and/or ignored by modern servers
(see <http://tools.ietf.org/html/rfc5321#section-3.6.1>.)
--
Bill Cole
