-----------------------------
In reply to: 95931114031659/0003765414NA2EM
Application message id: 0100044086220735
Importance: Normal
Grade of Delivery: Normal
-----------------------------
We can only achieve real convergence if we really converge. <<
-- Einar Stefferud, Nov 14, 1993
X.400 just needs to use INTERNET routing technology, which is working
just fine all around the globe. It is here now, and it works. It can
work for X.400 just as well as for SMTP if X.400 uses SMTP addresses. <<
"If X.400 uses SMTP addresses" ... is another requirement to
*replace* the current network platforms in use. The way to go, and the
common point of convergence is X.500 based directories. Besides
pointing to the unique mailbox address (be it RFC-822, or Terminal
Address, or X.400 Mnemonic Address or ...) for a non-unique personal
name, it can point to a PSAP address for a domain. X.500 can
incorporate SMPT addresses, DNS can't do the reverse.
I think that the two-level domain structure as specified in X.400
was an advance in the early eighties, since it allowed very simple and
small static routing tables. But this is now bypassed by the dynamic
routing tables as offered by the DNS -- or even more powerful, X.500.
I think intermediate MTAs and domains will still be required, but
only in some specific cases: a) when incompatible Transport/Network
stacks require an intermediate hub (e.g. ISO-TP4/CLNS, ISO-TP/CONS,
proprietary transport stacks, or RFC1006); b) when some conversion is
required, including between RFC-822 and X.400 body parts;
c) as a security measure. And, of course, for a transitional period,
when a node is not able to dynamically find a route.
My proposal is to handle this as a Defect Report, on the basis that
someone forgot to include RFC822 adress types in the beginning ;-)... <<
Nice idea. This might be repeated with CLNS (ISO 8473) addresses,
where they forgot to reserve an AFI for ARPA-IP addresses.
Of course, if this is accepted, it will break the PTT Monopoly on
X.400 Address Assignments. <<
I think this is a wrong perception of reality. There is no such
monopoly. For one, X.400 was designed as a distributed system, without
a single national "mapping authority". The ideas of a "national
backbone" and a "single national registration authority" came from
other quarters, not the CCITT and ITU members. Secondly, wherever you
look (with very few exceptions), there are more than one X.400 ADMD per
country (code). Have a look at my ADMD listing, available by FTP from
the ITU-DOC server in Geneva or in some forums on Compuserve.
The Internet is actually much more centralized than the public data
networks and the international X.400 e-mail network. The physical
network structure is more centralised, especially in Europa. In the
Internet there really is a single authority for assigning network
addresses ([begin footnote] the US military in final analysis, right?
Some will not want to depend on that since they don't want to
depend on the military wing of their main competitor on the world
market, others object to the record from Hiroshima via Korea and
Vietnam to Iraq and Somalia -- which doesn't keep Cuba from opting for
RFC-822 mail and the US army to use X.400 :-). [end footnote]);
this centralization is not undone by the fact this some of their
functions is delegated to subordinate authorities; these subordinate
address distribution agencies can only assign addresses which have been
assigned to them for redistribution, a strange concept in the world of
the "PTT monopolies".
The CCITT based network and e-mail addresses have always been based
on a voluntary international cooperation of equal partners.
Kind regards,
L.Willms