Publication of EMSD document

[braden@ISI.EDU]

Folks,

The RFC Editor has conscientiously examined and re-examined all the
facts, allegations, and issues concerning Banan's EMSD document, and
we have concluded that on balance, it would be better to publish it
and move on.

We received an advance copy of the proposed IESG statement from Steve
Coya.  While we agree with all the points made in that statement (and
in fact we raised most of these points ourselves), we believe that the
statement as written was a bit too long, a bit too redundant, and, yes,
a bit too punitive.  We believe that a slightly shorter and more
moderate form can make the same points and make them more forcefully.
We have therefore made an editing pass over the statement (we are not
the IESG, but we ARE editors!), resulting in the text below.  Details
of this are certainly negotiable, if you feel that some particular
points are not made clearly or strongly enough.

We propose to now proceed with publication of EMSD as an RFC, with our
edited version of the IESG Statement.  We sincerely hope that the IESG
will accept this as filling our obligation to publish the IESG's
statement.  This action is taken in good faith as we try our best to
work productively with the IESG and to serve the best interests of the
IETF community.

Bob Braden
Joyce Reynolds

___________________________________________________________________


IESG Statement
_____________________________________

The IESG recommended against publication of this document, for the
following reasons.  The protocol specified in this document may be
satisfactory for limited use in private wireless IP networks.  However,
it is unsuitable for general-purpose message transfer or for transfer
of messages over the public Internet, because of limitations that
include the following:

- Lack of congestion control

  EMSD is layered on ESRO [RFC 2188], which does not provide congestion
  control.  This makes EMSD completely unsuitable for use across the
  public Internet.  EMSD should be considered for use in a wireless
  network only if an EMSD-to-SMTP gateway provides email connectivity
  to the general Internet.

- Inadequate security

  The document specifies only clear-text passwords for authentication.
  EMSD should be used across a wireless network only if sufficiently
  strong encryption is in use to protect the clear-text password.

- Lack of character set internationalization

  EMSD has no provision for representation of characters outside
  of the ASCII repertoire or for language tags.

- Poorly defined gatewaying to and from Internet Mail

  Because Internet Mail and EMSD have somewhat different and
  conflicting service models and different data models, mapping between
  them may provide good service only in limited cases, and this may
  cause operational problems.

The IESG therefore recommends that EMSD deployment be limited to narrow
circumstances, i.e., to communicate only:

- with devices that have inherent limitations on the length and
  format of a message (no more than a few hundred bytes of ASCII
  text), and

- over wireless links with adequate link-layer encryption and
  gatewayed to the public Internet, or

- over private IP networks that are either very over-provisioned
  or have other means  of congestion control (e.g., medium access or
  flow control).

In the near future, the IESG may charter a working group to define an
Internet standards-track protocol for efficient transmission of
electronic mail messages, which will be highly compatible with existing
Internet mail protocols, and which wil be suitable for operation over
the global Internet, including both wireless and wired links.