SMTP operational experience in mixed IPv4/IPv6 environements
draft-ietf-ngtrans-ipv6-smtp-requirement-08

[Note]: 'Rev 8 just says that doc has expired.
I am also checking with APPS ADs for the APPS area review that Randy suggested back in October' has been cleared by Bert Wijnen

[Note]: 'Rev 8 just says that doc has expired.
I am also checking with APPS ADs for the APPS area review that Randy suggested back in October' added by Bert Wijnen

Recording discussion I (Bert) had with itojun:

-----Original Message-----
From: itojun@itojun.org [mailto:itojun@itojun.org]
Sent: vrijdag 19 december 2003 2:46
To: bwijnen@lucent.com
Cc: iesg@ietf.org; pekkas@netcore.fi; Jonne.Soininen@nokia.com;
bob@thefinks.com; itojun@itojun.org
Subject: RE: Re: draft-ietf-ngtrans-ipv6-smtp-requirement-07.txt


> Itojun, I discussed this with v6Ops WG chairs.
>
> It seems to me/us:
>
> - you update to get rid of "open issues" of which
>  one seems resulved by your email below
> - you may want to check the security considerations
>  section and beef it up a bit
> - you submit new rev, and then inform v6Ops Chairs
>  and myself, and we ask them to ask v6Ops WG to do
>  a review/check of the document.
> - If they have comments you rty to address them.
> - You submit to RFC-Editor as an individual submission
>  (Informational I assume), and then when it gets
>  to IESG I will shepherd it through IESG. Should be
>  OK by then
>
> Sounds like a plan?

yup.  will work on new rev.

itojun

it may not reflect all the coments i have received, but anyways
better than nothing.  i think it should really be integrated into
update to RFC2821 (instead of being standalone document).  it is
still using "ngtrans" in the name, however, it is suggested that
the document should be discussed in APP area (there's no
email-related WG exist at this point).

Date: Sun, 18 Aug 2002 14:30:55 -0400
To: Itojun@iijlab.net
From: Margaret Wasserman
Subject: Fwd: Re: Fwd: draft-ietf-ngtrans-ipv6-smtp-requirement-06.txt
Cc: Rob Austein , Randy Bush

Hi Itojun,

I was talking to Rob Austein the other day, and it turns out that he
has significant real-world experience configuring mail systems and
MX records.  So, I asked him to review the ipv6-smtp-requirement draft.

Here are his comments (forwarded with his permission).

It looks like there is definitely something here that should be
published, although the document may need some editing.

After the v6ops charter has been approved by the IESG (hopefully this
Thursday), Randy will contact the Apps ADs to discuss how we can move
forward on this document.  Not clear yet whether it makes sense to
do this in the Apps area or in v6ops, but either way it'll probably
be a collaboration.

Margaret

>Substance:
>
>a) MTAs must look for both AAAA and A RRs, unless they have no idea
>    what to do with one of those, in which case they should only look
>    for the one they know what to do with.
>
>b) One should not set up one's MX RRs so that one's backup mail
>    servers cannot possibly forward mail to one's primary mail servers.
>
>c) Getting SMTP over IPv6 to work may require one to upgrade broken
>    DNS name servers that return the "Server Failure" response code
>    when asked about AAAA RRs.
>
>d) Scoped addresses suck, particularly when used in applications that
>    have no good reason to know or care when they are crossing a
>    address scope boundary (in other words, scoped addresses suck in in
>    almost every known application, including SMTP).
>
>If there are substantial points beyond the above, I missed them.  Much
>of the document is introductory material that could and probably
>should be replaced by a single sentence: "This document assumes that
>the reader is familiar with RFC 2821."
>
>Point (a) should be pushed off to a small standards track document,
>since it's a modification of RFC 2821 section 5.  I don't see any
>obvious current APPS WG to which this could be passed.  Looks like
>about two pages of text, plus boilerplate.  Sad that this isn't in RFC
>2821, given how recently that was published, but c'est la vie.
>
>Points (b) and (c) are operational advice.  They could be kept in the
>same document as (a), but they're not really standards track subjects.
>Dunno if Randy has an opinion on this.
>
>Point (d) is the IPv6 WG's problem.  Well, at the rate we're going
>it's going to be the entire human race's problem, but the "right" fix
>would be repeated application of a clue-by-four to the heads of
>certain members of the IPv6 WG.

To: iesg@ietf.org
Cc: ngtrans@sunroof.eng.sun.com, Ned Freed ,
        Ted Hardie
Subject: draft-ietf-ngtrans-ipv6-smtp-requirement
From: itojun@iijlab.net
Date: Wed, 09 Jul 2003 17:31:26 +0900

a couple of IETFs ago, it was suggested at the ngtrans wg meeting
that the document draft-ietf-ngtrans-ipv6-smtp-requirement-xx "should
be handed to SMTP experts in application area".  since then, there
has been no actions taken.  the fact is, there is no active working
group for SMTP and email-related activities.  also ngtrans wg is
concluded.  therefore, the document is in "zombie" state.

I still want the document (or the contents of it) be published, as
it contains critical guidelines for IPv6/v4 dual stack SMTP operation.
I would like to ask IESG to decide and take appropriate actions,
such as:
- poke SMTP experts, let them work on the topic and produce an RFC
- let me handle the document as an individual submission and publish
  it as an RFC

thanks.

itojun