What happened at the IETF?

Is the IETF dying as an institution?

Many may think so, as they gathered together for the largest IETF ever: more than a thousand networkers had gathered together for four days of talk and work at the San Diego Fairmont hotel.

Again and again, a pattern repeated itself: thirty people in a room, three or four talking; the rest listening, contributing little, perhaps learning, perhaps not.

The talkers seemed impatient with the listeners; "this is not an educational conference", they said, and launched long diatribes spattered with acronyms against positions supported by equally incomprehensible acronym collections.

Yet - work got done.

Standards were worked out. Compromises moved closer to reality. Some people understood problems a little better, learning respect for each others' positions through the give and take of free-flowing discussion.

For many, the simple fact of being away from their offices gave them the first time they had had in months to focus on the technology that would lay the premises for their work in the future, rather than to focus only on making products that cashed in on last year's agreements.

Of course, you could always hear the click-click of the keyboards in the background; no self-respecting IETFer seems to be able to think without a laptop in his hands, and it seems perfectly acceptable to sit back and read E-mail while someone else is trying to defend a point that, in the typist's opinion, does not need his attention.

For these are busy people; Vint Cerf has moved to be senior vice president of data architecture at MCI ("they have 4 times as many people as Microsoft"), who seems to be hiring top-notch IETFers left and right. Ned Freed is rarely seen without a pager; the phrase "in my copious free time" is always uttered in the ironic mode.

Yet, there seems to be less of the enormous driving energy that could be found at previous IETFs; dinner conversations might center around skiing as easily as on the relative merits of Mosaic and Netscape; one rarely heard voices raised in anger even in the heat of the URN confusion.

But one should go back and look at the concrete events; only the future will show how effective we were.

Events in the networking area

Here I can report no more than vague feelings and overheard grumbles; I didn't attend any meetings, and listened only glancingly to presentations and corridor conversations. Still, IPv6 seems to be basically on track; it is my belief that one can start pushing products Real Soon Now.

And we DO need it; the Address Lifetime Expectation working group has moved the date for the total exhaustion of IPv4 addresses from 2005 +/- 3 years to 2003 +/- 3 years.

Back to what I know better - the E-mail area.

E-mail: Moving forward

Joy! Jubiliation! Progress!
The latest proposal for MIME/PEM integration, as usual published 2 weeks before the IETF, was accepted by the working group!!!!

The true significance of this event may only be apparent to those who have watched the offended feelings in this group for a number of IETFs, but still, it removes an important stumbling block on the way to a standardized, secure E-mail infastructure (and, by the way, clears the way for using PEM in the way that PGP is now used: without a central authority).

Less remarked upon was the conclusion of the MHS-DS (routing of X.400 messages using X.500) working group, which had its last document handed over to the IESG for publication as an Experimental Standard.
It is quite possible that this work will be continued under the auspices of ISO, if it continues; the future of X.400 in the Internet does not seem too rosy.

Nothing moved on the surface of the RFC 1327 (X400 to Internet mapping) review team; talk in the corridors indicate that Steve Kille, the author of RFC 1327, is very busy, and may not have time to do the necessary rewriting for some time yet. We need to do something about this - but what can we do?

EDI is getting somewhere too; it turns out, as expected, that the description of how to do EDI on the Internet using MIME is a 2-page document, and that the hardest section was the references.
It also turns out, equally expected, that the companion document envisioned describing how to sensibly evaluate the problems and promises of using the Internet for EDI is showing very little progress, as usual due to people not having time to work on the document between meetings.

NOTARY, the work on standardizing positive and negative delivery reports in Internet mail, was not satisfactory. The existing documents were widely regarded as "too complex", but extensive discussion both before and at the meeting failed to uncover a simpler model upon which a document rewrite could be based. Keith Moore will try to rewrite the documents using the existing model, but deleting most fields whose definition seems to leave doubt about what they mean, leaving a "lean and mean" (but fully extensible) report format.
I hope this works out OK - we all need this work, and soon.

The MAILEXT working group, which concerns itself with standardizing "odds and ends" of E-mail work, reviewed documents as usual. The document for language tags got a thumbs-up; the documents on mail based servers (the "bombs" series) were thought central enough that a new working group would be set up for it.

Little work had been done on the Binary, Streaming and FTAM-Body-Part documents since the last time; still, little work was expected in order to get them to standard-class status; expect to see them soon.

This about concludes the E-mail area. The rewrite work on the basic MIME documents is rapidly approaching its end, leading to the adoption of MIME as a full Internet Standard; the biggest change expected is that the "Binary" transfer-encoding will be removed if one cannot show multiple, interworking implementations before it goes to Full Standard, which will in turn mean that one has to get the Binary SMTP extensions finished and tested.

Other matters

I sat in on some of the URI sessions.
It seems that this group is now overcoming its constipation with the fetails of URL syntax and meaning, and is now doing productive work on the tools we need for the next generation; names and name-to-location translating services.

Security is also an important topic; there was a (quite productive, I believe) session on guidelines for CERT incident handling, a meeting of the Common Authentication Technology group (of which I understood little to nothing - this area is as acronym-infested as most!), considerations of DNS security and incremental DNS updates (making sense out of getting a new IP number every time you connect, for instance), considerations of security in other areas; most of which I did not attend.

All in all, if you look at the productivity, it seems that the IETF is fairly well along; the old body might have a few kicks left after all!

Or, as one working group chair said it: "We've been worrying about the death of the IETF for the last 8 years - let's contiune doing that!"


Harald.T.Alvestrand@uninett.no
Last modified: Fri Dec 9 21:06:55 1994