Attention all hands - this is your area director speaking :-)
This IETF was my first meeting as an area director for Applications. It was an enormously learning experience, and after five intense days of wall-to-wall meeting discussions, corridor discussions, bar discussions, dinner discussions, terminal room discussions and discussions in dreams, this area director is now totally, utterly exhausted.
The fact that my hands keep rambling over the keyboard as the plane rises into the skies over Stockholm is a testament to the inspirational powers of the IETF; my logs show that I wrote 84 E-mail messages and 500 lines of other text - ideas, actions and reactions just poured forth like a flood.
Stockholm's Grand Hotel was a fantastic place for an IETF - and was made even better by the incredible networking arrangements of the KTH crew; where other IETFs have had two cameras and three mikes for a plenary, these guys had installed five mikes, each with individual lightning and remotely controlled camera - as a question was being asked, you could see the face of the questioner on the giant screen behind the stage; slides weren't shown on the overhead, but on the multicast whiteboard controlled by an onstage Sun machine. And if there was any problem, a green-shirted KTHer wearing a headset with a radiolink to Central Control would be there before you even started wondering about it.
And of course, the fact that the NORDUnet crew had managed to get the first ever transatlantic 34-megabit Internet line installed just before the meeting didn't exactly make people feel cut off from the network!
There'll probably be some changes in the terminal room for the next IETF - the RFCs for how to do encrypted and/or authenticated IP (in IPv4, not the future IP version 6) are now finished, and we were challenged to build crypto into our laptops and use it at the next IETF, combining security with interoperability in unique ways.
The suggestion of installing "the official IETF packet sniffer" with a large wall display in the terminal room, just to remind people of what they were doing, also met with instant approval.
But no matter what they do, I don't think Dallas has any hope of matching the incredible style of the Stockholm multicast and networking teams!
But back to technical issues; my view of this IETF was even more myopic than usual; an area director simply has no time for going to other groups than those in his own area. If you want to know what happened outside, go ask someone else; I can give a few tantalizing hints, but no more.
The basic protocols, RFC 821 and RFC 822, are now showing (by contrast) how much we have learned since 1982 about writing precise specification (and the value thereof!); a new working group, DRUMS, gathered to start the work of revising those documents.
The group's charter lets it do two things:
Another groupt, RECEIPT, has started in on defining standards for emulating one of the most commonly quoted features of the X.400 standard, receipt notifications or "read notifications". Work on two other extensions, NOTARY for delivery notifications and MOSS for a way of using MIME-style messages in a secure manner, is finished, and the documents are in the very last stages of the approval and publication process.
The relationship to X.400 is also the business of MIXER, the group charged with updating and revising RFC 1327, the basic document about how to interoperate between X.400 and Internet mail. This will incorporate the features of MIME and NOTARY, leading to the ability to use more functions across the borders of the two networks.
The keywords for work in this area are:
Anyway, nobody served intent to become my enemy for life, so I guess they aren't entirely unhappy with the development :-)
The rollout of the WHOIS++ service, centered around Bunyip's DIGGER product, served as a nice backdrop; one of the two documents defining the standard has already been approved as Proposed Standard, with the other being close to final edits, and Bunyip reports 40 or more servers connected to their service in the weeks since the announcement on June 20.
We expect to be hearing more from this field; a new activity was suggested, called "Common Indexing Protocol", giving a common means of finding information in X.500 databases and WHOIS++ databases, and loosening the constraints that have plagued the worldwide X.500 service with regards to searching.
The X.500 Root service, managed by DANTE, is in the throes of finding out
exactly how to change protocols from X.500/88 to X.500/93; there may
be work for the IETF too in this area; the changes in X.500 over these
years seem not to be small, and the implications may not be well understood.
It is also clear that X.500 is enjoying steady growth; DANTE reported
that the increase in number of entries was approximately 40% per year
in its responsibility area. X.500 is probably still by far the biggest
working directory on the Internet.
Another interesting number came from Tim Howes, who explained that since 95% of requests to his service came via the "lightweight DAP" protocol LDAP, he was considering rewriting the database engine to support LDAP directly, and then gateway "full DAP" onto it, rather than gatewaying the LDAP requests into DAP as he does today. Numbers are so much fun!
The "requirements" documents for Directory, output of the former WHIP group, will be rewritten to limit their scope and clarify their requirements, and finished - soon - in the IDS working group.
It is by now clear that we don't at the moment have specifications for a working, scalable, acceptable directory service for the Internet, but I would say that we are getting closer.
The good news is that they now regard HTTP/1.0 as a finished specification.
The bad news is that there is as yet no agreement on what features should go into HTTP/1.1, there is no significant sign of agreement on what the letters "HTTP/NG" mean, if anything, and there has been discovered one problem with HTTP's usage of TCP that is quite harmful to the proper operation of the protocol under some circumstances (the "connecton reset by peer during POST" problem; it's also known to occur on GET requests).
The HTML group turned out to be quite harmonious in comparision; the HTML 2.0 spec is in Last Call; other proposals (tables, maths, style sheets) are being prepared. One, Tables, was actually to be discussed at the meeting; very few people had read it, but most assumed that it would work OK. It's mostly compatible with the Netscape tables, with some added functionality that Netscape missed.
The greatest events happened in the URI group. To cut a long story short: The URI group is now closed down, to be replaced with a set of groups for smaller subtasks that need to be done, with short lifetimes and limited charters. These include:
Chris Weider, who is now on the IAB, is also initiating work to see if he can use the Internet Research Task Force (IRTF) as a platform to form work teams for exploring issues that it seems to be premature to write standards for now; these issues include an envisaged "information architecture for the Internet", "the meaning of information about information", the possibility of devising reasonable rules about when two objects are to be considered the same or not, and so on.
Quick results are not expected from this work; once it arrives at results, these may be used to form new activities and workgroups within the IETF for standardization based on the understanding gathered in the IRTF.
One group that will not form yet is the group for work on caching and replication; discussion between several people who are actively trying to make progress in this area indicated that they would have to do some more months of research before they could enter the standardization process with a goal they were reasonably sure would be useful, so this activity will not be progressed in the IETF at this moment.
There seems to be something happening in the US with regards to the
relationship between the Internet, children and politicans.
While it is somewhat hard for me as a Norwegian to understand the
excessive importance attached to recreational procreation in a country
where several thousand of its citizens die from gunshot-related causes
every year, it seems that several politicans are convinced (at least
in public) that the moral fiber of the nation will fray and snap if
the Internet is allowed to make any desktop PC a possible means of access
to images that could incite lust in the minds of the younger
generation.
(Yes, I know the problem is a bit more complex than that, but some of
the public posturing that some of these politicans do, and the
hardships that the laws they may pass because of this posturing can
bring down on our community fills me with such rage that I have little
incentive to tone down the particular piece of totally unbalanced invective
above).
Anyway, it was deemed terribly important by several distinguished members of our community that the IETF should Come Forward And Show Its Concern - and the result was the so-called "Read The Label BOF".
Held in the largest room in the Grand hotel, multicast over the world by the MBONE, and brilliantly led by our grand guru and founding father, Vint Cerf, the BOF did exactly that.
The outcome of the BOF was rough community consensus that:
Off the record, the BOF was often mentioned as the "RTFL" BOF, for obvious reasons. It was also a nearly comletely unanimous consensus that any group who tried to take this on in any way, shape or form would get nothing whatsoever done outside of this topic due to the political dynamics of the situation; separate WGs were a good idea.
A good many people also suggested that ratings were a very salable product, not because of the potential for protections, but rather in the opposite direction - "you need protection against what I want to find" seems to be a not uncommon modus operandi.....
The organizers had done a great job of getting just about every serious player on the field to come to Stockholm; these people know that they need to talk to each other, and the IETF isn't the worst venue to meet in.
Anyway, after the dust had been left to settle for a couple of days, the opinions of concerned parties were converging around a set of ideas, more or less:
This time I had along 4 more UNINETT people; I hardly saw them all week. Once they report what they hear, I may think something about it; in the meantime, I shouldn't say.
There was an OSI harmonization BOF (disharmonious), discussion on multicast routing protocols' merits (inconclusive), discussions on key exchange protocols for the IP security stuff (biggest surprise: A real proposal from NSA, the National Security Agency, that nobody could see any back doors to), mobile IP work, site security handbooks, user glossaries, IPv6 address allocation (quarrelsome, it seems), and so on and so forth. The CIDR Deployment group thinks that IPv4 won't hit the roof on address space until some time early next century; the massive CIDR investment has indeed given us some breathing space!
Oh, and yes, there was ATM work done. Nobody knows yet if ATM will work in the real world, but if it works, we can run IP on top of it.
All in all, a Good Time Was Had By All.
See you in Dallas, Texas next time!