This is my very personal notes from participation in the IETF meeting in Maastrict, July 2010. I am Jacob Palme, professor emeritus at Stockholm university.
Much was the same as previous IETF meetings. Lots of intense discussions about minute technical details of standards which are, or may become, central for the Internet of the future.
I have not been active in IETF work for the last ten years, but I succeeded in getting a contribution, paying for my attendance in one meeting, so I went to this meeting, mainly in order to get a feeling of what is happening, how IETF is developing. Surprisingly most was siimilar to ten years ago. The meeting rules, organisation, issues. The most radical was that they were discussing to abolish "draft standard" since so few standards move to any higher level, and "draft standard" is not a good name for the final stage of most standards.
They said that they intended to appoint an Applicatons Area Director, and asked for proposals.
All which is happening in IETF can be found in a calendar, which you can search using the data tracker.
There are three BOFs in this IETF meeting belonging to the applications area.
- Web applications security (not transport security). Coordinate with W3C which has similar activities.
- Persistent identifiers, used for example in national libraries, which handle documents to be kept for a log tine, in particular ISDN, which has increased from ten to thirteen digits. Existing documents must be aligned with internet standards.
- FTP X2 - surprisingly, there is still work going on to extend FTP.
Patrick Fältström said that many applications reinvent the same thing, and a BOF will be formed to discuss how to reduce this, on wednesday morning at 8:00 AM.
There needs to be a processs on how to do when an RFC gets obsoleted. In such case, something else may need to be triggerred.
11 documents to be produced. 235 issues on issue list, about 30 design issues, 10-11 hard issues. Has been working for several years, and now wants to get ready.
Tries to reflect reality in the new standard, not bother about problems not problematic in any known implementations.
CONNECT is specified in much more detail than before. UPGRADE infrastructure defined better.
Most parts of 2817 are not implemented by any active implementations..So 2817 might be obsoleted, moving the only thing used to other parts.
BASIC and DIGEST into two separatre documents.
Internationalization important but might require a change to the charter of the wg.
Not very many people in this meeting, I wonder if there will be any work started in this area.
Show of hands: Three people will implement or write paper in this area.
Making FTP work over IPv6.
If every client used EPSV there would be no problem.
Servers should ignore the IP address in protocol, and use the IP address provided by the control channel.
Three laptops and projectors were stolen from the premises. Participants are warned to keep their equipment supervised.
22 people present. Very few people on this, as in many subgroup meetings.
Millions of people are using URNs, we were told at the start of the meeting. National libraries are the largest users. URNs are persistent identifier systems (PIDs).The naational libraries have huge digital collections to be kept "forever". The national libraries are legally obliged to keep these collections. Multible parallel URN independent implementations, so it can be put on the standards track in IETF..
Another usage area is XML name spaces. There are 40 registered name spaces. One more example is SIM cards.
Examples of library usages: Every digitized journal gets a URN. Every articles within it gets a URN. Every imagewithin each artciles ges a URN.
This working group seems to have been doing a lot of work already, the documents seemed have had much work already done on them..
Abuse types are: spam, phishing, spoofing, fake sender address, unathorised sender/recipeint, suspicious network/domain, message flooding, denial-of-service attack, malware.
Spam reporting formats for email, sms, mms.
There seems to be competing standards. Need for automatic translation betrween the different formats.
The mail standards are draft standards, should they be forwarded to full standards? But IETF will maybe abolish draft-standards, and move all draft-standards to full standards.
Goal is to have high HTTP compliance.
Initial length may not be known, but chunking will be used with an explicit length of each chunk.
Should we allow subprotocols, other protcols embeeded within HTTP?