• United States

IETF: Not a teenager anymore

Mar 27, 20063 mins

By the time you read this the 65th IETF meeting in Dallas will be over. This meeting represents a milestone, as the IETF is now more than 20 years old. The IETF has been the major player in almost all important – and quite a few not so important – Internet and IP network technologies for the past two decades. Not bad for an activity that has never had any legal existence.

The first IETF meeting was held in San Diego on Jan. 16 and 17, 1986. Twenty-one people attended, four of whom are still very active in the task force. That meeting focused on some topics – including routing and QoS – that were discussed by 1,200 or so people during this year’s Dallas meeting.

The Internet has come a long way since that first IETF meeting. In 1986 the Internet did not exist for most of the population. At the time it was made up of two backbone networks, mostly for government research – the then-17-year-old ARPANet and the recently created National Science Foundation Network (NSFNet) – as well as a handful of regional research networks created along with the NSFNet.

For the most part, access to these networks was limited to researchers receiving federal funds. The number of Internet hosts would not pass 10,000 for another year. Today there are thousands of ISPs and there are IP networks in millions of enterprises and residences worldwide. And there are more than 350 million Internet hosts and close to 1 billion Internet users.

Since 1986 the IETF has developed or maintained all the core Internet protocols running “above the wire and below the application” (in the words of an old description of the task force’s role). Attendees at the first IETF meeting were mostly from academic and research institutions. The attendees in Dallas this year will be mostly from corporations involved in Internet and IP network equipment or services.

For at least the last decade some pundits have been saying the best days of the IETF are over and the important new standardization activity will take place somewhere else. So far, these predictions of irrelevancy have proved false, because the IETF continues to develop technologies that become widely adopted in IP networks everywhere. Among the important technologies from the last decade are Enum, iCal, IPv6, iSCSI, Multi-purpose Internet Mail Extensions, MPLS and Session Initiation Protocol.

There are numerous protocols under development, as can be seen from the Dallas agenda. Not all will wind up being accepted by the marketplace, but I expect many of them will be.

There are a few key features that make the IETF as important as it has been over the years. One is its mode of operating mostly on mailing lists; another is its openness. Anyone can participate in the IETF standardization process by joining a mailing list; there is no fee or membership agreement. Join, read the documents, which are open to all on the Web, and start participating. You do not have to spend money attending the face-to-face-meetings in order to participate effectively in standards development.

Time will tell if the IETF will be supplanted by other groups in the Internet standards biz, but there is currently no specific contender for the role.

Disclaimer: Some pundits have claimed Harvard’s best days are behind it; I did not ask the university for an opinion on that idea or about the IETF, so the above birthday report is mine alone.