After four years at the head of the premier standards-setting body, Harald Alvestrand talks about his tenure at the IETF and his plans for the future
Cisco Fellow Harald Alvestrand is stepping down in March as chair of the Internet Engineering Task Force, the Internet's premier standards-setting body. Alvestrand led the IETF for four years as the all-volunteer group of network engineers reeled from overwhelming workloads that stretched from the height of the Internet bubble through the corporate bankruptcies, unemployment and slashed travel budgets at the depths of the dot-com bust. Network World Senior Editor Carolyn Duffy Marsan spoke with Alvestrand about his tenure at the IETF and his plans for the future.
What progress has the IETF made in improving the timeliness of its standards development work?
We've cut down on the number of [documents] stuck for dumb reasons. The RFC Editor has a backlog of several months. They've suggested hiring a new person to cut down the queue. That will probably happen next year. We're experimenting with changing the way objections are handled at the document-approval level. We're giving the working-group chairs more responsibility here. We also have shut down 11 working groups since August primarily because their work was done.
What will be the legacy of your four-year stint as IETF chair?
If the restructuring [of our administrative processes via cooperation with the Internet Society] is successful, that will be my legacy. What I hope is that my successor will be remembered as being the chair when the IETF ran more effectively and had a greater impact on the world.
What are the most important protocols developed at the IETF during your tenure?
SIP and iSCSI because they both created emerging markets. [Editor's note: The Session Initiation Protocol supports real-time communications and is the foundation on which the IETF is creating voice, video and instant-messaging applications. ISCSI enables universal access to storage devices and storage-area networks over standard Ethernet-based TCP/IP networks.]
I notice you didn't mention IPv6, the long-anticipated upgrade to IPv4.
IPv6 was a done deal before I came onboard. It's not something that I consider to have influenced. But I do enjoy seeing it moving from an experiment to being a common checkbox for network products. It does have further to go.
In several high-profile areas - most notably instant messaging and spam - the network industry turned to the IETF for a technical solution, and the IETF failed to deliver.
The IETF has gotten its act together recently on an instant messaging and presence protocol. We already had made progress on [a SIP-based approach]. And it seems that the hype level has gone down on instant messaging. Everyone at the IETF meetings uses Jabber. But you can't talk between Jabber and MSN because MSN sees value in keeping a closed network. Instant messaging is just like IPv6. Just because we have specifications doesn't mean the market will adopt them.
What about spam? Many people in the network industry hoped the IETF would come up with a technical solution to this problem.
I was at the Federal Trade Commission hearing on spam [this month], and the conclusion was that there is no silver bullet. Technology is one component, but we have to try multiple approaches. Spam is different than all but our security protocols in that you're not facing random chance, you're facing an intelligent attack. The question is how will spammers change their behavior when a technical solution is released. Technical solutions have been oversold. We have many [protocol] drafts that have come in related to spam. We closed [one spam-related working group] because we couldn't get consensus, but we will open others if we see enough interest.
What will you do when you step down as IETF chair?
I will still be a Cisco fellow, so I'm sure I'll find something interesting to do. I move back to Norway on July 4. I'm glad I came here to California for a year, but I'm glad to go back.
With all its problems, is the IETF still significant?
Yes. It's relevant because it makes the standards that people use. That this many people still come to our meetings [1,300 attended this month] shows that they're getting benefit out of it. The Internet revolution is over, and we have gone from the Internet being new and exciting to it just being part of the infrastructure. The technical work required for making the Internet work is still being done here.
What is the status of the IETF's multi-year effort to restructure its administrative processes?
We had a couple of specific proposals, and we had a community discussion about them in September. We received strong feedback that we should ask the Internet Society to organize an activity to be the contracts manager for the IETF. Then they would sign contracts for the IETF for administrative functions. We need to be one step removed from awarding contracts because the IETF has not declared itself a legal entity. [On Nov. 10], I officially declared consensus on one plan. We [soon] should have a transition team in place that will oversee the process and hire a person to be our administrative director. By March, when I step down, I'm hoping this process will be done.
Can the IETF fund and run its thrice-yearly meetings in the interim?
Yes. CNRI will continue to support the IETF until the IETF decides what it wants to do. [Editor's note: The Corporation for National Research Initiatives is a nonprofit organization run by Internet pioneer Robert Kahn.] Right now, we have no contract with CNRI, just a gentleman's agreement going back many years. All our meeting fees go to CNRI, which contracts out with Foretec Seminars [a for-profit event planning company]. We get about $2 million a year in meeting fees that we use to run the RFC Editor and other functions. We want to put that under a common control model and make the relationships ordinary. That new process should be normalized by mid-2005. Then how we do contracts - whether there is rebidding or not - will be ordinary business practice.
Several high-profile people resigned from the IETF leadership during your tenure, saying the workload was too great. Is the IETF having more success at attracting volunteers for leadership positions?
We are hoping the new administration structure will help. Part of my workload has been dealing with the IETF secretariat. By moving those functions away from the IETF chair, I'm hoping that my successor won't have to deal with so much administrative work. As far as the nominating committee's process is concerned, people are willing to take on the job. They don't seem to have trouble finding candidates.
Four years ago, you thought the Internet's biggest challenge was scaling. What do you think it is today?
Scaling is less of an issue than it was at that time because we've built faster routers to keep up with the growth, and we've had industry consolidation. There are 600 million people on the Internet. What that means is that 5 billion people are not on the Internet. The whole push toward mobility is going to drive usage. The biggest challenges the IETF faces now are mobility, zero configuration (so that devices are useful and secure out of the box) and scaling.
How has the IETF changed in four years?
We had 2,900 people at our San Diego meeting in December of 2000. We had 1,300 at our last meeting in November 2004. One of the things that has gone away is the expectation of growth. To that I say "good riddance." In 2000, we were still working in the Internet bubble, where people had five new ideas per minute and no one cared about business plans. Now we're taking something that works and making it work a little better, like with iSCSI.
Were there any disappointments for you as IETF chair? Work that you wanted to get done and didn't?
I feel like I should have started the administrative restructuring process sooner. I didn't get around to making the technology plans that I thought I would.
Learn more about this topicIETF's quest to be quicker moves slowly
Network World, 11/17/03IETF begins self-assessment effort
Network World Fusion, 03/21/03Standards bodies meet to coordinate efforts
IDG News Service, 07/28/03