By: Andrew Sullivan
Date: March 19, 2017
IETF 97 was my last meeting as Internet Architecture Board (IAB) chair and IETF 98 will mark the end of my time on the IAB. From the start of my appointment, I hoped to achieve three things. The first was to strengthen the IAB’s programs in relation to Internet architecture. The second was to sort out the IAB’s role in the IETF’s relationship with other organizations—what I sometimes describe as being the “foreign office” of the IETF. The third was to try to root out traces of the “great person” theory of IETF leadership. Now is a chance to reflect on how well that all worked.
Architectural Oversight and Programs
The IAB organizes its long-term work into programs. Since the IAB is unlikely to have experts on everything, programs give the IAB a way to call on outside expertise when the organization needs it. In addition, programs enable an IAB member to engage with work that the IAB starts, even if that work might extend past the end of the member’s term. While not every program is about architecture, the IAB should organize its architectural efforts into programs.
Perhaps inevitably, architectural programs have sometimes worked better in theory than in practice. In my view they work well when there is at least one and preferably more than one IAB member highly engaged with the work of the program. In that case, it can be an effective amplifier of that interest and a way to get the IAB to contribute something useful to architectural discussion. But the IAB can sometimes keep alive a program that is not really working. Often, this is because the IAB recognizes that the topic is one that has big implications for the IETF, but either does not have any members with an abiding interest in the topic or does not have enough participants with enough time to engage with the program’s needs.
Over the course of my tenure, we did make some improvements in this area. We managed to close or reconstitute some programs that were no longer producing results. We increased the frequency of program reviews, and attempted to ensure that program leads were fully engaged by asking them to produce topics for the IETF technical plenary or Birds-of-a-Feathers (BoFs). But some programs floundered, and the floundering ones appear to be the ones least likely to schedule themselves for review.
Setting up a new program if there is energy is not hard. And, unlike IETF Working Groups, there is no particular procedural advantage to keeping a program around to “tidy up”. This suggests that the IAB might be better served by aggressive closing of programs on the principle that it is bad for the IETF to offer to do work and not complete it.
The IETF Foreign Office
When I was appointed to the IAB, I thought we needed to tease apart the roles of the IAB and the Internet Engineering Steering Group (IESG). I believed the increasing external profile of the IESG and the IETF chair was risky for the IETF because of the role of the IESG in declaring IETF consensus. We have a hard time explaining rough consensus anyway, and there is some danger that people will mistake “IETF leadership” for people who are in control.
For better or worse, the Internet Assigned Numbers Authority (IANA) stewardship transition that happened during my IAB tenure gave me immediate exposure to the relationship between outside organizations and the IETF. It changed my mind about the “foreign office” approach. However we might like to organize ourselves and our work, most organizations are used to dealing with one another through the leadership. So, they want to talk to our leadership, even though our leaders are not really in control. We can either spend a lot of energy trying to change the way others understand us, or we can ignore the mismatch and try to achieve our more important goals. I have come around to the view that the second is more valuable.
The Community Is the Leadership
If our relationship to other organizations needs to conform to convention, then we must ourselves ensure that we do not allow that conventional mode of thinking to undermine our own ways of working. It is important that our leadership does not misunderstand itself as having control.
One way of promoting that kind of thinking is to try to reduce the importance of the IAB chair and instead spread that work around. We have tried to do that. Communications from the IAB do not always come from the chair, but instead are sent by whichever IAB member leads the work in question. We separated from the chair some tasks that had previously come with the job, such as that of stream manager for the IAB’s RFC stream. And, of course, the continued emphasis on programs means that there are more opportunities for the IAB to reflect views from outside itself.
There is still more to do here, however. Most worrisome to me is the linking of membership in the IETF Administrative Oversight Committee (IAOC) and the IETF Trust with the position of IAB chair. Not everyone who is chosen for IAB chair is likely to be the best candidate to work on issues for the IAOC or the IETF Trust. Because of the IANA stewardship transition, the Trust is more important than it used to be, and has a greater outward role than it originally had. I hope and trust that the “IASA 2.0” effort that the IETF has started will, among other things, permit greater flexibility in how those roles are filled.
In closing, I thank those who have been my colleagues on the IAB, and especially those who put their confidence in me by selecting me as their chair. I also thank my employer, Dyn, particularly for its steadfast support during the period when the IANA stewardship transition took much more time and work than forecast. And I thank the community for recommending me to the Nomcom for appointment and for the good counsel I received so often during my tenure. I am honoured to have served; I hope I have served you well.