IESG Retreat


An IETF meeting is a busy time for the Area Directors. We do not have much time for discussing IETF-wide topics or getting to know new team members. Every year we meet for a couple of days as a team, outside the IETF meeting. This year we met in London, partially interleaved with the IAB meeting that also took place in the same location.

This article is a summary of the discussions at the retreat. The main topics were how the IESG could work better, how to make sure the IETF community is working most effectively, how and when to introduce the new RFC format, and the progress so far on addressing pervasive monitoring. And, there were a number of other important topics discussed about how to make sure the IETF is working well today, and into the future.

We spend the retreats not so much with the day-to-day document or working group issues, but rather try to ask higher-level questions about our organisation and how we work at the IETF. This year we started off with the ongoing re-organisation of areas. With the new ADs in place, we are now ready to formally switch to the new area structure. The biggest change is the creation of the Applications and Real-Time (ART) Area with three ADs. Revision of BCP 67 defining the DISPATCH process will be submitted, desirable qualifications for next year’s IESG selections for the nomcom are being worked on, and IETF tools are ready for the new area. I expect to see a formal decision for the creation of the ART area in the next IESG telechat. The primary reason for the creation of this new area is that it allows us to be more flexible in taking care of rapidly changing topics in this domain, without being too constrained by the area structure.

But that change is just organisation. It is good to have organisation match the kind of work we do, but there are even more important things. For instance, ensuring that we work in the most efficient way, and that work happens where it most naturally fits in. The IESG has discussed “smallerising” its workload. The idea is that if AD tasks were not as full-time as they are today, good things would follow. For instance, many more people could consider taking on the AD role. But the crux in work load reduction is that AD’s work consists of many small things that together take a lot of time. It is difficult to find a single task with significant savings potential. Nevertheless, several ADs are planning to run an experiment in their areas around document review tasks.

But how the IETF community as a whole works is even more important than the IESG’s workload. The community is where the real work happens! One of the trends that we see is a more prominent role of open source efforts. The IESG noted the early good experiences from the Hackathon, and we plan to continue and grow these events. We also noted good experiences in focusing IETF work related to data models better. A lot already happens with data models at the IETF, but there’s plenty to do to make open source and IETF worlds work even better together. As an example, at the Hackathon, Benoit Claise worked on tools to integrate checking and I-D generation of YANG models. When data models can be easily moved to different formats or checked against each other, producing high-quality models becomes easier for everybody. We are spreading the word about the upcoming Hackathon in Prague, planning the next year, considering whether it would be possible to invite students from local universities to these events, and so on. If you have further ideas around the Hackathon, let us know. And do sign up, as we will still have limited number of seats in Prague for this.

Similarly, in the past year, the IESG has discussed how the increasingly common document repository collaboration model with Github and other popular tools has been successfully used in several different Working Groups. At the retreat, the IESG discussed both having a wiki that contains descriptions of tools that have been successfully used and the details of how to set them up properly for a Working Group so that WG chairs can easily investigate using one. As part of this, the IESG discussed the work necessary to clarify how to apply the IETF Note Well and various IPR notices to such tools.

Heather Flanagan, the RFC Series Editor, also participated our retreat. Her focus has been the introduction of new RFC formats. The question for the IESG is determining the how and when to introduce those formats into the IETF process. We’ve already asked our tools team to do the first easy step of accepting drafts in XML format and not requiring a text version as well; of course, a text version will always remain acceptable. Two advantages from the new RFC formats are the ability to accurately include authors’ names, regardless of character set, and the ability to include diagrams in SVG. We’re planning to take the first Internet Draft with non-ASCII author names through the process soon, and proceed to diagrams once that has been successfully completed. Of course, there are still a few questions around determining which display versions of a draft are reviewed when and how to tell if there are substantive differences between the different display versions. Similarly, there will be additional tools work needed; for instance, the ability to compare XML and text versions, if both are submitted, and verify that the differences aren’t an issue (e.g., copyright changes) still lies ahead.

One of the actual technical topics in the retreat was a review of where are we with pervasive monitoring and defences against it, two years after the Snowden revelations. While improving Internet security is hard and slow, a significant amount of work has gone on in this space during those two years. Our first technical plenary on the topic, the STRINT workshop, RFC 7258, continuation of our work on improved and more easily usable security in TLS 1.3 and HTTP/2, chartering the UTA, TCPINC, and DPRIVE working groups, deprecation of RC4 and SSLv3, the introduction of ChaCha20 and Poly1035, the RFC on opportunistic security, the new IAB program on privacy and security, and countless examples of working groups carefully analysing their technology in light of possible privacy dangers. And at the same time, the world seems to care more and more about security, and is employing secure protocols in an increasing fashion. Pervasive monitoring continues to be a significant concern, and communications security is just a part of the defences. But the IETF is clearly committed to doing what it can to improve the technology.

We also talked with Greg Wood, ISOC public relations expert, about information that different interested parties need from the IETF, in view of the Website renewal project and other efforts.

The IESG responded to concerns that the datatracker sent too many e-mails to too many people on too many draft state transitions, by working with Robert Sparks and the tools team in reviewing what draft state events should lead to e-mails being sent, and where they should be sent.

Ray Pelletier talked about services for remote attendees, the anticipated growth of this service, and need to integrate effectively and efficiently into the participation of the WG sessions.

We also talk about IETF educational efforts. The EDU team has told us that they need new members and that this is also a good time to re-assess how the team works and what are the highest priorities for the work. Martin Stiemerling will be participating the EDU team from the IESG, but again it would be good to have volunteers from the community! We are planning to run a session at IETF 93 about this so that the community can weigh in on priorities and we can brainstorm about the best ways of running IETF EDU sessions. My personal thought is that the world is changing and maybe our focus should be on building a great YouTube library of IETF educational materials. This would fit a perhaps more growing remote participation, and enable people to learn about the IETF in more targeted ways, say, for someone who needs to come in to develop a specific piece of technology. This would imply getting excellent tutorials shot once rather than necessarily run repeatedly. It would also imply more work on organising the library of information and access to it. If you have any thoughts on this, let us know!

Finally, the IESG discussed situations relating to authorship of documents. The RFC Series Editor has made a statement about expectations relating to authors of documents. An additional case that we discussed in the IESG was that we’ve seen a few cases where it has been a surprise for people that they have been added as co-authors in documents. Please work with your co-authors to ensure that everyone is on board with being an author (and aware of all the responsibilities involved in being an author). An IESG statement on this topic will appear soon.

Overall, this was a very useful meeting for the IESG, and will be followed by a series of new arrangements. In London, we were hosted by Ted Hardie and Google, and I would like to thank them for their hospitality!

Jari Arkko, IETF Chair