Internet Architecture Board


IAB Minutes 2021-09-01

Home»Documents»Minutes»Minutes 2021»IAB Minutes 2021-09-01

Minutes of the 2021-09-01 IAB Teleconference

1. Administrivia

1.1. Attendance

  • Jari Arkko
  • Lars Eggert (IETF Chair)
  • Wes Hardaker
  • Cullen Jennings
  • Mirja Kühlewind (IAB Chair)
  • Zhenbin Li
  • Cindy Morgan (IAB Executive Administrative Manager)
  • Tommy Pauly
  • Karen O’Donoghue (ISOC Liaison)
  • Colin Perkins (IRTF Chair)
  • David Schinazi
  • Amy Vezza (IETF Secretariat)
  • Martin Vigoureux (IESG Liaison)
  • Russ White
  • Jiankang Yao
  • Deborah Brungard
  • Ben Campbell
  • Jared Mauch
  • Daniel Migault
  • Greg Wood

1.2. Agenda bash and announcements

An executive session on the ICANN NomCom appointment was added to the agenda.

2. Technical Discussion: Consolidation and Discovery Document

Jari Arkko led the IAB on a discussion about the role of discovery in ensuring that Internet technology is broadly applicable.

In previous discussions, the IAB noted that centralization and consolidation are driven by many factors, including economic and network effects, but also by some aspects of the technology itself. Discovery is another area where technical work might have an effect.

Jari Arkko noted that in the specific recent case of encrypted DNS transport (DoT, DoH, DoQ), the main interest was in improving communications security. Discussions about DNS-as-a-filter requirements were polarized, and there was not much focus on deployment models or whether one trusts the serving entity. As a result, no discovery mechanisms were developed as part of the effort. The resulting technology fits extremely well in the deployment and business models for some parties (e.g. global DNS providers, browser vendors), but is much less useful for things like upgrading OS-to-ISP resolvers to use DoH. There were many debates about the trustworthiness of various parties in the DNS system, and later more work emerged on the technical means to trust servers better.

Jari Arkko suggested that ignoring deployment models can lead to results that do not provide as broad a service as they could. One take is that IETF chose to serve the needs of the big parties in the Internet industry over some other needs; an alternative interpretation is that IETF does solutions for those who take things forward.

Tommy Pauly asked whether it is that the IETF chose to serve the big parties, or that they chose not to serve the smaller parties.

Cullen Jennings said that it is difficult to talk about the IETF having made an explicit decision; the work is done by the people who show up.

Colin Perkins replied that the IETF does not do a very good job at getting input from parties outside the IETF community.

Russ White noted that the bigger parties are more likely to have the money to spend to send people to the IETF.

Jari Arkko said that different potential goals often get mixed up and have conflicting requirements, e.g. setting up a connection to a specific service versus upgrading a connection to better security versus setting up connections to a class of things.

There are underlying questions about who is in control:

  • Is something decided by an OS, ISP, or an app?
  • Who decides that a particular service is trustworthy?
  • Who gets to own the data about what users do?

Tommy Pauly asked whether the motivation should be to make it impossible to use any data that is collected.

Jari Arkko replied that that’s one of his big motivations.

Russ While observed that another aspect of this problem is that services that have more data tend to get more use, which leads to even more data being collected.

Mirja Kühlewind noted that the users are generally not aware of who should be in control, and it is difficult to present the information in a way that allows for meaningful input from the user. Website banners about cookie preferences are ostensibly intended to give the user control, but are often so confusing that it makes the situation worse.

Jiankang Yao said that Internet protocols are mainly about the client and server structure. Whoever controls the server controls the data.

Wes Hardaker said that there is a subtle nuance in that running code is easier to do when it is less secure or more simple, and by demanding that all implementations have running code, the IETF may end up producing standards that do not have the best architecture. Organizations with more money are more able to implement things.

Jari Arkko will update the draft document on the role of discovery in ensuring that Internet technology is broadly applicable based on the feedback received during this conversation.

3. Executive Session: ICANN NomCom Appointment

The IAB discussed the IETF appointment to the ICANN NomCom in an executive session. Wes Hardaker, Tommy Pauly, and Jari Arkko will interview the candidates.