IABopen at IETF-109 online

When: Tuesday, November 17, 2020, 14:30-15:30 UTC+07 (Bangkok time)

Chairs: Mirja Kühlewind, Wes Hardaker

Note taker: Cindy Morgan

Welcome and Status Update (10 mins) - Wes and Mirja

Slides: Welcome and Status Update

Spencer Dawkins: Two things I said on the mailing list, one is that calling these "administrative support groups" usually "administration" means something else, so I suggest "oversight" The other thing was, I thought one of the things with technical programs was they were more headed to open membership, where the oversight ones have needed to have closed membership. Including that in the description.

Mirja Kühlewind: We want to be transparent about this, but it is true that the ASGs sometimes need confidentiality.

Dave Thaler: You're saying that there are no long-lived technical programs because the IAB is deciding there are no long-lived architectural issues. Is there any long-lived architectural issues.

Wes Hardaker: I'll take the blame; Stephen suggest we change this to "frequently short-lived" and I missed that. This is a framework, not a hard and fast rule. Things may not fit exactly in these lines.

Dave Thaler: If it's "typically short-lived", then that's fine.

COVID-19 Network Impacts Workshop (10 mins) - Jari

Slides: IAB Workshop: COVID-19 Network Impacts

Eliot Lear: Jari, I have one question that I hope you can delve into in the report. Who is this "we"? You talk about "we have limited information" who has the information, and who doesn't, and in what circumstances should that be shared, and with what authorization, because I think "we" have a lot of information, depending on who "we" is.

Jari Arkko: "We" can be a number of things, we can clarify that in the report. Also, in the more broader sense, society would benefit from understanding how well the Internet is doing in our country. It's a good question.

Mirja Kühlewind: It was discussed at the workshop, and it was good to see all the measurements at different angles.

Toerless Eckert: It would be good if there could be more ongoing consideration about this, a lot of questions about reliability of critical services. I think just limiting this to the Internet we see at home isn't very inclusive. We've seen in any type of catastrophes, the critical services providers might break down, which may or may not impact the Internet, 911 services. Would be good to figure out if this could become part of a larger process for the IAB.

Jari Arkko: I do agree we need more focus on this in the world.

Michael Richardson: You said there was a lot of conversation among operators, and I think the public needs to know the nature of that, and who is the "we" and are we getting the Internet we deserve. I think someone needs to write the story about some of the things that happened and the repairs, because I think the public needs to know. I think that the public needs to know a bit about what the operators were doing behind the scenes. Even if this means semi-fictionalized to protect the guilty (Chatham House Rules). A radio play podcast perhaps. Maybe Jari, through this workshop and RIPE/ARIN/Apricot/etc. we could collect some stories and then share them in some kind of dramatized way.

Stuart Cheshire: I wanted to agree with what Jari said. I think it would be really good if the IETF could work on defining network quality. Users go to speedtest.net and think it will tell them if Meetecho will work on them. I think it would be helpful to start some conversations about network quality. IPv6, AQM, ECN. Do they block multipath TCP. Many aspects of network reliability and quality that go beyond how many Mbps upload and download speed. If we could build some tool that people could run that gives them a dashboard about what services they're using and how they are working.

Evolvability, Deployability, & Maintainability (EDM) Program (10 mins) - Tommy

Slides: EDM Program

Internet Threat Model (model-t) Program (10 mins) - Stephen

Slides: Model-t technical programme

Open Mic (10 mins)

Mark McFadden: Encourage the IAB to think about when you get to the end of a Program, how you get an actionable result into constructive work in the IRTF or IETF. The doc on GitHub is explicit about what Programs don't do, but I hope you'll reconsider about what those Programs provide for the rest of the organization.

Wes Hardaker: I think that's an issue across the industry. If you have ideas, please bring them up on arch-d.

Mirja Kühlewind: Programs are to get work started and we've had trouble in the past knowing when they should close; it's not a sharp milestone point, but that is definitely something to consider.

Brian Trammell: Well done on the COVID workshop. There was some discussion at the end that I would encourage the IAB to think about. The Internet worked well for the things the Internet works well for. Anecdotes are saying that the limiting factor that this network had isn't the bandwidth in most markets. The limiting factor has been room. If you have a family of 5 and they are all expected to be online at the same time, a lot of houses don't have 5 doors you can close. Or devices -- kids don't need tablets until they do. From an Internet perspective, we would say this is a success. I am not sure how to turn this into a thing the IAB should consider that isn't a rabbit hole, but I think considering what success for taking everything remote, beyond just the pushing packets part of it. I would suggest that the IAB think about that in possibly a follow-on workshop.

Jari Arkko: Thank you, that was a good comment. It is true that this is much broader than the narrow technical things we like to think of.

Brian Trammell: There was no congestion from people who weren't connected to the Internet. Possibly figuring out how to use the IAB in its capacity as a technical advisory body to the Internet Society.

Michael Richardson: What Brian said about not having enough doors to close, people have devices and they don't always want to use the same ones, and they tend to be tied to one user, and I think there is a question of identity and data and how to move it from device to device. We see this in IOT as well. Need to get configurations on and off devices. A bunch of work there that is end-user focused but is essentially bits on the wire. Needs to be done right.

Wes Hardaker: I agree with you completely. It's interesting that there's a lot of comments about edge networks. We've concentrated for a long time on the core infrastructure. The COIN RG is also dabbling in this space.