When: Thursday, March 24, 2022, 13:00-14:00 UTC
Where: Park Suite 9 & Meetecho
Chairs: Mirja Kühlewind, David Schinazi, Jared Mauch
Slides: Chairs’ Slides
Newly Published: RFC 9170 (was draft-iab-use-it-or-lose-it): Long-term Viability of Protocol Extension Mechanisms
Newly adopted: draft-iab-path-signals-collaboration: Considerations on Application - Network Collaboration Using Path Signals
Ready for Approval: draft-iab-rfcefdp-rfced-model-12
draft-iab-mnqeu-report: IAB workshop report: Measuring Network Quality for End-Users
New: draft-iab-aid-workshop-00: Report from the IAB Workshop on Analyzing IETF Data (AID), 2021
Slides: Report Analyzing IETF Data (AID) Workshop
Eliot Lear: Thanks for the update on the workshop. I saw a lot about what transpired at the workshop and what the next steps are. Can you provide your top 2-3 insights about what you learned about how we participate or what misconceptions people might have?
Niels ten Oever: Yes, but not with a full answer. People came in with different research interests, and we tried to understand the problem and how things can be approached. By no means a full understanding of the particular issues and I wouldn’t want to talk for the other participants. But, some of the code is being used to try to generate dashboard-like features and more of a web interface to do data science on mailing lists and standards. It sparked a lot of interesting things but standards are not.
Lars Eggert: This was very much a workshop, getting people together to understand the interests and motivations. I expect this will be an ongoing thing in some way, but it’s certainly not done after one event. We are trying to get some insight into what we can do in efforts like this.
Shane Kerr: This is very cool. I wonder if during the work there was attention given to areas where it was difficult to extract information from the data we had?
Niels ten Oever: Datatracker: very nice. Mailing lists: very hard. But many other organizations also have mailing lists. We can see if we can enhance the mailing list data with the Datatacker data. It continues to be work. There were challenges, but it was nice to have Robert and others there so it felt like we were working on a common problem.
Christian Huitema: Viet, I have been following your work for some time and I would like to see it published. The one thing I would like to be clear about is whether we should measure centralization by itself or as part of the web delivery process. When I do my own measurement, I find that the top sites, one of the biggest providers is Wikipedia, because it has an infrastructure that is in many countries. Same for Facebook and Google and others. Hard to compete with the various CDNs in some kind of closed market. Should we also assess the importance of stuff like that? How do we define these things?
Mirja Kühlewind: That is an open question for the community.
Dominique Lazanski: Mark McFadden, Eliot and I also have a draft on this, and maybe we could look at our draft and Mark Nottingham’s together and see where there are holes. The other thing I wanted to bring up more generally is that Stephen in SAAG mentioned consolidation around 2-3 protocols, and I don’t think that is something that has been looked at here in the IETF. Thanks a lot.
Eliot Lear: I think my name has fallen off that draft, Dominique. The other point I want make is channeling Cullen, in there was a workshop last year in DINRG on centralization in the Internet. I will post the link in chat, and it’s probably good for anyone who was interested in the topic to take a look at those. And while Mark and I have been talking about taking this to the Independent Stream, I hope that will not be the end of the conversation. There are many places, I think we should reach across to form better shared understanding.
Vittorio Bertola: I just wanted to suggest that the IAB take a look at one of the key issues of this centralization stuff is the relationship between applications and the network. This is really the key point, in the end by taking functionality that was originally provided by the network, and now these functions are being sent to 4 browser makers. Globally this is one of the reasons we see centralization. There should be an IETF position on what we see as the relationship between the network and the applications in the long-term.
Mirja Kühlewind: The IAB is working on a document here.
Phillip Hallam-Baker: Centralization is the frame we received, but it may not be the right frame to look at this problem. I worry about walled gardens and switching costs. On the social media side, I am not so worried about the publication side of things; I am worried about the curation side of things. It was the curation side of Facebook that was used to promote an agenda. I think we need to be skeptical of the frames we were given and address the real frame. We don’t do economics here, but we do do interpretability. Have a decentralized curation capability. RSS was originally designed for that.
Mirja Kühlewind: As a reminder, this is the start of the discussions, not the end. Have a good rest of the week!