**Session Date/Time:** 19 Mar 2026 06:00

This is a verbatim transcript of the Research and Analysis of Standardization Processes (RASPRG) session at IETF 119.

**00:00 - 02:15 [Chairs' setup and initial conversation in Spanish]**

**Alvaro Retana:** ¿No tenemos a nadie de MeetEcho por aquí, no?

**Ignacio Baguena:** Sí, voy a preguntar. Eh, Simón. Simón.

**Alvaro Retana:** Eh, les pregunté, pero si quieres empezamos y después vemos a ver qué.

**Ignacio Baguena:** Empezamos pues? ¿Quieres que lo lea yo si funciona bien el sonido?

**Alvaro Retana:** Sí, el sonido funciona bien.

**Ignacio Baguena:** Okay. Well. 

**02:16 [Official session start]**

**Ignacio Baguena:** Hello everybody and welcome to RASPRG, whether you are locally, remote, or in the nearby remote Tokyo. So let's start with the Note Well. The IRTF follows the IETF Intellectual Property Rights disclosure rules. By participating into the IRTF, you agree to the IRTF processes and policies that you can see on the slide. These include being aware that every contribution is covered by patents or patent applications that are owned or controlled by you or your sponsor and must be disclosed. That the IRTF expects you to file such IPR disclosures in a timely manner. The IRTF prefers that the most liberal licensing terms as possible are made available for IRTF stream documents and definitive information is in the relevant RFCs. 

Next slide. 

You should also be aware that the IRTF routinely makes recordings of online and in-person meetings including audio, video, photographs, and publishes these recordings online. If you are participating in person and choose not to wear the "do not photograph" lanyard, then you are consenting to appear. If you speak, you approach the microphone, you will be recorded. If you participate online and turn on your camera and/or microphone, then you will consent to appear in such recordings as well.

As a participant or attendee to any IRTF activity, you acknowledge that written, audio, video and photography recordings of the meeting will be, maybe made public. Personal information that you provide to IRTF will be handled in accordance with the privacy and policy of the IETF. As a participant or attendee, you agree to work respectfully with other participants and please contact the ombudsman if you have any questions or concerns on this regard. You have the code of conduct and the anti-harassment procedures in the respective RFCs as well as the IRTF code of conduct.

The IRTF focuses on long-term research issues related to the internet via the parallel organization IRTF to the IETF, and focuses on—the IETF focuses on short-term issues of engineering and standard making. The IRTF conducts research. It is not a standards development organization. And while the IRTF can publish informational or experimental documents in the form of RFCs, its primary goal is to promote development or research collaboration and teamwork in exploring research issues that are related to internet protocols, applications, architecture and technology. And you can have a primer on IRTF in the respective RFC.

And now going back to RASPRG, from the general staff of the IRTF. The RASPRG, Research and Analysis of Standardization Processes research group, works on understanding the standardization processes via evidence-based reproducible work. The main outputs are joint papers, tools, data, and open source software. And the goal is not to make hierarchical comparisons between SDOs or directly influence IETF operations. And the chairs are Alvaro and myself, Ignacio. And the charter is available in the IRTF page of RASPRG.

And please join the mailing list. The house rules: please scan the QR code, use the MeetEcho for questions and answers. The speakers please be ready. And I don't know if we have any volunteer to take notes of the meeting? Any volunteer in the room or away? 

**06:23**

**Alvaro Retana:** We just need one person to take the notes. 

**06:42**

**Rod Van Meter:** Hi, it's Rod Van Meter. I can—I can do it. I'm still getting logged into the datatracker, but I'll do it.

**Ignacio Baguena:** Thank you very much. Really appreciate it. 

**06:53**

**Ignacio Baguena:** Okay, so thank you for that. And let's now jump to the agenda. So first, we are going to have Noga Rottman from University College London, who is going to be talking about RFSeek, beyond ASCII art, making RFC protocol logic auditable with RFSeek. Then we will have—

**Alvaro Retana:** Actually, why don't we skip until we get the video ready because she is not here. 

**Ignacio Baguena:** Yep. Then we will have Priscilla from University of Sao Paulo, who's going to be talking about the Brazilian experience with global standards, local voices, mapping Latin American participation in the IETF meetings and within a project that she is part of in promoting participation of Latin Americans and in particular Brazilians. And then we're going to have another regional perspective from Mi-Jin Kim from Yonsei University, from standards adopter to standards author, the evolution of Korean national standardization system with a focus on ICT, which comes from a paper for the World Bank. 

**08:29**

**Alvaro Retana:** Priscilla.

**08:36 [Introduction for Global Standards, Local Voices]**

**Ignacio Baguena:** So while Priscilla gets to the microphone, let me make a brief introduction. Priscilla is a researcher in computer science working in the intersection of data analysis, visualization, and internet governance. She recently completed her PhD in the University of Sao Paulo in Brazil, and she's currently involved in studies of participation dynamics and representation in the internet standardization processes, with a particular interest in the IETF and IRTF, geographical diversity, and institutional engagement in internet governance. And without further ado, I hand over to you, Priscilla. 

[Priscilla's slide deck: [Global Standards, Local Voices: Mapping Latin American Participation in the IETF Meetings](https://datatracker.ietf.org/meeting/125/materials/slides-125-rasprg-global-standards-local-voices-mapping-latin-american-participation-in-the-ietf-meetings-01)]

**09:23**

**Priscilla:** Thank you. Thank you Ignacio for the introduction. And hi everyone. I'm Priscilla and I'm from Brazil. And today I will show you the results of a work that we have been working in the last few months. We are analyzing the Latin America and Caribbean participation in IETF. And I would like to start explaining why we are doing that. We know that IETF is an open and volunteer community, but open to all doesn't mean equals to all. We would like to know how Latin American participants have been contributing with IETF. And in this particular study, we are analyzing the attendance data from IETF. 

Well, talking about the attendance data, we know that all the IETF data is open to everybody to collect and analyze. But specifically about the attendance data, if you look for the recent works analyzing this data, you'll see a very short period of time because the recent meetings, we can get really easy the data, the name of the participant, the organization, the country, the type of registration. But if you go to the oldest meetings, this is kind of hard to collect, especially the first ones. I think that from meeting one to meeting 29, everything is in PDF, really hard to read to be honest. And also depend of the period of time, IETF collected different data from the participants. So sometimes we only have the name, other—in other meetings we have the name and their organization. For the oldest ones, we have email and phone numbers and fax that we don't use anymore. But we can—we did a complete process collecting, cleaning and pre-processing this data to in the end have the name, the organization, the country, and the type of participations. 

Okay, I would like to show you the big picture first, then we follow to see the Latin American participation and then I will show you specifically the Brazilian example of participation in IETF. So not surprisingly, if you look for the entire data of the participants in IETF, most of them, almost a half, are from US and Canada, followed by Europe and then Asia. And if you look to the south global, you'll see the lowest participation. But this is again in the attendance to the meetings, okay? Not the email or RFC or drafts. So, but you can see something similar if you analyze the other data from IETF. 

Well, this is the participation of all regions over time. The yellow line is US and Canada. We separate Mexico from North America just to analyze Mexico with the other countries in Latin America. And of course, you will see that the yellow line have the most number of participants. But removing US and Canada, we start to see some insights about the other regions and how they have been participating in the IETF. So we can see that in the red line is Europe, and they start to contribute really early. And the orange line is Asia. And in the bottom, you almost cannot see at the beginning, you can see the Latin America and Caribbean, Oceania and Africa. Let's look as an example, the Asia participation because this is kind of model to us. We can see that they start with a small number of participants and then they start to increase the number of attendees until that peak that is the meeting in Japan. So we can see that the meeting in Japan have a huge impact in the attendance from Asia, which is like not surprised, right? The meeting was in Japan, so people from Asia come to participate. But the thing is, after this meeting in Japan, we can see that—oh, there is something that I need to highlight here. From meeting 55 to 75, we couldn't infer the country of the participant—the most of the participants, okay? This was really hard to get. That's why you can see the number decreasing, but this is actually because we couldn't identify the country of the—of the participants. So after the meeting in Japan, we can see that Asia created kind of baseline of participants. So looks like these meetings in the region have some impact in the future participation. 

Okay, so let's look like to—let's look to Latin America and Caribbean participation. As you can see here, we have a very modest number of participants in the meetings. We start at the beginning, the first—the first attendance from Latin America was like three Brazilians in the meeting number 20. After that, Peru, three or four meetings after that. And we can see that we always have like less than three participants during in the meetings until around the meeting number 75, we start to increase the number of participants until we have this peak in the meeting 95, that was the meeting in Argentina, that we have like a huge number of participants from Latin America. And after that, we could like create a baseline, similar of—of the number of the participants like Asia did, but of course, our number is slower than the Asia. But same thing happened with us. After we have a meeting in our region, we kind of increase and create a baseline of participants, around 35 participants by meeting. 

Okay, let's look for the global institutional composition of IETF, and this is interesting to look for Latin America after. So it's not surprise this, looking for the organization from the participants in the meetings. This is the historical data, right? So Cisco has the most number of participants, followed by Huawei and Ericsson. So basically, the participants of IETF is most of them working for some technology company. And this is the global view. But let's look to Latin America view, and you will not see any company here because in Latin America, the participants are, most of the participants are not from technological companies. They are from universities or organizations like the—the internet register organizations in Brazil or Latin America. So if you look this, NIC.br is the organization in Brazil that register internet domains. LACNIC, UFAL is a university in Brazil and UTN is a university Argentina. So the—if you look the global picture, you will see most of the participants come from technological companies. And if you look for Latin America, most of the participants are from the academy or from one of the organization that internet—internet organizations in the region. 

Okay, so let's look the access and how the participants interact in the meetings. Because from the moment that IETF start to provide remotely access, that people could participate in a remote way, this make us—this give to us more—one more opportunity to participate in IETF. So as we can see here, people in Latin America from universities or even technological companies or government agents, in most of the—in almost a half of the participation, they are remote participations, okay? So this happened because in Latin America, most of these organizations, universities, we don't have budget enough to send our participant to onsite participation. So the remote participation is really important for increase the number of participation in our region. There is only one exception here. It's call "O" organization—it's not a good name. But it's the LACNIC and register organizations in Latin America. They usually send people to go onsite to the meetings. 

Okay, so I talk about Latin America, let's talk a little bit about Brazil. Brazil is by far the main actor when you talking about participation in IETF from Latin America. We have more than the double of the second country, that is Argentina. This mean—we can look to what Brazil is doing to increase the number of participation and maybe the other countries can follow our steps to do the same. And also I would like to see what other countries outside of Latin America doing so Brazil can also try some things. One fun fact about this is Peru was the second country in Latin America to send someone to IETF. But after that, they couldn't keep like—like Brazil did, increase the number of participation. In some moment, Peru start to not—not decrease, but they couldn't increase the number as Brazil or Argentina did. 

Okay, this is the participations in Brazil and you can see the shape is—is basically the same shape of the Latin America participation. And here the colors represent the type of registration. So you can see in orange, we have the remote participations and in green, we have the onsite participation. So as I say before, the remote participation option make us possible increase the number of Brazilian participants in the IETF meetings. And there is a reason for that, I will—I will talk about this later. 

Okay, this is the entire number of participations and this is only the onsite participation. So we start in the meeting 20, and then we start to grow—increase the number of participants from around the meeting 75, and you have this peak of attendance in the Argentina meeting. So what's going on here and why we increase this number? How we increase this number? You can see in blue and red, two initiative that make us increase this number. One of them is the ISOC fellowship that doesn't exist anymore, but a few numbers of Brazilians applied to the ISOC fellowship to participate of the meetings. And in red, you have the CGI program fellowship, that was a program that occurred like 10 years ago to provide financial support for Brazilian students and researchers went to IETF meetings. And as we can see, this was super important to make to increase the number of Brazilian participations. So the remote options and the fellowship are two of the main reasons because Brazil could improve their participation. 

Something really important about this fellowship is because we really have good results about it. One of the fellows from CGI fellowship in this time is now a chair in a research group. And two of the ISOC fellows in this period of time are responsible for the increase the number of remote participants in Brazil because two of them are from University of Alagoas and University of Pernambuco and they create hubs for the undergrad and grad students participate of the IETF. So in three—in these three meetings, these two ISOC fellows, they create these hubs for undergrad and grad students being involved with IETF. And this—this was really important for Brazilian community because we start to have more students interesting to participate and engage in IETF. 

Well, to finish the presentation, two lessons that you can get from this. One is the geographic proximity is really important. The fact that we have a meeting in our region shows that our people show up. So after—after the meeting in Japan, this make Asia create a baseline of the number of participants, and after the meeting in Argentina, of course we decrease the number of participants, but this make us create a baseline of participants here. Of course, there is any reflection of this in the drafts, RFCs and engagement in the main list? We are investigate this right now. But at least in the number of attendance, we can say that yes. 

And the second lesson here is that the combination of these three factors make Brazil increase the number of participants. So international—sorry, institutional leadership is really important. So NIC.br and LACNIC are very important organization to send people from Latin America to participate here. The second thing is strategic funding, because in Latin America, we don't have—we are not huge technology companies to provide the budget that the people need to come. So this type of fellowship is really important to help people from other parts of the globe, especially the south global, to participate in IETF. And the last thing is community access, and I particularly think that this is really important, that people from our region that come back to our countries do something with your community, like the two ISOC fellows that create hubs to students participate of IETF. This is really important. I mean, this year we are also organizing an event for young students participate and understand how IETF works. If anyone wants to collaborate with this, please talk with me later. But this is really important so we can create the new—new students interesting in that, so we can have more people interesting participate. 

And well, our work doesn't stop here. We also want to investigate some things that I kind common sense in IETF. For example, some people think that maybe have a meeting in a particular country or region will not impact the participation of this region or country. We would like to analyze that. I mean, we know that we have some impact in the participation in the meetings, but we don't know if this—there is any reflection of that in the engagement of mail list and in the RFC and draft authorship. So we—we would like to investigate that. We are doing this right now. The second thing is we are developing a visualizations tool so any organization or country that want to do the same analysis about how your country or your organization are contribute with IETF and without do any code or anything that we did here. So if anyone is interested in that, you can talk with me later. We almost have in the first version of this tool, and we'd like to have like people to—to test. And to finish, we are planning starting a study about technical consensus. This is something that I'm really interested, how a consensus is built inside of a working group, how a topic that is a huge disagreement in the—in the working group became a consensus over time. This is something that I really interested study, if someone is also interested can talk with me later. Thank you everybody.

**27:26 [Q&A for Global Standards, Local Voices]**

**Alvaro Retana:** Thank you so much, Priscilla. Please don't go. We have a queue of people. I have a lot of questions too, but I'll let the queue go first. Jean?

**27:46**

**Jean-Francois: [Jean-Francois (John) was not specifically introduced but is identified from earlier conversation and Q&A context]** Hi, Jean-Francois. So typically when I hear this kind of presentations—and this is no indictment against you or anything, it's just a general feeling that I have—I'm always concerned about where the consideration is in terms of how distributed is or how global is the participation in how do we make work the internet and why is it that the US or the West is always the one who's participating the most. And then I'm thinking, well sushi became global as well, do we expect not to be the Japanese who push forward with new sushi? As in, there's something to be said about the first one who arrived somewhere is probably going to be the one who's moving more forward. So the idea is that how you incorporate other things, but I'm not surprised in the list that the US and the West is still remaining the ones who move forward with what they invented. 

Um, in—there's another thing that I'm—that I'm always surprised with this kind of analysis and this is something that I see in all of them, okay? And I feel that there is, when we do the comparisons in terms of participation, never mind the fact that you pointed out that yes, of course the very likely the most representative analysis, the participation in the mailing list and the actual outcomes and it would be great to see the results that you have with that. It's also—I never see any incorporation of data such as GDP. As in, do those countries have enough internal revenue and internal capacity to generate an educational system that would provide enough engineers so that in the aggregate you can have enough people in to be able to send to other places? Or whether they have other more pressing concerns to develop their own country rather than to send people to other places for international standards? All of those things seem to be very valid elements of comparison, and if I want to see for instance why Peru did not send more people afterwards, maybe they had other things in their plate and it was just not a priority for the government and therefore they did not push to have more engineers. And so all of those things may also be interesting to add into the analysis, which I understand would be very, very difficult and complex to add on top of that. I'm feeling that there is more to the story than just the bare numbers of participation. Thank you.

**30:06**

**Priscilla:** Yeah. What's your name? 

**Jean-Francois:** John. 

**Priscilla:** Oh, hi. Yeah, I totally agree with you. It's obviously that US and Canada will have more people participating. It's pretty obvious. I think that the—I would like to put here the global picture just to look to how this is different from Latin American participation right now. But I totally agree with you with this. And also to about the second, about the educational system, I think this is because I put the second lesson here is how we can make our region or our country increase the—be more interesting in this and computer networks and the work of IETF, so what we can think and what we can do to increase this, especially because as you can see, we have some small results but still results. I mean, from the fellowships and people interested in the past, we have a today we have a chair, we have professors that came back and start to create a small community with their students. So it's something more that a Brazilian experience that I would like to share with other countries in the south global that maybe they are interested too and try to do the same, and maybe we can do this together. 

**31:24**

**Jean-Francois:** Yeah, I don't want to give the impression that I don't want more people to participate. My organization in fact helps Southeast Asia engineers to participate more in the IETF. So we are very much in the same, like it's just what I'm talking about the numbers, it doesn't always ring as something that is data that would affect my—the way that I want to proceed. And second, the collaboration that you were requesting, I would definitely want to talk to you after that for some things that we are also doing that might be overlapping. 

**Priscilla:** Oh yeah, totally agree. I will be happy after analyze the engagement of the mail list and the draft in RFC, I will be happy to present here again these other results. I didn't finish yet. Excellent. Thank you for the work.

**32:04**

**Sarah:** Thanks, Priscilla. Can I just check I'm coming through okay? 

**Alvaro Retana:** Yes, we can hear you, Sarah. 

**Sarah:** Amazing. So first thing to say, I—I massively commiserate on the issues with the data perhaps being slightly imperfect, particularly as someone who is Irish and then comes up in the registration data as being from the UK because that's where I live and work. And I understand that particularly the—the most recent data being self-reported might make some of this a little bit of a challenge. I was wondering if you might have thought about tracking some specific people whether they move country might be interesting. I imagine there are lots of people from Latin America who might start their career in one place and then move subsequently and then won't show up in the country data in that regard because they're working in the US or the UK, for instance. And then this is just more of a thought. I'd be really interested in seeing how this remote hubs experiment might be able to kind of map across to supporting some of this work. I'm currently in Tokyo. It's worked really well, it's been really successful in terms of kind of having people be able to take part in the meetings without necessarily having to be in the country. That's for slightly different reasons than in this case, but I think it's something to—to kind of think about and bear in mind when we're looking at encouraging participation. And then finally, really, really interested in what you mentioned about fellowships. Would love to reach out. It's something we're really interested in in UK government. So I will drop you an email. Thank you.

**33:29**

**Priscilla:** Yeah, thank you. I mean, we will be happy to share our data set with you, if you are interested. And about how to identify people from Latin America that are working for other organization in other countries, the answer is we cannot. We cannot do that. I mean, was really hard—was kind of good idea use the—for the oldest meetings, take the phone numbers to identify the countries and the emails, but this is the only that we can do. Thanks.

**34:06**

**Andrew:** Thank you. Hopefully you can hear me okay. Really interesting presentation, quite fascinating. And it does, in my opinion, underscore the importance of rotating—of the rotation of venues for capacity building purposes in the community. I think it'd be quite interesting—and I'm not suggesting do this—but if we were to sort of somehow overlay visa constraints into the data, I suspect that often times local proximity to wherever the meeting's taking place may well equate to easier entry requirements given we heard about that obviously yesterday in the plenary, some of the challenges of people getting visas to travel to places. So that may also sort of simplify participation for people within the region. And likewise, if the sort of relative time differences are not so big, which makes remote participation a lot easier. Shout out to anyone in Europe getting up at 1:00 AM this morning for the first session, for example, and I appreciate that happens a lot for people all over the world from time to time. But anyway, that aside, I think this is just another data sort of piece that tells us there's good reasons to move locations to improve diversity of participation and therefore hopefully get to better standards. And not a rhetorical point rather than a question to be answered now, but be fantastic to perhaps, say, initiate a conversation with ISOC to see if there's any appetite to reinstate that fellowship program as again a further method to sort of build a bit of momentum in target countries. But as I say, really interesting presentation. Thanks for doing the work. And hopefully the community will look at it and take note. So thank you.

**36:18**

**Priscilla:** Okay, thank you Andrew. This visa thing is really—is really something. But I mean, at least in our analysis about Latin America, even the meeting in Argentina, we have more people from Latin America being a remote participant than onsite participant. And most of them we don't need a visa to go to Argentina. So I think that the budget is for Latin America is the main reason. And then of course we have visa problems. 

**36:51**

**Rio:** Hi, could you hear me okay? 

**Priscilla:** Yeah. 

**Rio:** Okay. My name is Rio from University of Standards. First of all, thank you very much. This is a very interesting work and I think it's very important work. Could you go back to the slide that was looking at the sort of the annual sort of tracking of the participants in the earlier part of the slide, if that's possible? 

**Priscilla:** Which one? The global one or Latin American? 

**Rio:** Comparing different regions. Around here I think. Yeah, that's right, yeah. You mentioned there was a period that was quite lossy. Which period—could you remind me which period that was? 

**Priscilla:** Oh, yeah. It's the purple line here. The participants that we couldn't identify the country. Especially the meetings from 55 and 73, I guess, was kind of—most of the participants didn't provide any information about country. And when I say about country, I mean, an email or phone number that we can get the information of the country. 

**Rio:** Was there affiliation left in that by any chance? 

**Priscilla:** Sorry, can you repeat? 

**Rio:** Were there any affiliation data available in that period? 

**Priscilla:** Oh, yeah, but the thing is, if they put like Huawei or Cisco, from which country? I mean, there are some organization that we can know, but like most of them we can't. 

**Rio:** Excellent. Thank you. I just wanted to know which period it was just because I'm also looking at these kind of data, sort of time-tactually. 

**Priscilla:** Yeah. One thing that we think about it was like looking for the next meetings for this person to identify the country, but maybe this person change organization or change the country. So we didn't—what we did foi look for two the two previous meeting and two the next two meetings to try to figure out the country, but if this was not possible, we just put that we don't know the country. 

**Rio:** Yeah. No, no, no, totally understand. Thank you very much. 

**Priscilla:** Thank you. 

**39:26**

**Colin Perkins:** Hi, Colin Perkins. So I'd like to echo the comments, thank you, that's a really interesting talk. Two points, I guess. The first, especially for the recent years, the LLC has a lot more demographic data than is made public. And I don't know if it would be possible to get access to that, but if it is useful, you might want to have a conversation with people in the leadership to see if, under various restrictions, they might want to make that data available for research purposes. And I don't know if it's possible, but it might be worth having the conversation. The second point is that the—obviously the IETF has had remote attendance in various forms for many, many years, and it's been gradually improving in quality and features and so on over time. It would be interesting to know if there is something which is perhaps still causing problems with the remote attendance, or if there are features which have significantly improved the remote attendance over time. Partly to help the IETF improve the remote attendance features, but also if we can learn lessons from what is really helpful and what isn't, that can maybe then be helpful to other organizations or other SDOs that are trying to provide remote attendance to their meetings to help others learn from the lessons we have learned over many years. Thank you.

**40:59**

**Priscilla:** Yeah, thank you, Colin. It's a good suggestion. I'm going to do that, talk with LLC to see if they have more data that they can provide. And yeah, I think that IETF improved the remote participation can help people from the south global to participate. 

**41:19**

**Irina:** Hi, my name is Irina and I'm currently working for the IO Foundation. I'm the Indonesia associate. And what you present here is very interesting and informative as well. I probably just want to have a bit of discussion or maybe input. Sorry for my English. Um, your data here is very beneficial to see how the participation is in the level of awareness probably. But maybe because you also have a data for the companies itself, and I think several of the participants before already asked whether the one working in the in that company also comes from a certain country, which is will have a better influence to raise not only awareness but the participation, the real participation in the future. So I was thinking because you took the data from mailing list or probably participation list, probably the the participant of IETF themselves would not mind to be included in the research more actively. As for Indonesia, which happens one of them is being mentored by Alvaro himself, he is very happy. And then when we find try to find other Indonesian in IETF is so rare, but I found Indonesian names in the participation list that is actually working in New Zealand and other places. And I was trying to chat with them through MeetEcho actually, and then whether they would wanting to being involved more into reaching out the students as well. And they most of them are very happy because they see that one of the biggest problem that comes from this country is not whether there is fellowship or not, but because they are not aware how technical standard is actually quite reachable for them to get involved into, especially in IETF. So most of the university that we were working on in Indonesia, their lecturer just come to us and say, oh, it's not our problem, we don't give those kind of curriculums here, that's the problem of the higher level government or something like that. So they don't pay attention. But the ones who eventually we managed to bring into our fellowship and see IETF, they say, hey, this is not that far, it's not Mount Olympus, you can do more about it. So maybe Jan already said that we would like to work together if that's possible, but yeah, probably that is just my input that this data can we can develop this data to be more than just awareness, it can be used to educate other people to bring more student, more young generation inside the IETF itself because I see like Alvaro is here, he is so openly wanting to teach. And I'm sure more, most of us is like that as well. Thank you.

**44:51**

**Priscilla:** Yeah, thank you. Yeah, I hope that other IETF participant can give us feedback and help us with all this data. I mean, we don't need to put the names on it, just give us feedback and new ideas and you can see what we can do with all this data and collaborate. 

**45:11**

**Alvaro:** Thanks. I put myself in the queue. This is personally very interesting to me because I'm one of the Latin American people who doesn't show up in the list of Latin American people because I live somewhere else, right? And I work for a company that is based somewhere else. Um, I did a lot of work with LACNIC before IETF 95, trying to get to that peak. Um, one of the things that we noticed and everyone keeps talking about geographical distance. And yes, that is important. At the same time, in regions like Latin American, the region is very big. So going from Mexico to Argentina, for example, is very far. I mean the same can be said for Asia, for example, or APAC as we see it. You know, going coming from Australia here or vice versa, right? Going from China or Japan to to Australia when we had the IETF there is—is far, right? So we can't always assume that because we are in the region that everyone is going to come. And I think in my opinion that's why we saw so many remote people in Argentina. 

Um, last night, if any of you went to the plenary, Jay said something that to me was the highlight of the night, which was that for IETF 133, I believe, they were looking at a country in South America. So I'm not sure where, and I can guess and hopefully, you know, where it is, but, you know, this gives an opportunity, right, to ramp up again. So you said that if anyone was willing to work with you guys on getting the word out, I am willing to do that. Anything I can do to help in Latin America, I'm very, very happy to do that. Um, I do have one question. And the question is, um, definitely fellowships and financial support are important. Um, we've talked about ISOC, we've talked about the NICs, you know LACNOG or LACNIC and of course in Brazil the CGI helps a lot with with money, all these organizations that are very interested of course on the internet that already have participation in the standards. Um, there are many companies that participate in the IETF. My company, for example, has provided financial support for two or three people to come to the IETF, but we haven't done it consistently to be honest. Now how—how can we motivate other companies to do similar things? Um, you know what is the—and you know maybe there's not a quick answer, you know what is the business case for them or what is the justification, the motivation for them to provide financial support for people that are probably not part of their company, you know, to participate, especially in as you mentioned several times the Global South where we don't see that many companies, manufacturers or operators participating, but it is more academics and and other people that participate there. Um, so if you have something that we could use, that would be great, but I think it's something that is going to be important for us to to push in the future so that we can get more people providing fellowships or, you know, financial support here and there to get more people to participate from the Global South. 

**48:43**

**Priscilla:** Oh yeah, I mean, I don't have a answer to that, but I think that technology companies and especially the people from IETF and the organization of IETF can help us with this, so we can have improved the participation of the south global definitely. 

**49:03**

**Rod Van Meter:** Hi, Rod Van Meter. So Priscilla, thanks for coming all—all the way. I realize that's basically opposite side of the planet for you. Um, one question but preceded by a comment, which I promise is actually related. Um, so I'm from Keio University in Japan and from the WIDE Project, and with APNIC, one of the things that we have done—I don't get much credit for this, the others get all the credit for it—but a few years ago we created the APIDT, the Asia Pacific Internet Development Trust, and it runs a number of educational programs as well as the actual as well as an actual REN, you know, a network throughout Southeast Asia to Indonesia, Malaysia, Singapore, Guam. Um, and that's targeted at—the—one of the things that's with it is the is a program called APIE, I forget exactly what that stands for, Asia Pacific Internet Engineers or something like that, but it's targeted at undergrads and getting them involved in internet operations and understanding the protocols and things like that, bringing them in with the hope that eventually they'll join the community like this. Um, but you know, it's a very long road from that undergrad level all the way up to real contributor here, and there are lots of places where people fall by the wayside, and one of the things that with all of the different organizations that are involved, moving from being supported at one level to being supported and mentored at the next is—is I think one of the places where we lose people in—in that overall end-to-end pipeline. Um, so that's the comment. Related to that, the question: one of the reasons that we created this is, um, at least in Japan, and I think this is true in other places, over the last 10-15 years, we have seen sort of a decline in interest among the students in L2, L3, L4 issues, the—the kinds of things that, you know, most of us cut our teeth on and grew up working on. They're all interested in blockchain and AI and building websites and Y Combinator and getting rich some way, shape or form. So it seems the students have many options, they have many different things, and it's—but those of us who work on the core internet stuff, we feel like we're not done yet, we feel like there's a lot of work still to do and we want young people to come up and take over and do this stuff behind us. So, question: are you seeing that same thing in—in Brazil or across Latin America more broadly, and do you have sort of a strategy for dealing with that or recommendation?

**51:52**

**Priscilla:** Ah, yeah, thank you. I was waiting for someone talk about this. Uh, yeah, I mean, I feel the same, that is really hard like to catch the attention of the students to work with computer networks because right now we have so many things happening in computer science and different things catch their attention like AI, soft development, this type of things. Uh, but it's my personal opinion, okay? I think that students, especially undergrad, they are usually follow not just what is the trend of the moment, but also when they see that they are professors really engage in the theme, in the topic, and then they usually engage together with the professor. That's why what you say about have something like a educational program that can help us with this, because right now in Brazil, we have like isolate initiatives from professors in different in different parts of the country—Brazil is a huge country—but for now is not—there is nothing like coordinate like something that's like a computer network school or summer school or something like that that could bring the students together, and also that they could see other students that are interesting in the same. Sometimes I have a student that talk with me to work with this, but it's the only student and they feel so alone. But if you have like something integrate like the program—I would like to talk about this later—this could like help us to improve. And I really believe that is not something that we can do and next year we will have results. I mean, if you are—if we are talking about working with undergrad students, this student—you'll have some time until this student start to have like a start to contribute with IETF and start to do like important things. It's something that we start to do now and you will have the results in a few years, not next year. But yeah. Thanks.

**54:00**

**Ignacio Baguena:** Thank you Priscilla for the really nice presentation. I'm very glad you could make it. Thank you. Um, so we also went through the hurdles of doing affiliation and location mapping, so I'm very happy to discuss about those because yeah, it's—it's quite of a painful process and very imperfect. There was also a presentation from Justus Baron in Dublin, also looking at the geography of IETF participation that you might find useful. And the question that I had was actually on the line on the one of Rodney. Have you look or are you planning to look on what are the factors that not just foster participation but foster retention? People that manage to stay within the IETF orbit for a certain amount of time or make a contribution? I was wondering if you look into it up to what extent that is related not so much to the excitement of the topic but to the ability to make a contribution.

**54:58**

**Priscilla:** Ah, yeah. We are working this right now, but it's kind of difficult because we can identify that okay, this is a new participant in IETF. This participant will participate only this meeting in Argentina or here in Shenzen or this participant will continue contributing going to IETF. This is kind of really hard because depend of the attendance, they put different names, they change the last name, and sometimes it's Robert then Robert become Bob. We are still doing cleaning pre-processing this data to try to—

**Ignacio Baguena:** That is true for the participation in the meetings, but through the data tracker for the mail list and anything that is recorded through MeetEcho, you can use the ID of the data tracker. 

**Priscilla:** Ah, yeah, yeah, exactly. That's why we start to to collect the data from data tracker because only with the attendance data we cannot do that, but we are interested in that like identify the participants, okay, this participant start in, I don't know, Argentina meeting and then three years later he proposed a draft, something like that. Yeah, this is something that we want to to do, especially because we would like to understand and identify like okay, this was a, like Jefferson, that was a fellow in in the CGI program, now he is a chair, what which was the path? But he is not the only one. In Latin America we have other examples, so we would like to identify this type of things. Yes, definitely something that we want to do.

**Ignacio Baguena:** Thank you. 

**56:41**

**Arnaud:** Hi all, Arnaud. So I will speak as a Study Group 17 chair in the ITU-T and I'm here for sharing some input and as well to—to perhaps propose something. So I heard a couple of issues regarding of course Latin America is a prime for me as well. So, but in the ITU-T we have the concept of regional groups. So as Study Group 17 chair yesterday I bootstrapped my SG17 Africa group, soon I will bootstrap my Study Group 17 Arab group, but I don't have a Study Group 17 regional group for Latin America. Um, when people work with a formal regional group, that's very helpful because suddenly they can share their own problems as a region, right? And when I go to their meetings, I'm not the chair, I'm just a guest. They have full rights, this is their meeting, they do they can make their decisions and so on, so there is a little bit of that. Now, good thing is that people do contributions and they are doing good technical contributions. So they have started one of the first work items from Africa passed through—amazing. So point here is that what it is for Latin America and Anatel is there. I mean, Brazil is there, right? And strong, they—they are doing a good work and so on. But the point is that I'm wondering, it's difficult to organize these things. Is there is there not a way that in Brazil, for example, you guys could start a group where you mutualize and start to send people in different standard bodies because you would have a more coherent view about how you want to address standardization and you could organize yourselves to understand more holistically because not every standard body will will provide the same value point or value—value proposition to the students you have, to the people you have, to the engineers you have, to the companies you have. That would be a first perhaps thinking to have. 

Second is about the point from our chair about industry engagement. That's both a problem for developing countries and a problem as a whole. And in fact, the problem we have is actually a problem for developed countries because the the industry engagement is crashing everywhere since 20 years. It's—it's harder and harder and it's going to accelerate to keep our own roles in our own industry. We in ITU realized the problem three years ago, it was really hard, we lost a first resolution 68 at the assembly of WTSA in 22, we lost the resolution in Plenipotentiary in 22, but we won all the action plan. Next week, if interested, there is the second workshop on industry engagement with ETSI in Sophia Antipolis on Friday. And I think there is remote participation. The point is that as I'm coming from the industry and Broadcom, this is a very difficult issue. I cannot make it the whole discussion now, but we we have a full action plan, we could share it, you could see all what we learned, we have the report, the the report of the first workshop was so remarkable that it went through the whole of the ITU for all the three sectors. And the next other thing you have as an issue is the problem of next generation. And these are connected but separate problems. So we managed to create the first resolution 107 on next generation. Point is that we learned a lot on our side in the past three years to re—to recreate the engagement, right? And maybe there are things we could discuss offline I could share. But the first thing is if perhaps Latin America could create an entity or something to organize itself because you—you need the multi-stakeholders, you need you need governments, you need industry, you need academia, you need civil society and this is hard. So perhaps mutualizing the resources for attacking standardization is a proposal. Anyway, that's my feedback. 

**60:50**

**Priscilla:** Ah, thank you. But talking about—I mean, I don't think I can talk for the entire Latin America, but it's a huge—it's a huge part of the globe and I don't know if you have like a coordinate something coordinate between all the countries. But in Brazil, we start to to do some initiatives to do that, like for example we have a computer network conference in Brazil that we we will have a day in the conference to talk about IETF, and I think this is start—we are starting something to engage the community. So yeah, but still babe steps. 

**61:36**

**Ignacio Baguena:** Fantastic. Thank you very much Priscilla. 

**61:46 [Introduction for From Standards Adopter to Standards Author]**

**Ignacio Baguena:** And for our next presentation, we have Mi-Jin Kim. Kim, are you there?

**62:07**

**Mi-Jin Kim:** Hello everyone. Can you hear me? 

**Ignacio Baguena:** Yes, very well. Yes. So let me let me introduce Kim. Kim is a research fellow at the Graduate School of International Studies at Yonsei University. She got her PhD at the international management from Graduate School of International Studies at Yonsei University. And she was also senior researcher at the Korean Standards Association. Her research focuses on the political economy and geopolitical dimensions of international standardization, particularly in critical and emerging technologies. And I think it's a good continuation to the previous presentation because her presentation looks at the—at the how Korea moved from being standards adopter to offering standards, not just in the IETF but more broadly in general to ICT in general in this presentation, and this comes from a background paper for the World Bank. Thank you very much Kim and over to you.

[Mi-Jin Kim's slide deck: [From Standards Adopter to Standards Author: The Evolution of Korea's National Standardization System with a Focus on ICT](https://datatracker.ietf.org/meeting/125/materials/slides-125-rasprg-from-standards-adopter-to-standards-author-the-evolution-of-koreas-national-standardization-system-with-a-focus-on-ict-00)]

**63:14**

**Mi-Jin Kim:** Yeah, thank you for having me here. And maybe this topic is not really closely linked to the standard activities that you do in IETF, but today I just briefly talk about how Korea evolved from being a standard adopter to becoming a standard contributor in global standardization landscape. 

So let me begin with a simple question: what is a standard? Um, I realize that most of you here are very familiar with the concept of standard, so I just briefly revisit the definition. So in simple terms, standards provide a common reference that allows technologies, products, and systems to work together. For for example, as you can see in the slide, paper sizes, shipping pallets and USB connectors are all based on standards. So without standards, modern industrial production and global trade would be extremely difficult. So as economies become increasingly digitalized these days, the role of standards becomes even more critical in ICT sectors because ICT sector is the place where interoperability between devices, networks, and platform is essential.

So why are standards so important? So the first, standards function as a key infrastructure for economic and industrial activities, which ensures the quality and safety of products and services. And second, standards enable compatibility and interoperability between technologies and systems. So it is especially important in ICT sectors where multiple devices and networks must seamlessly interact with each other. And third, standards support the diffusion of innovation and the expansion of market. So in ICT industries, standards often played a decisive role in which technologies are adopted globally and how markets evolve. In particular, for developing countries, standards are important because standards help firms to enter global markets and participated in global value chains. So they they can help build trust in global markets by demonstrating that products and services meet internationally recognized requirements.

So to understand how standards function at the national level, we often use the concept of National Quality Infrastructure that we call NQI. So NQI refers to the institutional framework that ensure the quality and safety of products and services. It consist of, usually consist of four main components: standardization, methodology, conformity assessment and accreditation. So together, these elements form the institutional foundation that supports industrial quality and international trade. Especially in ICT sectors, this institutional framework is particularly important as it supports the development, standardization and global deployment of digital technologies.

So I think in this perspective, I think South Korea provides a very good example of economic development through an export-driven industrialization strategy. So in this process, improving product quality and meeting international standards became essential. As a result, the development of Korea's national standard system and infrastructure played a key role in supporting industrial growth and export competitiveness. In other words, Korea's economic development was closely linked to the evolution of its national standard system.

So as I mentioned in the beginning, Korea's economic development has been widely studied from many perspectives. But this in this presentation, I revisit Korea's development from the perspective of standards and standardization. 

So in perspective of, perspective of ICT sector, since the late 1990s, Korea has become increasingly active in international standardization activities, particularly in ICT sector. So this marks Korea's transition from adopting international standards to actively contributing to the standard setting.

So this slide illustrate the relationship between standardization and industrial development in Korea. So Korea's economic development strategy relied heavily on export-oriented industrialization. So to export products successfully, manufacturers in Korea needed to meet quality and safety requirements that defined by international standards. So standardization helped firms to improve product quality, production efficiency and technological capabilities. So as you can see in the graph, the growth of Korea's industrial standards that we call KS has been closely accompanied by the increase in Korea's export volume. So this suggests that standardization played an important role in supporting Korea's export competitiveness. So this experience laid the foundation for Korea's later expansion into the ICT standardization and international standard setting activities.

And this slide shows the number of new work item proposals submitted by Korea to international standardization organizations like ISO and IEC. As you can see in the graph, the number of proposals has steadily increased over time. So it indicates that Korea is no longer simply adopting international standards, but is increasingly proposing new proposals at the national level. In other words, Korea is moving from a standards adopter to a standards author in international standardization. It's not mentioned in this graph, but a significant portion of this increase has been driven by Korea's active participation in ICT related standardization activities.

And this slide shows Korea's increasing leadership roles in ISO and IEC. These leadership roles including the positions such as chairs, secretaries, and convenors in ISO and IEC. So as shown in the figure, Korea's presence in these leadership roles has steadily increased over time. So these positions are important because they allow countries to actively shape the direction and content of international standards. So it reflects Korea's growing institutional and technical capacity in standardization. And many of these leadership roles are concentrated in ICT related technical committees and working groups. So this is very important for the evolution of Korea's ICT standardization.

And this table is that I summarized the evolution of Korea's national quality infrastructure across four stages. So I I just briefly mention about this. So the first stage is the founding stage, which began in the early 1960s. In this period, Korea focused on building the basic institutional foundation for industrial standardization. So in this period, key legal frameworks such as the Industrial Standardization Act and the KS mark system was also established to ensure the quality of industrial products. These initiatives were very closely linked to Korea's export-oriented industrialization strategy where improving product quality was essential for entering global markets. And the second stage is the growing stage, which focused on strengthening the institutional and technical foundation of standard system. So in this period, Korea also established several key institution that later became central to ICT development and standardization. For example, the Electronics and Telecommunication Research introduced that we called ETRI was established in 1976 and also the Telecommunication Technology Association was established in 1988. So these two institutions later became key organization of Korea's ICT standardization ecosystem. And the third stage is the upgrading stage, which began in the late 1990s. In this period, Korea significantly strengthened its national standard system and international engagement in ICT standardization. For example, the Framework Act on National Standard was introduced and Korea also expanded its participation in international standardization activities in this period. And at the same time, the collaboration between ETRI, TTA and industry sector became more institutionalized in this period. So this linkage between ICT R&D and standardization and also industry participation played a key role in Korea's ICT innovation. And finally, the most recent stage is the expanding stage, which focuses on further strengthening national standardization strategies and governance. One important initiatives during this stage is the Pan-Government Participation Standard System which is introduced in 2015. So under this system, standardization responsibilities expanded beyond the single center authority that we call Korea Agency for Technology and Standards from many other ministries. So it allows different government agencies to participate more actively in standardization activities related to their policy areas. So this helped create a more coordinated and cross-government approach to standardization. And at the same time, Korea has been developing more strategic approaches to ICT standardization. For example, government introduced the National Standardization Strategy for High-Tech Industry in 2024. So this strategy aims to strengthen Korea's leadership in standard setting for advanced technologies. And in and also TTA, one of the key organization in ICT standardization in Korea, this organization published the ICT Standardization Roadmap every year. So it identifies the priority technologies and guides Korea's participation in international ICT standardization. So together, these initiatives reflect Korea's effort to systematically link technology development, industry policy and international standardization together. So as a result, Korea has strengthened its position in ICT standardization at the global level.

So from here, let me now turn to the several example from ICT standardization in Korea including CDMA and WiBro and DMB. So the first case is CDMA mobile communication technology. So let me briefly talk about this case. So in the late 1980s, Korea experienced rapid growth in mobile communication demand, especially after the 1988 Seoul Olympics. So it created severe congestion in analog cellular system. So to address this challenge, Korea began developing a next-generation digital mobile communication system. So importantly, Korea made a strategic decision to adopt CDMA as its national standard for digital mobile communication. So in 1988, ETRI proposed a national R&D project and the government launched a national development program. And then in 1991, ETRI signed a joint development agreement with Qualcomm to develop CDMA technology together. So CDMA was adopted as Korea's digital mobile standards in 1994, and then it followed by successful commercial trials in 1995. So in 1996, Korea launched the world's first commercial CDMA mobile services. I think this case demonstrate how standardization combined with R&D enabled Korea to significantly enhance its technological capabilities and strengthen its domestic mobile communication industry. So this is the one of the most important ICT standardization case in Korea.

And the second case is WiBro. So I think everyone know, but WiBro stands for Wireless Broadband. So originally, the Korean government allocated the 2.3 gigahertz band for fixed wireless communication, but the actual demand was very limited. So to utilize the spectrum more effectively, Korea introduced the concept of portable internet which enables wireless broadband access on the move. So WiBro was developed as part of Korea's IT839 strategy which aims to strengthen national ICT capabilities. So the development project of WiBro began in 2003, so it is led by ETRI and Samsung Electronics. And also there is a participation from major Korean telecom operators. And by 2004, prototype systems had been successfully developed and demonstrated. And commercial WiBro services were launched in 2006. And importantly, WiBro was adopted as an international standards under ITU-R IMT-2000 in 2007. So I think this case shows how Korea not only developed a domestic technology and but also actively promote this domestic technology as an international standard.

And the third case is digital multimedia broadcasting that we called DMB. So in the early 20s, there was growing demand for multimedia content on mobile devices. So Korea aimed to develop its own mobile broadcasting technology to reduce reliance on foreign technologies. So there is on there was an research on Terrestrial DMB began in 2002 and the system technology had been successfully commercialized by 2004. So DMB was adopted as a standard by the World DAB Forum and later by ETSI. And in 2005, Terrestrial DMB broadcasting services were launched in Seoul and DMB was selected as a recommended international standard by ITU in 2007. So this case illustrate how Korea combined technology development and standardization to expand its global influence in ICT industries.

So I think based on Korea's experience, we can draw several key lessons. So the first is that standards can serve as a foundation for industrial development, particularly for export-oriented economies. And second, strong government leadership and policy coordination played an important role in developing Korea's standard system. And third, Korea developed a coordinated institutional ecosystem for standardization involving research institutes, standardization bodies, government agencies and also industry sector. So this close linkage between technology development and standardization was essential for successful ICT innovation in one country and also in global. And finally, Korea's active participation in international standardization enabled countries to transition from a standards adopter to a standard key contributor to global standards.

So in conclusion, Korea's experience demonstrate that standardization can play a strategic role in economic development and technological upgrading in one country. So the development of National Quality Infrastructure provided an institutional foundation for industrial growth and export competitiveness in early stage. And since the late 1990s, Korea has strengthened its role in international standardization, particularly ICT sectors and a coordinated institutional ecosystem that linking government research institute standardization body and industry has been a key driver of success in Korea. And for developing countries, I think Korea's experience highlights the importance of strategically linking technology development, industry policies and standardization. So in this sense, standardization is not just a technical activity for engineers, but also a strategic policy tool in ICT and beyond. So this is the end of my presentation and thank you very much for your attention.

**87:31 [Q&A for From Standards Adopter to Standards Author]**

**Ignacio Baguena:** Thank you Kim for the for the very interesting presentation. Thank you everybody for the really good questions. Sorry being remote makes it really difficult to know that there is going to be clapping. I find myself a little bit off-foot in terms of timing all the time. I was wondering what lessons do you think that can be taken from Korea to other countries more generally. So before just we had a presentation about participation from Brazil and how to engage more broadly Latin America in standards participation. I was wondering like are some of these lessons exportable or they are very Korea specific? 

**88:21**

**Mi-Jin Kim:** Oh sorry, there is some congestions. Oh, the question was what of the lessons that you have learned with regards Korea you think can be exported to other countries? Ah. You mean what's the implication of Korea's case for other countries? So I already mentioned in maybe slide this slide. I think maybe Korea's lesson just focused on very export-driven economies, but I think the point that Korea that strategically linking the technology development and standardization in very early stage of technology development and standard setting is applied to other countries. I think that's the key lessons to other developing countries. 

**89:40**

**Ignacio Baguena:** Okay, thank you. I think Jan is next in the queue. Appreciate it if people can be a little bit succinct with the questions because I'm realizing we are starting to run out of time. 

**Jean-Francois:** Jean-Francois again. Um, wouldn't be much of a question, more of a remark. So Korea is quite remarkable in general in how they have become a receiver of support internationally just 50-60 years ago to be an actual exporter of goods, standardization donations and so forth internationally. What I find very interesting is when I look at the in three SDOs: so the IETF, ICANN and ITU-T. It's very interesting to see how different countries seem to have different degree of of participation. I'm not that much familiar with with ISO, but definitely South Korea participates quite a big deal with with the ITU, as Arnaud can definitely witness. The first one being China and then South Korea and Japan are pretty much at the same level. But then for instance in the IETF you don't see that much of that. But I was also thinking ICANN, you also don't see that much of that. You see them at the GAC because they have to have a government representation, but there aren't that many players in from Japanese or South Korean interest, whereas there's somehow a little bit more on the on the Chinese side. So I would be very, very aligned with what you mentioned as the the government alignment and support is fundamental for any of this to happen, if we are linking this to the previous presentation. If you want to think as to why South America may not have brought so many people, well there will probably be big differences in their national laws or acts for standardization, whether they were looking inwards or outwardly at worst. China for instance has a very, very public international strategy when it comes to technical standards and it is what drives them to to present 80% of the contributions to ITU-T for instance come from China. So the government and also whatever industry has been developed in that particular country or jurisdiction seems to be the main drive to whether that particular country will have engineers participating in standardization process or not at the international level.

**92:07**

**Arnaud:** Yes, so Arnaud again. So thank you very much for this really brilliant presentation. It's really impressive. I'm not going to repeat what Jean-Francois just said for for us in ITU, Korea is in my study group, this is partner number two just after China. To give an idea, Korea brought 75 contributions just to my study group in the plenary of December. The—and the former chair of Study Group 17 was from Korea. So there are many you are right, there are many positions, leadership positions that Korea trusts in in a number of places, and that's a good way to lead your national standardization strategy. So this is a good KPI. What I find really amazing into your presentation is slide seven. Not sure if you can go back to slide seven, but the correlation you make between standardization to the output in terms of your export increase. I I find this this slide absolutely amazing. And I would be really happy to understand more how you build it because one of the issue for industry engagement that we discussed before is exactly that. If we could show our industry that type of slide, that is a no-brainer. Everybody would understand the return on investment, the impact it has on the on the economy. Now, can this be applied to all countries? I doubt, because not all countries do have the same flourishing industry as Korea has. So I suppose that a country in developing a developing country would have a very different footprint and therefore it would be more difficult for them to get these results. But nevertheless, this this slide is remarkable. So just thank you for your really great presentation.

**94:10**

**Colin Perkins:** Okay, Colin Perkins. Sir, thank you for the interesting talk. The IETF has met in Korea a number of times and we have certainly had a significant amount of participation from participants from Korea over the years, both from industry and from the various research institutes. It would be interesting to see and to understand if the if that participation tracks with the types of programs and activities mentioned in the talk and to understand if these types of national programs drive participation in SDOs as a whole, or if they are just target just a small number of targeted major SDOs. And I suspect you don't have the data to hand, but it would I think it would be an interesting follow-up. So thank you.

**Mi-Jin Kim:** Ah, I think most of maybe—I'm not sure because I'm not politician or policy maker, but the most of Korea's national program are targeting the technical committees or working groups in ISO, ITU and IEC kind of big international standardization organization. But these days Korea also targeting the de facto standard development organization too. So maybe there are some two track for de jure SDOs and de facto organizations. So Korea tried to um try to send more people to standard development organization, but usually targeting the major three international organization. 

**Colin Perkins:** Yeah, yeah. I I see that. I also see the participation in organizations like the IETF varies over time and I'm wondering if it correlates and if if these large scale activities are incidentally helping engagement in other organizations. It would be interesting to find out. Thank you.

**96:55**

**Ignacio Baguena:** Hello, I'm Ignacio. Yeah, thanks very much, really fascinating presentation. Um, so South Korea, you know, seems to have the world's most organized national standard strategy. So um so here in the IETF we have also seen, you know, good contribution from Korean institutions and especially in our area you seem to have this special setup with the ETRI as you know one like almost centralized standards contributor. Um, I'm just wondering: so this seems to be also like a big investment somehow by the by the country, by the state, by the nation. Um, and you are measuring contributions. Do you also have some data on maybe kind of return on investment or impact? I I'm sure—I know this is hard to measure, but um I'm just curious, you know, how how is this data you know evaluated? How do you know it's kind of working for the economy, for the for the companies? 

**98:08**

**Mi-Jin Kim:** Oh, you mean the how to measure the effect of standardization for country or company? Is it yeah, so overall for the country. So I know like if you work in a company, you often try to achieve certain impact, right? So with your standardization activities. Um, I'm just wondering whether you are also measuring this on the on the national level. Ah, I'm not sure about there is some maybe Korean government government, they try to measure the impact or they develop indicators to tracking the effect of standardization activity, but as far as I know, there is not there is not there is no kind of such data to the public. They do not open to the public. Okay, okay, fair enough. Thanks.

**99:32**

**Sarah:** Hello, this is Sarah. My question is how critical to the growth of the ICT work was the ETRI R&D lab with such strong efforts? 

**Mi-Jin Kim:** Ah, you mean the ETRI, right? 

**Sarah:** No, ETRI, the the R&D lab. ETRI, how critical was it to the growth of standards to have a strong R&D lab in the organization ETRI for Korea? 

**100:30**

**Mi-Jin Kim:** So you mean the how ETRI, how how how did—you mean the ETRI's. I have worked, let me try a different way. I have worked with the engineers at ETRI over the last 20 years and they were very good and they had great ideas and they worked aggressively to see those ideas go into standards. I have been impressed with them. What I am looking for is they were a uniquely powerful researchers and how did this impact Korea's growth in standards? So actually, the technology that developed in ETRI and standardization that is developed in ETRI it goes to the international SDOs such as IETF or many other international SDOs. So I'm not sure but actually the technology and standards that created or developed from ETRI, it it goes to the international development organization as proposal. So the impact of ETRI is bigger, very big in Korea's standardization ecosystem.

**Sarah:** Okay, that was my personal reflection having experienced ETRI and the commercial companies in Korea is that ETRI the research done at the ETRI research took a enabled quite a bit of commercial growth. I was looking for your statistics on that growth. 

**Mi-Jin Kim:** Ah, commercial growth. I think I maybe I have to check about the commercial growth maybe later. 

**Sarah:** Thank you very much. 

**Mi-Jin Kim:** Thank you. 

**103:12**

**Ignacio Baguena:** Well, Kim, thank you very much for the very interesting presentation. Thank you everybody for the really good questions. So now we can move on to the last presentation. So for the last presentation, we have Noga—Noga Rottman. Unfortunately, Noga won't be able to make it in person. She had a possibility that she would not be able to make it, so she recorded a video just in case and it looks like that's been the case. So let me introduce Noga while Alvaro sets up the video. We hope that the quality of the audio is good as it's a bit complicated. Noga Rottman is a research fellow in machine learning and networking at University College London and a recipient of the Eric and Wendy Schmidt postdoctoral award for women in mathematics and computer science. Her research explores the intersection of networking and machine learning with a focus on building reliable and auditable systems. And she previously held a Bitterby postdoctoral fellowship at Technion and completed her PhD at the computer science at the Hebrew University of Jerusalem. And today she's going to be presenting RFSeek, which as she will show in the video, tries to take the RFC specifications and make sense out of them using different LLMs and machine learning tooling. Thank you very much. I don't know if we have it there. Yes, we have it there. Okay, let's go ahead. 

[Slide deck for [Beyond ASCII Art: Making RFC Protocol Logic Auditable with RFSeek](https://datatracker.ietf.org/meeting/125/materials/slides-125-rasprg-beyond-ascii-art-making-rfc-protocol-logic-auditable-with-rfseek-00)]

**104:52 [Noga Rottman's video for RFSeek]**

**Noga Rottman:** Hello everyone, and thank you for having me. I'm Noga Rottman and the title of today's talk is Beyond ASCII Art, Making RFC Protocol Logic Auditable with RFSeek. This is a joint work with Thiago Ferreira from UCL, Gilad Parag and Mark Silberstein from Technion, and Alexandra Silva from Cornell. 

The outline for today's talk is as follows. First, I'll motivate the work. Second, we'll discuss the methodology. Then, we'll look at some results. And finally, we'll discuss future plans. 

I'm sure I don't have to tell this crowd that modern RFCs are challenging to reason about. They are made out of dozens or even over a hundred pages of specifications, and protocol logic is spread across the document, often many pages apart. To help readers, RFC authors often include finite state machines, or FSMs, in the form of ASCII diagrams. However, partially due to the limitations of the format, these diagrams are typically simplified abstractions that omit many details. And, as we'll see, they can sometimes miss important pieces of protocol behavior. 

Let's look at an example from TCP. RFC 9293 is a unified specification that consolidates about 40 years of TCP evolution. Many of us probably recognize this diagram, the TCP FSM from that RFC. The note at the top clearly warns that this diagram is only a summary and must not be taken as the total specification. Many details are not included. Even so, while TCP is probably the most studied protocol on the internet, a transition described in the RFC text never made it to the diagram. Using RFSeek, the tool I'll be discussing today, we were able to identify that transition. Interestingly, the transition already appears in real implementations. For example, in the Linux TCP stack. This eventually led to an editorial errata being submitted and accepted. 

So, how did we achieve this? Our general approach is to create an auditable protocol visualization grounded in the RFC text. We call this summary visualization. This is done by taking the RFC prose—here, for example, is a fragment from TCP—extracting the protocol logic, and then turning it into an interactive diagram. The key property is that each component remains linked to the RFC text that produced them, allowing users to audit the extracted logic. While summarizations and visualizations are common in natural language processing, applying them to protocol specifications with provenance links back to the RFC enables a new kind of auditability. 

Let's talk about our methodology in more detail. A summary visualization differs from a traditional FSM diagrams in four ways. First, every node and transition is grounded in the RFC text. Second, the transitions themselves include richer semantic information, such as conditions, triggers, actions and rationale. While some of this information may appear in traditional ASCII diagrams, there is no consistent convention across RFCs. One RFC may only show triggers, another may include partial action description, and in other cases, the constraints of the ASCII format limit how much detail can be shown. Our intention is to make all the relevant transition information visible and accessible to the user. In other words, instead of compressing protocol behavior into a minimal diagram, the goal is to expose the structure that is already present in the RFC prose in a form that can be inspected and explored. Third, the diagram includes recommended and inferred components that are derived from the prose, not just the mandatory ones. And finally, the visualization is interactive, allowing users to inspect and review the RFC in more depth. Next, we'll see how these properties were realized in RFSeek. 

Extracting protocol state machines from RFC prose is actually quite challenging. Recent work such as PSM-Bench highlight how difficult this task remains even for modern language models. To construct our summary visualization automatically, RFSeek uses the following pipeline. The first challenge is simply that the original RFCs are too long. A single protocol specification may span over 100 pages. So, rather than feeding the entire document all at once, the text is divided into structural chunks based on the RFC structure: sections, subsections and so on. We use embeddings to identify the relevant context for each chunk and use them together as input to a large language model. The model, in turn, produces a structural summary for each chunk. Because the context can originate from different sections of the RFC, these resulting summaries capture the protocol logic even when relevant information for a specific chunk is distributed across the document. This stage is called structural summarization. The output of this stage is not yet a diagram. Instead, it is a structured description of protocol behavior. In the next stage, these summaries are used to extract the protocol visualization. Using the summaries as input, the model identifies state transitions and conditions and assembles them into a protocol diagram, finishing this stage of the pipeline. Finally, we perform semantic grounding. The system retrieves the corresponding RFC text fragments that justify each edge and node. This way, the resulting visualizations remain linked to the original text, allowing users to inspect each component and verify the logic directly against the specification. 

Once we load a visualization summary for a given RFC, our screen splits into two. On the left side, we have the main view of the diagram. A user can zoom in and out and arrange the nodes in the way they see fit. This arrangement can be saved for later inspection. The text across each edge denotes the trigger for the transition. Hovering over each edge shows the action or actions that should be taken if such exist, as well as the original text fragments. Clicking on the arrows shifts the view on the right side, which contains the original RFC text. So, the user can automatically scroll between the relevant fragments to inspect and audit the extracted logic. By examining the visual summarization for TCP in this manner, we were able to identify the missing transition and submit the errata. 

Now, let's look at some results. This slide summarizes our results across several RFCs. When possible, we compared RFSeek against Prosper, a recent LLM-based FSM extraction system. For DCCP, RFSeek reconstructed all nodes and missed only a single edge compared to Prosper's missing node and seven missing edges. For PPTP, again all nodes were reconstructed with six missing edges compared to Prosper's 19. RFSeek also surfaced an additional state implied by the RFC text. For QUIC, RFSeek missed two edges but unified the send and receive FSMs and generated new diagrams for procedures that appear in the text but not represented in the figures provided by the RFC. For TCP, RFSeek missed a single edge but also surfaced a new one, which led to the editorial errata we discussed earlier. Broadly, these results show that RFSeek helps reconstruct and audit protocol behavior directly from the RFC text. Interestingly, this isn't limited to newer protocols like QUIC. Even mature protocols like TCP can still contain discrepancies between diagrams and text. More importantly, the tool goes beyond reconstruction. In several cases, RFSeek also surfaced behavior that was either implicit in the prose or missing from the diagrams. 

Let's now take a closer look at two examples. We'll first discuss the missing transition in TCP. This is a section of the TCP FSM as shown in the original text. There is one transition from LISTEN to SYN-RECEIVED, and another in the opposite direction, from SYN-RECEIVED to LISTEN. The note below the diagram states that this transition happens upon receiving a reset when reaching SYN-RECEIVED after a passive open. However, section 3.10.7.4 of the RFC also clearly states that when SYN-RECEIVED is reached after a passive open and a SYN is received, the connection should return to the LISTEN state as well. This transition, as you can see, is missing from the RFC diagram. RFSeek's visualization, on the other hand, surfaced this transition. As shown on the right, our visualization includes three edges between LISTEN and SYN-RECEIVED instead of the original two. Both edges to LISTEN show the joint part of the condition that is reaching SYN-RECEIVED after a passive open, and also the different triggering event being either receiving a reset or a SYN. When hovering over each edge, the corresponding text fragments are available for the user to inspect. 

For a second example, we'll look at QUIC, where we're facing different challenges. Though relatively new, RFC 9000 is widely deployed. The original document contains decoupled diagrams. This is the visualization created by RFSeek. As you can see, it unified the stream send and receive FSM on the bottom left and middle. It also generated visualizations for procedures that appear only in the RFC text and have no corresponding diagram, such as ECN validation on the upper left and connection termination on the upper right. 

To sum up today's talk, we introduced RFSeek, an interactive tool that extracts visual summaries of protocol logic from RFCs. Taken together, the examples show how RFSeek can surface protocol logic that it is otherwise difficult to see in long RFC prose. Of course, this is an LLM-based system, so human expertise remains essential when interpreting the extracted behavior. But tools like this may help make specifications easier to reason about for the benefit of both implementers and authors of new RFCs alike. 

As a next step, we are looking for protocol practitioners to help shape the next stage of the project. We're currently running short usability studies online and will also be conducting them live during the upcoming IETF 126 Hackathon in Vienna in July. If you'd like to learn more, please feel free to reach out or check out the paper on Arxiv. Thank you.

**119:20 [Closing of session and final comments]**

**Alvaro Retana:** Brilliant. Love it. I have many questions, so I will—I will maybe send them to you in email and I certainly look forward to the Hackathon at 126 so that I can see it live in action. I guess the biggest question is that you sort of answered it at the end that human, you know, review is unfortunately necessary at this point, even though it's discovered new things because almost every case had errors. So all of your ground truth is, how much effort did it take for humans to actually review the results? 

**119:54**

**Ignacio Baguena:** There are going to be no answers because unfortunately Noga could—the video! Sorry about that. Noga had some medical issues and they could overlap, and it seems to be the case, so we recorded a video just in case. But please make the questions because like that, she will be able to see them and get back to you. 

**120:25**

**Arnaud:** Yes, so yeah, again, this is really amazing to see that. We have the same problem elsewhere, so it's really inspiring. Um, yeah, so the the point it reveals as well and again back to industry engagement. I mean, this is with this type of tool that you improve the conditions of standardization. And this is really important, I mean, we measured how much it was important for us to improve our tools on the on the other side, because the industry is moving at another pace and not only that, the next generation expects something else. They they don't expect to continue having to work with the old tools we used in the past and sometimes systems are so complex that if you don't have an AI actually to help you see the mistakes, you may not see them anymore. Um, the other implication however is what if the AI makes a mistake and we don't see it? Or what if the AI makes a mistake with a bias? Because sometimes you have little things that come into standardization and then 30 years later you say, "Ah!" So anyway, I'm kidding but I'm not kidding, but it is a real point and maybe this is a point that we should put in a in a slightly bigger context. Thank you.

**121:52**

**Alvaro Retana:** I'm IETF Chair, so yeah, I think this is exactly the type of work that we'd like to see more of. And so be great if RASPRG, you know, could—could help, you know, bring more of these tools and test them and make them available for friendly user tests so that we can over time make some recommendations to Tools Team or the broader community. This would be fantastic. Thank you. 

**Arnaud:** Sorry I interject, I forgot one question on my side. The other question it brings is: what are the legal framework we have on these tools? Because when you use these tools, we we need a little bit of higher guidance to are we—it's not because we can use these tools that we should use them, and what are the exact tools behind and what do we allow ourselves, what we don't allow ourselves, so there is a little bit more that we need to do but nevertheless that's that's great. 

**122:52**

**Ignacio Baguena:** Thank you Arnaud. And thank you everybody for the many questions and engaging sessions. Apologies for running a couple of minutes over time. Um, as I said, Noga could not make it, but I'm sure she will be very happy to engage and you can reach her in the email which is in the in the slides. Otherwise you can also write to the group or drop me an email and I will be very happy to relay the messages. Um, thank you everybody. Have a nice IETF. 

**Alvaro Retana:** Nos vemos Ignacio entonces. Chau. 

**Ignacio Baguena:** Chau Alvaro, cuidate. 

**123:38 [End of recording]**