Ask Me Anything! with Katherine Townsend, Director at M-Lab

Drew Clark: And we're live. Good afternoon. Welcome to another episode of Ask Me Anything. I'm Drew Clark, editor and publisher of Broadband Breakfast. Super happy to welcome Katherine Townsend, the director of Measurement Lab here with us on the program. Welcome, Katherine. Make sure you're unmuted, okay? 

Katherine Townsend: Thanks very much for having me on, Drew.

Drew: Oh, no, it's our pleasure, we're so happy we could get you on. And I think you've got a super cool Twitter hashtag or a Twitter handle. Diplocat, tell us a little bit about how you got that, because you are a diplomat cat, aren't you? 

Katherine: I try to be. I guess I try in how I communicate, how I approach or how I see problem solving as diplomatically as possible.

Drew: But, I'm actually asking a broader and yet more specific question. You have been a diplomat, you worked for the State Department and had a number of extremely high ranking positions, just tell us a little bit about yourself, Kat and your background and how you got into this world of global internet standards.

Katherine: Yeah, I appreciate that. I love the characterization of high ranking, when you are in your early 20s it is amazing what they let you do in DC, which should both shock and potentially give lots of us hope. So yeah, I started out, I was fortunate enough to be an intern at State Department, working in their information, communication and policy office, and primarily the role there was to ensure that we were sustaining the capacity of US tech companies operating abroad. So if a government shut down access to a service or a website we would call up and say, "Please kindly open this back up again." And then also often cited a study that many of you may be familiar with, which is still cited that if you increased internet connectivity by 10%, that GDP growth would increase about one to two percent. That study was done by the World Bank, I think, in 2006, and that is just currency that just continues to be used, and I'd love to see it updated, but there really was, and I think there still continues to be, this idea that just being online, getting connectivity, being able to connect and share your ideas, share commerce, connect with people around the world to get different ideas of how organizations could be run, how governments could be run, there was a lot of promise in that and a lot of potential.

Katherine: And I think a lot of the utopianism of that time has given away to more sober minds, but I still really believe that the internet is an incredible invention. It can be used for many things, it can be used for ill, it can be used for good. And as much as we can steward work to lead that towards positive outcomes we'd really like to. And from that internship, I was privileged enough to then work in the secretary's office on some of the policies that they had about how to extend internet access, how to extend actually being online and using the web in statecraft and diplomacy and connecting US goals around the world.

Drew: And what... I mean, you've alluded to this, but I think it's worth again, for setting the stage for our discussion about Broadband measurement, what are the things that are so important about connectivity? In other words, the Arab Spring was seen as like, "Twitter is like, yeah, it's so cool." And now we look at that and we go, "Oh my gosh Twitter's a cesspool." We've gone from one extreme to another in our view of social media and its capacity to change, but underlying social media, of course, is Broadband: Broadband infrastructure and broadband application use. So from your vantage point, having traveled some of this journey, what are the kind of global requirements for internet accountability and engagement? Does that question make sense, Katherine? Can you elaborate on that? 

Katherine: I really appreciate it. There's so many ways to think about that question and think about how to answer it. So I'll just, I will put a plug in that one of the projects that Measurement Lab is undergoing is something called an Internet Quality Barometer. We have not begun this, so there's a reason that you haven't heard of it yet, but really wanted to address this question of what does it mean to have quality internet wherever you are in the world. We use a lot of measurements now, we use speed, we use say Broadband and there are others who have created their own measurements. There's an index called Meaningful Connectivity that was put together from the civil servant perspective, sorry, from the civil society perspective. And we really wanted to be able to show from internet measurement, from the scientific perspective, what are you actually measuring, that is a proxy for people's experiences to be able to have a quality connection to reach the goals that they have through similar experience wherever you are in the world, regardless of your geography.

Katherine: And so I think that is a, the internet evolves, the web evolves, and our understanding of it continues to evolve and the platforms that we have on it, including social media, I think there was a lot of confusion about what social media was when it first started, blogs, I can just... Going back to the diplomacy days, there were all these rules about second Life, if anybody remembers Second Life and there was a requirement in the Foreign Assistance Manual that you had to be a human on Second Life, and you had to wear a business suit. So no animal avatars, and that was the most that we had to say about conduct online.

Katherine: And of course, it's evolved quite a lot since then but I really think the evolution of saying the internet is ubiquitous and it is current, we cannot wait to respond and go through all the normal, or sorry, the prior channels of getting 15 levels of clearance the world moves too quickly and those days are kind of over. So how do we trust our people to say, let's engage and be ourselves and try to connect across geographies and across boundaries in new ways and I still think that promise is there and I think, Arab Spring, before that you had FUCK protests, lots of social movements that were organized online and I'm very supportive of those who are noting that it is not a coincidence that the public comments where a lot of that social organizing and a lot of that those movements happen are then taken over and what happens with those conversations and where can they go? 

Drew: Right, right. Well, we may well come back to this. The last thing before we move full throttle into Measurement lab is data.org or data.org. You are the chief operating officer of data.org, what is data.org? What is or was and what's the status of that effort in this Civil Society initiative for open internet data? 

Katherine: That's a great question. Yeah, so data.org still exists. It initially started as data science for social impact, and so really the goal was to recognize that data science is a growing field, one of the popular fields of study and inquiry, and that there are many lucrative careers that will attract data scientists toward really important applications like stock trading or delivering packages faster or marketing ads, and those all have their place in society and they have a purpose, but there's a lot of really interesting problems out there there are clean water infrastructure or solving disease outbreaks or getting vaccines and medicines to people faster, that's the delivery that you're looking for. And so it was really this effort to say, how can we take this interest in data and this interest in data science and direct it towards social impact? Very similar kind of work that we have at Measurement Lab of Look, we have a lot of these resources, how do we apply it for the public interest and the public good. So that work still goes, it goes on and the team is only growing. I think they have a few openings, so if anybody's interested in working in projects like that, please check them out.

Drew: Well, obviously you've moved over in January of this year, you became the director of Measurement Lab and Measurement Lab's been around 15 years. I actually remember it quite well, we had a chance to talk a little bit about overlapping history and interest and the way, I've been following broadband data through broadband census, and obviously now, in the work we're doing here with Broadband.money. Talk a little bit about the history of Measurement Lab, just the short version of what it was initially and how it's evolved, like has anything kind of materially changed besides the fact that it's no longer run out of one of the three parties that brought together, you have a separate home, but tell us just a little bit about who it is and what it is now, and your role there, Katherine? 

Katherine: Oh, sure, thank you for that. So, measurement Lab actually started as a, still intends to operate as such, it started as a coalition and then a collaboration between different perspectives of what it means to have a better and healthier internet, so we have three founding organizations; one is New America Foundation, so that represents the public interest, what is in policy and public good; we had Planet Labs running out of Princeton University, so the research academic side, how did the internet form, what is its edge, where is it going in the future; and then the private sector and technology side, so that's Google, actively building, using the internet and relying very much on its evolution and growth. And the recognition there is that there isn't a shared language that is used and there isn't a shared understanding of what is actually happening with the internet, and what's happening online.

Katherine: And so, three different organizations and groups of people within those friends came together and said, how could we have a shared basis of information and understanding. That's the foundations of Measurement Lab as a project, and so the principles behind it were, let's make this an open data project, let's use open source tools so that we can operate in a public way, so that other people can benefit, and we have this sort of commons and awareness of what's happening in the internet and that sustains to today. I would say that that's the ethos that runs through. As you mentioned, we've gone through different iterations, we were housed at New New America for a while as a project, and then spun out into our own independent nonprofit entity, and so that's under an organization called Code for Science and Society, and they have a lot of other really great open source projects, that I recommend you all check out.

Katherine: But so we're still operating as a nonprofit open source, open data project and we really try to think of it as a project because there are a few of us who work on it every day, but we really rely and need the input from many different people, organizations, individuals around the world who support and sustain it and improve it, which is done through code, just people taking tests every day and then we are also fortunate enough to have some funding, but also just people seconding just some of their time, which we have from Sorbonne University, we have from Google's engineering team, and we have some bits and bobs from different people who are able to lend time here and there. So, M-Lab has always styled itself as a consortium, I think we would like it to be a larger consortium, which is really reliant on organizations seeing the value of having publicly shared information about the internet that everybody can build off of and plugging in. And so, one of my role as the more recent director is to try to build that out, yeah.

Drew: Well, and I should point out that we were very fortunate to have, 15 months ago, the then executive director or recent executive director, Lai Olsson and Dustin Loup of the National Broadband Mapping Coalition, who's now with ready.net, and the two of them spoke at length about this whole issue of data, of broadband data, of collecting broadband data, of using broadband data. And that, of course, is at its core, as you've just discussed, the speed data, the other types of data that's collected and let's talk a little bit about that. We've got a number of questions that I don't wanna get to them, but to set the stage, what data is collected, and if I go online and I search test my internet speed, well, if I'm using certain browsers like Google's search engine, it will direct me right to Measurement Lab right there, and I can take a test and it tests my up and down speed and my latency and I believe jitter, could you just speak about what is being tested when one does a Measurement Lab test? 

Katherine: Yeah, that's a great question. And mostly what people will see if they go, as you mentioned, is you'll see up down, and you'll see the latency of your connection as well. So what measure at Measurement lab is actually the amount of data that can be received through interconnection points, so I do think that one of the questions that we have on the chat is how are we different than, say, other speed tests or other measurement tests? And I would say that the difference is where we sit on the internet and so we really sit at interconnection, at the transit layer and that gives us a different perspective of what the quality of internet access should be for someone who is then subscribing to an ISP service, and what expectations they may have or what the ISP is able to provide.

Katherine: So we have that information gets collected, as many people know, we also collect IP addresses, and the reasons for doing this is to give a good sense of where in the world people are connecting, but not at the level of detail that perhaps other providers offer. We also have a few experiments that run on, we have Wihii, traceroute and a reverse traceroute, and we even have availability for a few other experiments. So we do try to, as much as possible, share out the platform in a way that those who would like to understand other kinds of data. So traceroute how your packets are actually getting delivered to you, that can be helpful to understand, it can definitely be helpful for the organization who is providing us a space for our servers to be located to see whether their network is as efficient as possible. So yeah, those are some of the things that we're able to trace and make available.

Drew: So we have some questions, some hard questions here, and I'm gonna dive into one by Alex Betty. And he's basically getting at, do you believe, do you genuinely believe that countywide median M-Lab data, which is the highest level of granularity available, is going to help with using M-Lab data to better allocate BEAD funding? And he goes into some more detail, and you can again, fact check this or correct it if he's wrong. He says that M-Lab and CloudFlare, which I gather is a partner of this, is refusing to publicly share location specific speed data under the guise, is stated, of consumer privacy and kind of again, making the point, he is arguing that how is being able to look up a stranger's home on Google Earth an invasion of privacy? If so, then how is someone's location specific upload and download speed an invasion of privacy? So I'll let you react both on the factual and the normative aspects of that question there Katherine.

Katherine: Sure. No, I appreciate it. Yeah, Alex always happy to chat and we can talk in more detail, you started out in your own very diplomatic way Drew of what is the actual question and then some of the phrasing that Alex uses of like "Is it under the guise of consumer privacy?" So the guise of consumer privacy is still an important lens or list of the word that's used for it, just to be clear it's not that we are refusing, another phrasing that was used, refusing to publicly share location, we aren't refusing... We don't collect that. So nothing that we collect do we keep private, and the reason that we don't collect that is because we've chosen and have talked with researchers and talked with others to collect IP addresses instead and make those available.

Katherine: Those and the way that we collect it and the way that we use it it's not possible to de-identify a specific person. Now, it is an active point of discussion, again this is a project, M-Lab is, it's an open project, if we are hearing from researchers and for users that it would be far more helpful and in the public interest to collect a different kind of data more specific latitude and longitude we'll collect it. Now it's not quite at the county level, it's more about the city level is about as accurate and local as it gets, then we can shift this, but what's lost if we then stop collecting IP addresses is the question that we're currently actively in, we held a community call about it last week or two weeks ago and would welcome feedback and participation with those who say I'd like to keep it as is.

Katherine: I'd actually like to have this far more granular data. User privacy is very real, and it's... I think everything in the public interest is a trade-off between individual privacy between progress and growth, but we also have to think of human rights issues and something that may seem innocuous if you're in as I am in New York or in Nebraska or in California, it's gonna look very differently around the world, and we are a global organization and so we have to think about how our work shows up at scale.

Drew: Well, the the other kind of piece I want to pop back into this, I think you were referring to the question that was asked about what are the differences between Measurement labs tests and other tests including Ookla, which obviously has a extremely robust database as well of speed tests, CloudFlare and and others. So like on broadband.money, the speed tests that many state agencies are using will actually do multiple tests. Will test using M-Lab, using CloudFlare using Ookla, what are the differences? And here's your chance to... This used to be a bit of more of a religious conflict between Ookla and M-Lab and everyone's mellowed on that point, but I know you do have a different methodology on tests, could you speak to that difference in methodology and how that translates to Alex's question about data and the specific of location data? 

Katherine: Yeah, yeah absolutely. And yeah, I was the... The religious era preceded me, so all I can say is that we've had great chats with Ookla team, they seem really nice, so we're very happy to work with them, they've got some new people on some new researchers who we've been in workshops with and been working on policy with. And I think that there's... The world of crowdsourced information and data is needed and it's small enough that there's room for everyone, but practically because there are different tests, it's genuinely the vantage point the location on the internet, so there's a few different ways to think about the differences, one is server placement. So the way, this is called is off-net or on-net. Off-net means that you're outside of the ISPs themselves and so M-Labs places it servers in off-net locations. Ookla as we understand it we do not work for Ookla cannot speak to their their process specifically but they are on that, so they work directly within ISPs. CloudFlare is on CloudFlare servers and Microsoft, I believe, are on their servers, so there's... And I think Netflix is fast as on it's servers, so it's just it's where you place the servers is one of the ways that you're viewing the internet. So then there's also methodology, M-Lab is a single stream TCP connection, and we measure, you asked what we measure, we measure that with as much as possible.

Katherine: So it's bulk transport capacity, and I can sort of write out as engineers have tried to help me better understand IETF RFC 3148, so for those who can follow these but that's a difference. Ookla actually uses Multi Stream and they're measuring link capacity. Link capacity is often associated with what service am I paying for? So that's the closest association. I know that there's perception that often that is more objective as a measurement, that's I think that's the FCC's terminology for it, but there's just, it's just that, it's not that one is better or one it's worth they are just worse, it is just different or different ways of understanding what's happening on this very vast and complex architecture called the internet. And then, sorry just to...

Drew: Finish.

Katherine: The geolocation granularity as has been noted by Alex, by others, Ookla's is far more fine grained because they do not make the data public. They use HTL5, another methodology of surveys to get more detailed data, and our data is publicly available, theirs is closed and proprietary. What we would say for BEAD and for anybody else who's trying to understand broadband infrastructure, where investor should go, what policy should make decision should be made is use both, if you have the capacity to do so use as many tests as possible. I think more data and more information is better, it will lead to better decisions, and this is a very complex topic, and so having more insights not just about speed tests but about other aspects of your internet infrastructure and how people use it and what they need it for is really going to give you the information that's gonna help you make the best decision.

Drew: We also have another critical question, it's actually probably the most politely worded critical question I've ever read from Brett Glass, he says "I'm an electrical engineer who helped bring the modern internet in 1983 and founded the world's first wireless internet service provider," he says that "So-called speed tests do not measure quality, cannot accurately measure capacity or bias in favor of large providers, make it easy for unscrupulous providers to cheat and penalize ISPs who maintain consistent performance and control prices and costs by regulating usage and preventing hogging and abuse," and then he goes on to state/argue that "Those who excessively run these tests," and I definitely fit in that category, "Who obsessively run tests, are actually congesting networks and hurting performance. So what's your reaction to that medley of complaints there Katherine? Oh you're on mute though Katherine.

Katherine: I did it again, okay. So the statement here, "speed test do not measure quality," is completely correct. I think that's interesting is that there is not a shared definition of what a speed test even is, and that's again one of the reasons that we're running this internet quality barometer is that we have all of these different things that we measure, we measure bufferbloat, we measure jitter, we measure latency we measure speed, but the question really is what does it mean as a person? How does it affect your experience? And so we think that speed and we kind of hope that speed and all these things that we measure serve as a quantitative proxy for your lived experience, but we can improve on that.

Katherine: That's really an active area of research and study and very clearly a debate, and so we would never say that, if you have this speed then your internet is fantastic, and you're fine. It is just, it's an indicator of what your experience is and Drew since you've developed your own speed tests I'm sure you yourself have a lot to say about this topic. As far as the phrase about edge monopolists and persistent promoting these, I think it's probably a facet honestly and, I don't know, I can't speak to who's promoting it, but I think it's... I think the internet is and understanding how it works is complex and yet, it affects all of our global lives, even people who aren't connected decisions about their lives are being made because of the internet.

Katherine: Because of satellites that are monitoring where they live, how they move, how the water is receding and reaching them. The internet affects us globally and I think it is a complex topic and people who are trying to say, "Oh this is really easy," can kind of do it as disservice but, it can also be extraordinarily anxiety producing for any individual, for any policy maker who is charged with governing people's lives to be faced with a topic that is difficult to understand and so I think people can understand speed, it makes sense as we try to use that as a proxy. Ideally more people would understand all of these other aspects and facets, but there is a lot that to be educated on so that we're able to better engage in the world around us, and so I think that we are all just trying to get to a place where people can be more informed and make better decisions.

Drew: Well, and I think I told you that when we started broadband census and the data campaign around it, I often use the term the broadband spark for speeds, prices, availability, reliability and competition and measuring these other dimensions are so important, and reliability if anything has gotten so much more important, so like you can view that as like okay downtime or do the text care to fix problems that go down like Detroit where the cable network went down for 45 days without getting fixed, or you can view it as simply like the fact that, you can get the the connection you want consistently lets you respond to my little rant there, but do so in the the context of David Tate's wonderful question. David was also an Ask Me Anything guest and talked a lot about Bufferbloat and all the other things that he's super knowledgeable about but he asks this one. Any chance we will see a test that measures upload and download at the same time? So what do you think of that question from David as well as this point about reliability, and how we measure reliability better? 

Katherine: So I think those are two separate questions. So how you could measure upload and download at the same time I think that's... How you can physically do that, is not something that I understand, I believe it's an area of study, and I think that's and if I'm wrong about that then come back at me or I'm happy to confer with people much smarter than me about that...

Drew: Well David has responded.

Katherine: Oh, great. [chuckle]

Drew: "There are some excellent tests now that test up and down at the same time like Crusader, none of the tests mentioned so far do that, so I finally have to ask this question." So anyway, that's his his addition there.

Katherine: Okay, understood. I think what I've been asked is that, is that an experiment that we can help run, that we can connect with on running? I don't think, I know that NDT... One of the things that we're trying to do is integrate what we call measurement Swiss Army Knife, which is our own Multi Stream test, but I don't believe it's up and down at the same time so I'd have to understand that a bit more. I think as far as the question about reliability it's very interesting because M-Lab, like Ookla and like other speed tests, is really a moment in time test or a point in time, and it only happens when somebody pushes a button and says, please test. We have run so how do we address that? 

Katherine: Because the internet fluctuates and what might work in the moment, your connection can drop when you are trying to take a class or be in an interview or talk to your doctor. I think it's a very different issue if internet is offline for 45 days or you're in a place where the government is intentionally shutting off the internet, but in general understanding the quality over time and not just a moment in time is quite important. The way that we've done this thus far is through measurement devices that are placed in specific locations, so we have something called a Murakami device, which we've placed in libraries, we had a project with IMLS to do so. And then that can give you a very consistent perspective about what is the quality of the service that you have what is the quality of the internet and how steady is it? 

Katherine: If you'd like others, I'm aware of UCSB has one called Net Unicorn, UChicago has one called Netrics, and I'm sure there are others who are developing these kinds of devices, but I will just put a plug myself and others who are interested in this space, there is a lot of investment that's happening right now in the US on infrastructure, there's these BEAD funds and in the next three more years there's gonna be a lot more investment in infrastructure and it will be very helpful if we have consistent measurements that show the impact of that investment. So we're getting better baselines right now, but if we wanna show the change over time, what worked, what worked well, what could have been improved, those network devices are much more effective than really anything else that we we have available. And I'll just offer that if we do that well and we collect that information we understand what works well in different contexts in more rural areas, with different kinds of connectivity then that is insights that we can bring elsewhere in the world, 'cause it's not just the US that needs better internet, everybody does.

Drew: Right right. We'll give one follow up for Brett Glass as well, he says his ISP specializes in reliability and is penalized for it, because we've prioritized that above getting the highest number on the speed test, is there any way through, I don't know, like a number of tests over a multiplicity of time in which you could have a have a metric for reliability? What ways have you thought about a reliability measure component to what M-Lab is working on? 

Katherine: Honestly, for me it just goes back to having a measurement device that measures consistency over time. Now if you... You could definitely imagine having multiple tests integrated and running, you could also do this on mobile phones just through an app that's constantly testing, if we can make it not drain battery life not saying that we would I'm just imagining, but I think it's a great provocation of how do you measure that? And if you're really only using a point in time you're really, you're going to get that snapshot in that moment and connectivity fluctuates.

Drew: Right. Scott Woods of Reddit asks, why are some but not all providers so against open data and publishing broadband speeds? What extent have these efforts hampered Measurement Labs advocacy worked and what's missed? Or what are they missing I guess from your perspective Katherine? 

Katherine: Well, Scott I really appreciate it, thank you for that. As Drew you know from my bio I've been working on open data and primarily it's for governments, open government data for about 15 years and I think it's really helpful and important to share information especially in subjects and for projects that are in the public interest and for the public good. It gives us not just a better sense of what information everybody else has but also helps us bring out better solutions, we have more creativity, more innovations, more ideas are able to flourish if we have that shared knowledge and resources, but so why do people not want to share it? I think there's often a a fear that there's going to be criticism and maybe backlash, and I don't know how often those things are founded, what we try to say is that if there's a few reasons why you would want to there join our network, make your data open, one can just be that it can improve trust in your product. If your product is very good you would want to show even more how you're doing so much better than anybody else's, and so that's one way that if you're more transparent about the information you have you can engender trust, also it's just a lot easier for people to find out more about what you're doing and and and how they could benefit.

Katherine: We also just try to appeal to the business case of the more you operate in the open, the more you're able to catch flaws. And you can also just rely on a community of users and a network to see what those shortcomings could be, how they could be repaired or fixed. And we find there's a lot of benefits. And organizations are even more efficient when they're deciding what they can make better. They have to be more organized in their information internally. And then finally, there's just a general appeal to knowledge and information for shared commons and for a better good. And we all want a better internet. That's a fairly universally shared global goal. And the way to get there is to better understand what is the current status of the internet, where is it working well, where is it not working well, and then better at diagnosing what could be improved. So if that is a shared interest, then we need shared information to get to that.

Drew: Let's drill specifically into BEAD and how Measurement Lab's tests, and data set can be used in BEAD. We have a number of great questions on this. One is here from Jake Neenan. Well, actually, that's a separate question. Let's actually ask the question about specifically BEAD and how can Measurement Lab's data be used for BEAD implementations. In other words, what's the role of state broadband officers who are watching this, Katherine, in using M-Lab data? What should they be doing right now to get this data and make better decisions in the BEAD implementation process? 

Katherine: Yeah, so the... How to get the data, the data is available for you to use at any point that you would like. So that doesn't require special access, but you're always welcome to reach out for us if you have somebody who is particularly adept at SQL, and who likes to play around with BigQuery and matching up of large data sets, it is helpful. Otherwise, we're happy to support. I think what we would say for the people who are trying to reach their BEAD process is using the data to actually understand what infrastructure you have, what's the quality of connectivity in your state. But truly, in order to reach the requirements that BEAD has, you'll need to have an integration. There are specific metrics that BEAD requires for speed, such as location. They require that people measure three times over the course of three days. And any of those has a lot of need for community engagement for outreach, for ensuring that you're getting more detailed information, particularly from those communities and spaces that aren't always represented, and are looking at the current NTIA maps and saying, how that's lit up isn't my own experience.

Katherine: And this is one of the reasons that we really advise and promote having multiple different vantage points and multiple different tests, because you're going to get better information about what's happening in your state. And I would just offer that anyone that says, "If you just do this, you'll have all of the information that you need," it's not true. There's no single source of truth of what people's experience is and what's happening online. And I'd also say that with the benefit of and the need of working with M-Lab and M-Lab's approach is that, in this moment, we are essentially having about 65 different responses to how to understand internet access in our states and our territories. And what we're going to come out with is 65 different kinds of maps or data sets. And how are we then going to use that as an improved public country-wide map of what the status of internet access is and how we're improving it? 

Katherine: And so whoever you work with, I would hope that all the state broadband leads work in a way that they're sharing that information publicly and interoperably so that we can have that information cut across. Otherwise, it's a huge lost opportunity. And we also had just shared yesterday with a few of the regional states and territories the diversity and equity teams, which I see as another question here.

Drew: Well, let's actually move into that. That was my next question I was going to ask. Adam Puckett asks, what do you see as the role for M-Lab in regards to digital equity? That topic is getting a lot more focus because of BEAD and the broadband equity access and deployment focus. "But it's important and long overdue," says Adam. How does M-Lab think about its role in this conversation? 

Katherine: Yeah, we really appreciated talking with the diversity and equity officers. They have a lot of... They have a big role that they're in and a big role ahead of them. And like so many civil servants, are overtaxed and under-resourced. And I think a few of the things that we would say are, there's a lot of push right now to make sure that the funds are allocated where they need to be today so that people can get internet access or improve their internet. But we also want to build for three years down the line, five and 10 years down the line. And thinking about what kinds of data you would want to collect to better understand what people's needs are. Because sometimes it's just basic infrastructure. A lot of times it's data literacy. How do I use these devices? What devices do I need? How am I updating them? Do I have malware? Is that the reason that I'm not able to get online and get the connectivity that I need? And if we're only asking questions, writing our surveys from the specific vantage point of, where am I going to send my funding so that I can build or improve infrastructure, then we're going to miss some of the opportunities that we have to support communities to sustain themselves in the longer term.

Katherine: And so I think that's... I don't know that that is an M-Lab specific perspective, but it's definitely one that we're interested in supporting. And I truly would just come back to the principle of openness and transparency in that work. It is very frustrating to see that map of where connectivity should be in the United States and realize that your community or your neighborhood or your street or your house says that it's getting an experience that it's not. And so the only way to really counter that is with better information and to really advocate for it is with better public information.

Drew: So the question I started to ask, but kind of want to come back to is from Jake Neenan. Do you see any use for Measurement Lab data in enforcement, in ensuring providers build up to the speeds they're obligated under awards? What is your thought on that? And again, maybe you could also speak to some of the other programs, the FCC and USAC have going to measure speeds because there are some of these going on, but the question specifically is how M-Lab could impact these enforcement discussions.

Katherine: Yeah, it's a very interesting proposal. I don't... From enforcement, I would say I really welcome accountability. And there is some precedent for M-Lab data being used for accountability. In New York, there's a legal suit just that was run a few years ago that people might be familiar with for some ISP accountability. I would say we're ultimately interested and supportive of any data collection. I use the process that ensures the publicly available baseline of current status and then monitoring this over time. Again, I said that the measurement devices are probably the best and easiest way to do that. And I think the accountability there, it's really, it's a collaborative process with private sector, with government, and civil society to ensure we're meeting these goals.

Katherine: So there is definitely... Sometimes accountability and enforcement needed of, did you meet the terms of the tender? You have a contract, did you build the thing that you said that you were going to build? But it is possible that we will build more infrastructure and still people won't have a quality internet experience. And they may not have a quality internet experience because they're using the wrong software, because they have the wrong device, because they have malware on their device, because they need different kinds of literacy. So I think the question is, are we having a whole of internet approach? Are we having a whole of user experience approach? And if what we're trying to get to is quality internet for all, then that has to be the ultimate and underlying goal, not just are we checking boxes on what we said we would do. And if we didn't get the outcome that we were going for, well, it's not my fault. I think we're really trying to get to what is the world that we're working towards and trying to build.

Drew: Right. Dr. Ron Suarez asked a question about, do we need changes like open access to ensure competitive opportunities for more affordable internet? Or is the program, the Affordable Connectivity Program, which may be regarded by some as welfare-like subsidies, what is your view on some of the ways to address the affordability and competition questions? Does M-Lab have thoughts on that? Do you have thoughts on that, Katherine? 

Katherine: I think it's more of maybe my self-thought than M-Lab thought, although that's another aspect that's going to go into the internet quality barometer work that we're doing. There's a lot of great researchers that we work with who think about affordability, accountability. There's some really interesting work that's out of markup. Recently, there's work that we did with Consumer Reports on what are people getting based on the service that they pay for. There's a professor, Arpit Gupta, over at UCSB who's doing his own research on affordability. Affordability is vital. I think I've heard it as sort of the other leg of net neutrality. If it's impossible for people to actually afford to get online, then you can't say that you're delivering a fair service. I bristle a bit at this idea that providing affordable internet for people who wouldn't be able to get online otherwise is charity. One, we provide plenty of subsidies for our private sector at much higher rates than an individual household. But I think that it's an investment.

Katherine: If people have high-quality internet access consistently, and they do not have to worry about this utility, which I think it is, I think there's work that's being done on access as a human right. It is too much integrated into our lives to consider it as a luxury and that a subsidy is not something that's an investment for all of us. So I do think the question of it is important to understand what truly does it mean to be affordable. And I think would be a lot of benefit to tracking the impacts of the programs that are currently in place. And just from the conversations I've had with the people who are working on insti-rating this, they would love to do that themselves. But like so many things, there's a lot of over-demand and under-resourced for staff, for teams, for all of these. So if we want those things to continue, we need to invest in them.

Drew: We have a question from Jace Wilson. M-Lab plays an extremely important role in understanding, in advancing understanding of the health and reach of the internet. What ways can organizations that use M-Lab tools support the work of M-Lab? And how best should they do that support? 

Katherine: I appreciate that. Jace is definitely an organization that invests in and supports, works with M-Lab tools. I really would regard M-Lab as a public resource and a public knowledge base. And I think like any other open source and open data resource, the question is, if you use it, are you investing back in your supply chain? And if you're benefiting from it, especially for those in the private sector who integrate M-Lab and benefit financially, we're a nonprofit, we're an open resource, and like so many, operate on a shoestring, just with very passionate people, to try to get a resource out there for others to be able to use. And because of the way that we operate, with making the data public, we don't always know who's using and who's using M-Lab data in their integrations. And our mission is to make the data available to anyone and for it to be useful. But I would just... The quality of the data, the diversity, the reach, the global reach of it would be improved with more support. And that can come in the form of fellows, it can come in the form of seconding time from a team, it can come in financial support, it can come in more experiments, more engagement. And the same is true if that support doesn't exist. Then the platform will shrink. I mean, that's sort of that story.

Drew: I want to come specifically to another question David Tate asked about videoconferencing. So we had the exchange about upload and download speeds at the same time. What about videoconferencing metrics? And again, what does the landscape look like in terms of what other entities are doing with testing video? And how can M-Lab contribute to that metric? 

Katherine: Yeah, I'm not... I guess I didn't this question specifically, but...

Drew: Is there a way to test video as distinct from the connection? 

Katherine: Sure, and actually, I don't know how much we say this, but we've talked with groups like Domos that specifically are testing video. And so organizations that have a specific interest in those kinds of uses and what those outcomes are, we try to partner and collaborate with them, see where our data can be useful, and then see as a reciprocity where they can share information more publicly so that others can get access to it. We don't have a... I don't have an affirmative forward from them yet, but I hope to. But yeah, I think if the question is, what are we interested in? We are interested in all aspects of what the performance is and what the function is of the internet and how to improve it, please.

Drew: Maybe we could talk just a bit more about that Swiss Army Knife approach. And another question from Benjamin Kahn is about bandwidth is often touted as the primary metric of internet service, but latency, packet loss, and jitter are also critical. So what would be the four or five or six top items that need to go into that Swiss Army Knife toolkit? 

Katherine: Yeah, so I'm going to... I think bandwidth is a strong starting point, but other metrics that he mentions, like the packet loss, jitter, they have other impacts on, as Dave mentions, on video streaming, video conferencing, on gaming. So I think everything that we try to do and what we say is the priority is really how it affects a person's experience. Again, we're really hoping that M-Lab and all of the measurements that we run can be conserved as a proxy for what people want to use and seek to use the internet for. So that's remote learning, working, telehealth, entertainment. I think it's also important to know that like how you measure these changes. There's not sort of a clear way that you say, now I understand what bandwidth is, just the same way as you don't, or broadband is just the same way you don't... You can't say, now I understand specifically what speed is.

Katherine: So again, engineers in my team would share that latency can be measured is latency under load. So that's how well a user will experience congestion, but it's not accounted for, for example, in bandwidth or unloaded latency. So I think, I mean, it's a great question of what is the most important priority? And I think it just comes, the diplomatic answer is it depends for anybody who's ever worked with a diplomat. That's your, like the most reviled phrase, but it happens to be the case that depending on what question you're asking, you're going to come at what information you need from a different direction. It's going to give you a different kind of insight. And so the best that we can do is try to bring it back to a proxy for what a person is experiencing and share the information that way.

Drew: All right. So this is my penultimate question and it's about the future of M-Lab and the future of like, what does success look like for M-Lab? You've been around 15 years, but another 10 years down the road, what does success look like for Measurement Lab? And with that do you see what's the future of the datasets that you're collecting. What new things do you imagine would be measured in 10 years? So anyway, that's kind of a two part penultimate question.

Katherine: I love that. I love the future visioning. I think as I mentioned in the beginning, M-Lab has really focused on transit ISPs and interconnection points. We are now expanding into cloud and have the cloud view of the internet. We're also talking with access ISPs so that we can also expand our own vantage points, better understand what's happening and how to measure online. So I think in the future, we'll have even more data to be able to share. Again, that does still come with, do we have the support to be able to pursue all of that ambition? But that's the intention that we have. I'm talking with quite a few partners about that potential. And that's all under this consortium model, a more flat partnership-driven structure to the shared project. I also would really like to see M-Lab data being used even more globally. We have partners, we have researchers around the world that really rely on M-Lab, and I'd like to see these kinds of discussions that are happening in the US about BEAD and the infrastructure happening there. The US, as I've mentioned, is not the only country that needs a better understanding of its internet quality and better internet infrastructure.

Katherine: So how might we be of better support to others around the world that are trying to improve their internet and are trying to improve it for all of the reasons that advanced communities and advanced societies. And I would just say that we're not precious about it has to be us. [chuckle] I think just the principles of M-Lab are sustained. Open information, open data, open platforms that people can connect with and support and share. That's really our true interest. And seeing that happen in a multi-stakeholder and collaborative way is what we hope to contribute to and have as an interest that's ongoing.

Drew: All right. Our final question comes from Ben Khan and it's about what does open internet mean? What is free and open internet? Is it about privacy? Is it about net neutrality? Is it about access to devices? Is it about Libya not cutting off the internet? What does it mean to you and what do you think its greatest obstacle is? 

Katherine: Very, very good question. It's also a fun one because I'm in a workshop right now with NSF with people who are trying to debate that exact issue is what is the future of the public and open internet? So I think it very much depends who you talk to and what their training and background is of how they'd answer that question. And I'm just, I come to it very much from a policy and sort of individual, a person's experience, and I probably really come to it from the internet freedom perspective. So if you have the freedom to connect, if you have the freedom to gather, if you have the freedom to express yourself online and to engage regardless of where you are, your geography, your personal identity, your language backgrounds that you have, if you have that same capacity as anybody else does, then that's what an open internet is. It's a resource, it is a platform and it's a tool that's available to everyone.

Drew: Well, the hour has sped by. Before we thank you and bid you goodbye, we wanna mention there's many, many events coming up. Next Wednesday is the Where's the Funding episode, the challenge of matching funds with Jochai Ben-Avi. That's at 11:30 till 12:00 Eastern time. The next day is actually a broadband breakfast BEAD Implementation Summit in Washington DC, but there'll be a webcast for that, you can learn more at broadbandbreakfast.com. On Friday, a week from today, we'll be hosting Marc-André Campagna, co-founder and CEO of gaiia. That AMA will be at 12:00 noon Eastern time, again, a different time, slightly different from our normal time. And then two weeks from today, we'll be back with Evan Marwell, the founder and CEO of Education Superhighway at our normal 2:30 PM Friday time. With that, we want to give a huge thank you to Katherine Townsend. Thank you for spending this hour with us and thanks for being part of the broadband.io community.

Katherine: It's a pleasure, Drew. Thanks for bringing us together. Really appreciate it. Thank y'all.

Drew: Bye.