IGF 2021 - Day 1 - Lightning Talk #69 Measuring Internet Resilience and Shutdowns: An overview of Internet Society's Pulse platform

The following are the outputs of the captioning taken during an IGF virtual intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.



>> SUSANNAH GRAY: Great.  I think we can get going then.  So welcome everybody to the IGF 2021 Lightning Talk on Measuring Internet Resilience and Shutdowns, An Overview of the Internet Society's Pulse Platform. 

     I'm Susannah Gray, Director of Communications here at the Internet Society, and I will be moderating this session today.  I will give a brief overview of the Pulse platform and then I'm going to hand over to my colleagues Kevin Chege, Mat Ford, and Hanna Kreitem who will dive deeper into the focus areas and how you can use that data in your work.  We'll have some time for questions at the end of the presentation so please feel free to type things into the chat and we will answer them as we go along or ask questions when we get to the end.

     Okay.  Thank you, Mat.  So The Pulse platform was launched in 2020 and presents curated internet measurement data from trusted sources to help everybody gain deeper data driven insight into the internet.  This data helps everyone to assess whether the efforts that we all are engaged in to ensure that the internet remains open, globally connected, secure and trustworthy are working.  And it also helps policy makers, researchers, journalists, network operators, Civil Society groups and others to better understand the health, availability and evolution of the internet.  Next slide, please.

     There are several focus areas on Pulse.  Enabling technologies, internet shutdowns, internet resilience which launched in November.  That is a new focus area for us.  And centralization, which is actually launching later this week.  So stand by for more information about that.  And we'll go into more details about these focus areas in just a moment.  Next slide, please.

     So where do we get all this data from?  We actually partner with multiple organizations.  And you can find out more about each of these data partners and the data that they provide to us on the Pulse platform at the link there on the slide.

     Next slide, please.  I'm now going to hand over to my colleague Hanna who is going to talk more about internet shutdowns.

     >> HANNA KREITEM: Hello, everyone.  So at Pulse, as Susannah mentioned, we cover multiple areas of focus.  One of them is documenting shutdowns or artificial internet limitations, as weather spectrum would be.

     We believe we need to understand shutdowns and the structure of the global internet to be able to assess the impact and the damage that they may be causing to the access.  We offer a curated database of shutdowns events on Pulse, and we work with key partners to verify and check and verify events as they occur.

     Our data comes from multiple sources including dedicated observatories like Kida Ayona, Uni, and other services dedicated to measuring disruptions of the internet.  We also have information from Internet Services Providers like Google and Cloudflare who allow us access to see where disruptions are happening to their services, where people are not able to access their services changes over time.  We also follow media outlets that report on shutdowns and events as well as local partners.  Next slide, please.

     Why do we care?  We believe that internet shutdowns harm societies.  They are harmful in general to the society and economies and global internet infrastructure.  And to be able to push towards an open internet that is globally connected, secure and trustworthy, which is our core mission, we need to urge governments and decision makers to support policies that keep internet on and strong. 

     We also believe that content blocking is generally inefficient and ineffective as well as causing unintended damages to internet users.

     So again, we encourage governments to rethink their policies and their options when they decide to cut off access to the internet, limit use of content, as well as limiting access to specific services.  We see that this is important for the internet.  Next slide, please.

     What are we noticing?  Over the past couple of years of documenting internet shutdowns, we have noticed that full shutdowns are less common but more blatant.  Artificial internet limitations in general are becoming more targeted, local and ever harder to detect.

     Common recent shutdowns targeted specific access methods, in particular, mobile data, which is -- which we noticed a huge jump in the use of cutting mobile data versus cutting fixed connection in 2021 compared to previous years.  Limitations are very targeted, are on the rise, and do not always mean cutting off access completely.  They also throttle or hindering specific services.  For instance, one of the examples that we have noticed is throttling Facebook Live streaming around a public square during a protest.  We can see that it is very specific in terms of area, service, and time, thus making it much harder to detect from the outside. 

     This artwork shows what AI thinks in terms of shutdowns which is basically layering different browsers but still unable to see what they are.  And that is -- that represents what we look at when we see internet shutdowns.  Inability of people accessing services they need, and they want in time. 

     You can contribute to our data.  Soon we will be launching a form for you to fill in.  But until then, you can always share information on shutdowns on [email protected]. I will share the link in the chat.  Next.

     >> SUSANNAH GRAY: Thank you, Hanna.  So we're going to hand over to Kevin now who is going to talk more about internet resilience.

     >> KEVIN CHEGE: Thank you, Susannah and Hanna. Greetings, everyone. My name is Kevin, and I will take you through the work we have been doing on internet resilience.  Next slide.

     So with internet resilience what we wanted to understand was how connectivity on African continent is for various parts of the continent.  We wanted to understand which countries are well provisioned, why they are well provisioned, what policies are beneficial and what infrastructure results in having positive experience when users are online. 

     We also wanted to get a good picture of statistics how they compared to internet usage on the ground.  You know, it's easy to present numbers and figures of how good an internet connection is, but how does that experience relate to the real-world experience of a user who is using the connectivity in an African country?  Next slide, please.

     So AFRINIC and ISOC worked together on a project called Measuring Internet Resilience in Africa.  In short, we call it MIRA.  And the main goal was to carry out sustained internet measurements in Africa so as to arrive at two main goals.  The first one being to determine what levels of internet resilience are in select countries.  And before we got into this work, we did a lot of background research to agree on what exactly we mean by internet resilience. 

     When we got to that conclusion of internet resilience as a definition for our project, we decided to seek for existing data sources, try to find a very scalable manner to carry any measurements we need to carry, and also identify which metrics we can use to identify internet resilience in a country. 

     And then we also needed to present this information in a very easy to understand manner.  So whether you are a policy maker or a technical engineer or just a simple internet user who wants to find out how good the connection is in your country in Africa or neighboring country, you can access that information easily.  Next slide, please.

     So using a combination of surveys and talking to a number of experts in the field of internet measurements, we came up with what we call the Internet Resilience Index which is an indicator that measures a country's performance against four key pillars on the internet.  

     And these pillars are infrastructure.  And infrastructure pertains to physical connectivity in the country.  For example, how many cable providers are there, how many submarine providers are there, how many managed stations are there. 

     Performance.  How good the throughput is in terms of speed in the country.  Enabling technologies.  How reliable is the DNS infrastructure, how is the routing IGs.  Are the ISPs in the country using good routing practices?  And finally, the fourth pillar there is local market readiness.  How competitive is the market?  Are there a number of internet service providers or is there a limited number of internet service providers?

     And these four key pillars we arrived at them using, as I mentioned, surveys.  And we also spoke to a number of experts, and we decided to shortlist to the most agreed on metrics.  This is by no means all of the metrics you can use to score the internet resilience in a country, but these are the ones that appeared highest in our research. 

     And just to summarize what the Internet Resilience Index is, it is basically a scorecard for connectivity in a country.  Next slide, please.

     So we are happy to announce that as of last month we have launched what we are calling the resilience page on Pulse, our Pulse platform which has been mentioned by Hanna earlier.  You can go to that site and rely on this section to try out the resilience page and see the various presentations in various countries. 

     You can see the metrics for the country, the whole continent, or core regions.  So you can say I just want to view the metrics in North Africa or Southern Africa and so on.  And the link will be shared as well in the chat. We just launched this, and we are looking forward to getting more feedback on what you think about the presentation of this data.  And we are looking to improve this as we move along. 

     So thank you, and we look forward to receiving your feedback.  And I will hand back to Susannah and Mat.

     >> SUSANNAH GRAY: Thanks, Kevin.  Next up will be Mat talking about emerging technologies and centralization which is the new focus area that we are launching this week.

     >> MAT FORD: Thank you, Susannah. 

     Yeah, I'm Mat, and this is a summary of the enabling technologies focus area on Pulse where we present data collected from a number of different sources on enabling technologies that in our view are critical to improving the scalability, security, and trust of the open internet.

     And so they are technologies like IP Version 6, which is really critical for the ongoing scalability of a global internet that serves close to 10 billion people and many, many more devices.  Security technologies like HTTPS, which secures the web but also many other services as well.

     And TLS1.3 which is the latest version of the transport layer security protocol which is both more performant and more secure than previous versions. 

     We also track the adoption of DNSSEC, both at CCTLD, so country code to main level, and the DNSSEC validation by recursive resolvers. 

     And most recently we added data about the adoption of routing security related technologies.  So the extent to which service providers are signing routing announcements and publishing.  And the extent to which those attestations are being validated as well.  So two different views of the adoption of routing security.  And we hope to add to all of these either with additional technologies or with additional data sources.

     And we hope that this data which we can present over time or by country is useful to local and regional advocacy efforts in encouraging ISPs in particular to adopt these technologies that are really critical for the future of the open internet.

     And there is a lot more information on the enabling technologies page about all of this.

     A new focus area that we hope to launch this month, so watch this space, is on internet centralization.  And we are using measurements of top websites as a sort of proxy for the internet in this case.  So we appreciate that that's a pretty partial view of the internet, but it allows us to get some traction on this question of how centralized are different aspects of internet service provision, how is that changing over time.

     And so we are going to launch this with an initial analysis of seven technology markets.  Those are content distribution networks or also known as reverse proxies, the security certificate provision, SSL certificate.  Top level domains.  Web hosting, data centers, DNS severs and server location.

     And we are looking at -- well, we are distilling all of those markets and all of this data which we have for different views of the top websites so top million, top 10,000, and top thousand, into two different metrics of centralization. 

     So we use a Gini coefficient which gives us a measure of equality or inequality in a market using the market shares of providers into the market.  And we also use the Gini coefficient to look at the equality or inequality of market shares by jurisdiction.  So we weight, we assign jurisdictions to providers in a given market, and we then weight those market shares by the internet using population of those countries.

     And then we derive a Gini coefficient across those market shares.  Again, that gives us a way of a proxy for equality or degrees of inequality on a per jurisdiction basis as well.

     And then we are also using the Herfindahl-Hirschman Index which gives us a measure of market concentration.  So it's a slightly different way of thinking about it to the Gini coefficient.  The Gini coefficient is more of an absolute measure of equality or inequality whereas the HHI, the Herfindahl-Hirschman Index gives a sense of is the market concentrated in a very small number of actors or jurisdictions, or is it distributed.  And it is not necessarily concerned with whether or not that jurisdiction -- that distribution is amongst all of the players in the market, more that there is at least some distribution amongst the bigger players in the market. 

     So as I say, we will be launching this really soon now.  And I hope it is interesting and helps provide some insight to people interested in this question of internet centralization.

     That's all I had.  Did you want to talk about contact points and so on, Susannah?

     >> SUSANNAH GRAY: Yes.  But first we have a couple of questions in the chat.  Could you go back to slide 12.

     And Kossi has a question on the internet resilience index.  Did you just want to see it again or do you need more information, Kossi?

     >> KEVIN CHEGE: Maybe just to add on, so Kossi the index is basically an indicator.  It is a series of calculations that we use to arrive at the score of resilience in a country.

     And if you go to the page you will see a whitepaper there that contains more information about how we arrived at the index and a bit more about the metrics and numbers that we are using. 

     But in summary, it is just a provider score for the country.  We identified the metrics we wanted to use for this stage of the work.  But it is the way we came up with the index is that it can be modular.  So in future we can add or remove a metric if necessary.  I would invite you to go to our web page and there is much more information about how the index was formulated.

     >> AUDIENCE: Hello, Kevin, please.  The first line you have 40, 40 and 20.  Why you don't take information measuring for one three, one three for everything?  Why are you taking 40, 40 like that?  Why enabling is load on capable ecosystem?  Why?  Why are you make option like this?

     >> KEVIN CHEGE: Good question.  So the question is relating to the weighting, how we came up with those numbers. 

     The weighting we used our, based on our assessment of the various metrics.  These are not all the metrics we use to calculate the infrastructure pillar, for example.  These are just here for demonstration purposes.

     So what we used were sort of figures which we arrived at ourselves and that is what we are using to present on Pulse.  But we are working on an interactive way to present this data. 

     So, for example, for you, Kossi, if you would like to adjust the weight so that maybe one is 20% or 30% for something else, you would be able to do that in -- probably from next year once we improve the dashboard. 

     So these numbers were just for our -- from our perspective, but from our analysis we do understand that it is not possible to arrive at a metric that everyone agrees on, so it is going to be adjustable once we are done with the front end.  I hope that answers the question.

     >> AUDIENCE: And last one also.  Local ecosystem.  The proposal to make half and half of a local ecosystem, market infrastructure and traffic localization is the same thing for me, it is not possible to have 60 and 40.

     >> KEVIN CHEGE: The metrics, as I said, we understand. It is probably not possible to arrive at weights that everyone agrees on.  We have made a couple off changes and additions to some of these metrics. 

     So if you go to the Pulse page on -- so if you go to the resilience page on Pulse you will see an accurate breakdown of each and every weight and each and every metric.  And, you know, these are adjustable.  If you think one or if you would like one to have more weight than another one you can adjust.  If one feels like -- for example, market structure is how many ISPs are in the country. 

     So if in your country the situation is different you can adjust it so that you get a clearer picture.  And as I mentioned, we are working on an interactive dashboard so that you can pick and choose which weight you think is more important than the other one. 

     But for the purposes of displaying the data we have collected; we did do a bit of research and we determined that for the purposes of presenting the data we have we need to submit some weights, and these are the ones we are using for now.  I hope that answers the question.

     >> SUSANNAH GRAY: Thank you.  We have A question from Mark in the chat.  I'm going to read it just in case people don't have access to the chat.

     How can national -- sorry, it just disappeared.  How can national policy makers and service providers use this valuable Pulse data in the resilience index to resolve problems and gaps in connectivity? 

     Is the data used to help develop community network projects in underserved rural areas, for example?

     >> KEVIN CHEGE: So I guess that is for me as well.  We just launched this last year, and I like that question because we are looking for various ways that this data will be helpful to regulators, internet service providers, internet users. 

     So, for example, we would like that if research is done using the data we provide on Pulse can shed light on policies that are beneficial in some countries.  So, for example why is it that maybe one country have more community networks than another?  Are there policies there that are good?  Why does one country have better round-trip time or better speeds or a better resilience score than another country? 

     So the data that is there is for use by everyone.  We are looking forward for the different use cases that will be derived from this.  And I think moving on probably from next year since we just launched this last year, the users of this data will be more apparent. 

     We will do a bit of research with AFRINIC and ISOC.  But we do encourage everyone to look at the data there and the questions you are asking is exactly what we are hoping to see for the use cases that we have on Pulse.

     >> SUSANNAH GRAY: Thank you, Kevin.  I think we have a question from the floor, and then we will go to Ose on the chat.

     >> AUDIENCE: Alexander Sutton from Oslo University.  As I have seen on the website, the map and the data is only about African region.  Do Internet Society plans to extend such research for other regions like eastern Europe, Asia, Latin America?  What kind of data maybe do you need or what kind of collaboration do you need to help to extend such researches to other regions?  Thanks.

     >> KEVIN CHEGE: Okay, thank you, I can answer that.  Thank you for your question.  When we started this work, the focus was on Africa.

     We have -- we did a lot of background research to arrive at the methodology for how we can measure internet resilience in Africa.  We did -- once we launched this, we did receive a lot of queries like what you are asking can this be applied in other regions.  We have had a lot of requests for using this in the Caribbean region and we are exploring that. 

     So perhaps next year we will see how to adapt this in the Caribbean region.  Just yesterday we had a call, myself and Amrish who is the expert who helped with coming up with the index and a lot of the work for resilience about how to adapt this to Europe.

     And to answer, it is possible to adapt the work, the research to any other part of the globe.  The index may vary so some of the information we received is maybe in one continent some metrics are more important than others.  Or one metric is perhaps not useful in that part of the world. 

     So all of these are useful and, you know, we invite all of you to have a look.  And if you feel that it is something that you would like to explore, we would be happy to hear from you.  And you can e-mail us at [email protected] and we would be happy to talk to you about that.

     >> AUDIENCE: Thank you.

     >> SUSANNAH GRAY: Thank you, Kevin.  We have another question in the chat from Ose.  And he is saying what has been the linkage between the Pulse data and community networks?  Has there been any challenges yet?  Back to you again, Kevin, I think.

     >> KEVIN CHEGE: For community networks, since we have just launched this I wouldn't say there has been a direct use case yet in relation to resilience.  But we are hoping to get those as we move along the coming weeks and months.

     So we are promoting the resilience page as much as possible.  We are hoping that more people get, you know, interested and, you know, try to use the data to answer some of the questions in relation to what data they've experienced on the ground. 

     So perhaps in the weeks and months to come there will be more information about how this will directly relate to assisting community networks.

     >> SUSANNAH GRAY: Thank you, Kevin.  We have one final question because we are almost out of time. 

     And this is another one from Mark.  On shutdowns, is ISOC hoping to create more transparency and accountability of shutdown interventions by governments?  Hanna, over to you.

     >> HANNA KREITEM: Thank you, Mark, very good question. Yes, this is virtually our aim.  We want to provide data and information for people working with the governments, people working with different parties that are related to the internet governance model basically different stakeholders we are working in terms to the internet to provide them with data and information that they can use to assess the impact of the interventions and to also promote accountability wherever possible.

     We as an organization that believes the words technical issues, that is our job to provide the data and information and put it out there for everyone to use towards the benefit of the internet.

     >> SUSANNAH GRAY: Thank you, Hanna.  And so now we are right at time so I'm going to close the session. 

     Thank you very much, everybody, for attending.  If you want to get in touch with us, please e-mail us, [email protected].  You can also follow us on Twitter @ISOC_Pulse.  Follow us for the Pulse mailing list.  We send out the latest news and updates just once a month and we have a shiny new video.  Click on that, have a look and see more information about Pulse and how the platform works.  Thanks very much and enjoy the rest of the IGF.