 Welcome, everybody, to our next talk, UNI. Let's fight Internet censorship together. When you think about Internet censorship, I guess that a lot of people would just have an association with countries like China or Iran or Saudi Arabia, but not a lot of people are aware that filtering and blocking of websites and Internet censorship also takes place in almost all the other countries around the world. And this talk will educate you about this. Our speaker is Aturo Philastro. He is the founder and developer of UNI, and he will tell you more about this super interesting project of his in this talk. Please give him a warm round of applause. Thank you, Miriam. Hi. I guess I would like to start with thanking you and sort of expressing how much this is an honor for me to present this project in front of an audience like CCC. I've been coming to these events since I was just over 18 years of age. And so I think this project and also a large part of my adult life is also due to the encounters and the things I have learned at events like this. And so to get right into it, I think the first thing that we should start with when talking about a project like UNI is understand what we are talking about. So to frame what we mean when we say Internet censorship, because these there is some video connection. Okay. Yeah. So what is Internet censorship? Within the context of UNI, when we talk about Internet censorship, what we mean is some form of network interference that comes with intent to disrupt or distort the reality of the Internet from the perspective of a country or a network. So it means that it requires intent from the sensor to create this altered view of what is the Internet. Another term that we use to define Internet censorship is a filter net. And what we will see during the course of this presentation is that there are in fact many filter nets around the world. And that depending on in which country you are in and depending on the regime or the ideology of those in power, different views of what is acceptable and not acceptable to be done on the Internet are enforced. And so UNI or the open observatory of network interference is this network measurement project that was started back in 2012. And it's a community-led project that's composed of a network of volunteers that install this software and collect evidence of Internet censorship around the world. Since until today we have gathered measurements from over 200 countries and we have measurements coming in from tens of thousands of different networks every month. But the sort of founding principles around UNI and the reason why we decided to work on this project is that we want to sort of bring to the field of network measurement and in particular the field of network measurement applied to detecting network interference and or Internet censorship the principles of openness and transparency. So UNI is open methodologies which means that all of the tools and techniques that we use in order to examine Internet censorship are openly documented and specified sort of like in RFC format. And these allow people to validate that what we are claiming to do is actually reasonable and fair. The methodologies that we specify are then implemented using open source software. So you can verify that what we define in the specification is in fact consistent with what is implemented and all the data that is collected by the thousands and tens of thousands and hundreds of thousands of volunteers around the world are made available openly as open data. So anybody can use this data for their own purpose and sort of from the beginning to the end of the chain validate that what we are doing is in fact correct and reasonable. And hopefully by doing all of this we will reach at some point a stage in which we can say that what we are doing is in fact reproducible research so that every step from the definition of what you do to the implementation to the publication of the data is entirely reproducible. All of our tooling like all of our backend infrastructure even our sys admin scripts are made available openly so anybody can set up their own uni if they wish. But before going more into what uni is I would like to give you a taste of what is possible with this sort of research as in what are the sorts of things that can be discovered by collecting network measurements around the world. And these are obviously just some small samples. You can learn more about each of these case by reading our full length reports. But this is sort of more to show you what is the kind of impact that a tool like uni probe can have in the world. So the first case I would like to talk to you about is the one that maybe you have heard in the news or maybe not because it was not covered as well and it's around the protests in Ethiopia in December of last year 2016. These protests sparked in a particular region of Ethiopia called the Oromo region and they are around some conflict related to the expansion of a certain city of the borders of the city of Sabada that would in a way sort of basically kick people out of their homes because they would suddenly have a big house like a big city in where they live. And so this led to some very serious protests that were very violent and the police was also very brutal to the protesters and it led to many deaths. And as a result of these protests also some network interference was done. So in order to stop the protesters from communicating and organizing WhatsApp was blocked. So the protesters could no longer coordinate or organize themselves. And with uni we gathered evidence of this so we could prove in an undeniable way that WhatsApp was blocked, that deep packet inspection technology was being used and we were also able to understand precisely how it was working that allowed us even to give some advice of what some site operators could do to even circumvent it in some way. And many websites from media outlets, human rights websites and obviously political opposition websites were found to be blocked during this period. Following this in Ethiopia the state of emergency was declared starting from October 8th 2016 just until a couple of some months ago August 2017 so lasting 10 months. But what's interesting from this sort of case is that what we found evidence of is actually blocking and censorship being implemented before the state of emergency which in theory was something that was not supposed to be allowed. And together with Amnesty International we published a research report outlining our findings and sort of providing evidence that this was something that was in fact happening and was clearly not something that a clear human rights violation. Another interesting case that we worked on together with our partner, our Malaysian partner, CNER project was published in May 2017 and it was sort of a study of internet censorship in Indonesia. We found that many sites of various categories were blocked but I think one of the most interesting things from this particular report was the fact that we collected measurements from around 60 different networks and based on these we were able to see that there was actually very much some difference between the different operators in terms of what was being blocked and what was not being blocked. So as an example Vimeo and Reddit were found to be blocked on certain internet service providers in Indonesia although the ban for blocking such websites was supposedly lifted two years ago. And this stems I guess from the fact that the Indonesian legislation states that internet service providers are granted the authority to inhibit or block content that is considered negative content and they are left on their own to make this judgment call of what constitutes negative content. And as a result you see quite some variation in terms of which sites are blocked and which are not blocked. And this is why it's important even in places where you may say you know that censorship is happening, yes you may know that censorship is happening but you don't necessarily know what the censorship of your neighbor looks like even if they live in the same country. Another very interesting report that I think we could spend an entire one hour talk just discussing about this particular country was published in September 2017 and it was on the country of Iran. And there we found an astounding number of websites to be blocked. I think in total we counted 1,020 URLs blocked ranging from all sorts of various different categories. And without going too much in depth into this what I should say on Iran is that it's definitely one of the countries that has the most state of the art internet censorship apparatuses and they implement the most advanced forms of internet censorship. And two interesting aspects that are unique to this country I think are one that the censorship that they implement is done in a let's say non-deterministic fashion. What this means is that you will visit a website and it's not that every connection to that website will be terminated immediately or every time you visit it. They will just selectively drop some packets every certain number of requests leading the user to think in some cases that the website is not in fact blocked but is just unreliable or not functioning properly. The other aspect that we found that was quite interesting is given the fact that we analyzed the corpus of data dating back to 2014, we were also able to track the evolution of censorship over the course of these years. And one of the things that we found was that certain websites such as Instagram that previously they were blocking only specific posts, so specific images or specific pages on the site that were deemed sort of unacceptable or against the morale of the majority. But as Instagram rolled out HTTPS on all of their websites, the Iranian government was put in a situation where they either blocked all the website or they could not block the specific pages. And so I mean I guess this is more something to reflect upon of how sometimes something that you take as just a positive thing and it clearly is a positive thing to have websites implement HTTPS and have secure connections to them. But it does also lead sometimes to some effects that are not expected maybe. And this is an example of it. So now Instagram is entirely blocked in certain networks in Iran. Another case that probably you have heard about happened in Spain just a couple of months ago and it was around the Catalonia independence referendum. And there was some movement to do a referendum to claim independence from Spain for the Spanish region Catalonia. And as a result of this, the Spanish government decided that one way to sort of create an obstacle for the success of this movement was that of implementing censorship and blocking websites. I think the reason why this is something important to talk about here in Europe is that we should not forget that these things also happen in our home and that oftentimes when we talk or think about internet censorship we think as it's something that happens only over there. But in fact many countries in Europe implement censorship for gambling or other reasons. And it's kind of a slippery slope because once you have the infrastructure in place that can be used to implement internet censorship, those in power can possibly abuse it to do things that are clearly not acceptable. And so during the Catalonia referendum, 25 websites were blocked. The .cat registry was raided by the police and they forced them to take down or block a series of .cat domains that were hosting just information about the Catalonia independence referendum. So I hope this has given you an idea of what are the sort of things that can come out of running a software like Uniprobe and collecting this sort of evidence. So now what I would like to talk to you a bit about is the sort of ecosystem of Unisoftr. So the various software components that make up Uniprobe and all of the things that you see here are open source. You can like, you know, get clone them and hack on them to your pleasing. And the first thing that I'll tell you a bit about is the sort of engine that Uniprobe, the client that we use to run the network experiments, is based on and it's called Measurement Kit. And this is a C++ library that we use to implement all of our network measurement experiments. And it's currently supports Android, iOS as our two primary targets, but it also works on any Unix system. And we are in the process of ensuring that it works properly on Windows as well. We have JNI and Node.js bindings, so you can include this library inside of a project of yours where you need to do some particular experiments that involve measuring networks. As I said, the tool that we use to collect network measurement data is Uniprobe. We have an Android and iOS app that you can download on your phone. It's also available on Android. We have Mac OS and Linux as well in the form of a Web UI. But we are also working on a more fully featured, proper desktop client that will support also Windows as well as Mac and Linux. And the sort of tests that we do, as I guess you somewhat understood from the case studies, are mainly around these sorts of categories. So we have tests that examine web censorship, so whether or not a particular website is blocked or is not blocked. We check to see if instant messaging apps like Facebook Messenger, Telegram, WhatsApp, et cetera, work or do not work. We have a series of tests for censorship circumvention tools like Tor, VPN, lanterns, iPhone, et cetera. Then we have some tests that sort of are more trying to understand certain characteristics about your network. So whether or not a middle box, so something like a transparent HDB proxy or something that is intercepting your connections and forwarding them for you is present on your network. And more recently we have started to have tests for examining speed and performance. This is how our mobile app looks like. You can run it on your phone and check to see what is blocked or not blocked on your network. And by doing so, you are contributing to our global understanding of internet censorship around the world. This is our desktop client that, as I was saying, is a web user interface. And this is another interesting thing that we have recently launched, which is sort of a way for you to also engage other people to monitor or stand up against internet censorship for the things that you care about. So with Uniran, what you can do is you can add a list of URLs that you want to monitor. So for example, let's say an animal rights activist group in Germany wants to monitor some German websites. They will create an Uniran link. And when people click on this, if they have the app installed, the app will open and will offer to test the sites that they want to test. So we obviously also accept contributions for or suggestions on what we should be testing. But you can also do this on your own with Uniran. And as I was saying, all the data is available as open data. So you can access the UNI API, visit the UNI Explorer and use the data in your own studies to uncover evidence of internet censorship. So this is UNI Explorer. This is a global data map and it shows all the measurements that we have. You can browse through it to understand what is blocked and what measurements and what data is available in the various countries around the world. As I said, we also have an API that's more oriented towards, let's say developers or data scientists. So you can query this API and obtain the raw data. And all the research reports that we write are actually all based on the same data. So you can write your own research report on a country that you care about just by accessing our data, which is all as open as it can be. So get out there and use UNI data to understand better what is happening in the countries that you care about. And to close, what I would like to leave you with is sort of another story, I guess, about a trip that some of the UNI team did to Cuba. And I have a video here that I'll use to support the narration of the story in a way. And yeah, so this was in spring 2017. And let's see a second if we can get this to start from the beginning. Yeah, so without further ado, I will tell you the story of our trip to Cuba. And the data that you see here and the facts here are obviously during the trip. Some things have changed since we went there. But you can find more information inside of our research report. Many of us take the Internet for granted. Many of the things here at CCC would not be possible without it. But imagine living in a world where access to the Internet is not ubiquitous. That's the type of world most Cubans live in. This is a story of our spring 2017 trip to Cuba exploring its Internet landscape. Simone, Maria, Leonid, Joe, and all of the other Unitarians that are here, this would not be possible without them. So this is also for all of them. Most Cubans cannot access the Internet from the comfort of their homes. Rather they need to visit designated public spaces that offer Internet access. Welcome to Parknet. This park, along with many others across the country, serves as a public Wi-Fi hotspot. Such hotspots were set up two years ago in 2015 when the Cuban government provided public access to the Internet for the first time. Most hotspots are located in parks, while others are located in other public spaces like post offices and airports. In a way, Cubans don't access the Internet. They visit it. And so it becomes a sort of collective experience. Especially at night, when the heat is more tolerable, dozens of Cubans gather in these hotspots to use the Internet and to chat online with their friends and family. To start browsing the net, you need to have an account issued by Atexa. Cuba is only telecommunications company. Atexa sells these cards at $1.5 for one hour of Internet access. You can also purchase these accounts from resellers who lurk around in Parknets, but they will mark up the price to even $3 or $4 for only one hour of Internet access. We are curious to explore whether and to what extent Internet censorship was implemented in Cuba. And so we run our network measurement software called Uniprobe, as well as a variety of other network tests. Based on our measurements, we found 41 websites to be blocked. More might be blocked as we limited our testing to a list of sites created by country experts. Most of the sites that we found to be blocked include news outlets and blogs, as well as pro-democracy and human rights sites. Many of them seem to have one main thing in common. Directly or indirectly, they express criticism towards the Cuban government. Other circumvention tools like the Toran anonymity network were found to be accessible across Cuba. Our software not only allowed us to confirm whether sites were blocked, but it also allowed us to collect network measurement data that clearly shows how censorship is implemented. Atexa appears to be blocking sites through the use of technologies that resets connections or serves block pages instead of the intended webpage. In many countries around the world, when a website is blocked, you know it, because it often looks something like this or this. Block pages usually include some information notifying you that the site is blocked and sometimes even include a reference to the legal justification, which means that it may not always be clear that the site is intentionally being blocked. Users might assume that it's a connectivity issue rather than government commission censorship. But Atexa isn't the only ISP in the world to serve blank block pages. You see this in many countries as well. Interestingly enough, we only found the HTTP version of the site to be blocked. And the HTTPS is unblocked, meaning that people could circumvent it using HTTPS. However, many sites that we found to be blocked do not support HTTPS. Chinese vendor Huawei appears to be supporting Internet infrastructure and you don't even need to run any tests to find this out. The server header of the block pages, for example, points to Huawei equipment. Even Atexa's captive portal itself seems to be written by Chinese developers since they left over some comments in the source code of the captive portal page. Overall, Internet censorship in Cuba does not appear to be particularly sophisticated. Given the high cost of the Internet, rendering it inaccessible to most Cubans, perhaps the government doesn't even need to invest in sophisticated Internet censorship. But Cuba's Internet landscape is changing, and so might Internet censorship. That's why we think it's important to monitor networks continuously. Users in Cuba and around the world can run Uni Probe to shed light on information controls. Thank you. I think I finished right on the dot, so I suspect that... So, everybody in the audience who now got curious about Uni, maybe you want to consider if you want to contribute by running an Uni probe in your own country or when you're on holidays in some interesting places. Just think about whether you would like to contribute and help Uni out. And we have an assembly in the CCL at the second floor, the Universe. If you have more questions about Uni, you should come and join us upstairs and we will be happy to answer any questions you may have. Exactly. Perfect. Check them out in their Uni assembly. Please give another big round of applause to Arturo. Thank you very much.