« More interviews

#4: Paul Gardner-Stephen - Serval Project

Paul Gardner-Stephen talks about the Serval Project, which lets mobile phones make calls without a cell tower. He gives real examples of it being used in disasters today. 14 Aug 2013



Kindly transcribed by: David Hansen

Francis Irving: Hello! Welcome to the redecentralization interviews, and today I’ve got Paul Gardner-Stephen from the Serval Project with me, which is about mesh networking of phone calls. Hello Paul!

Paul Gardner-Stephen: Good morning!

Francis: So, could you tell us a little bit about the Serval Project, where you got the idea from and how it all got going?

Paul: Sure. So really it was after the Haiti earthquake that triggered me to do something. And I still remember sitting in the car, driving to work and hearing about the earthquake. And human coping factors being what they are, I knew that the loss of communications — or even just the impairment of communications — in Haiti would be a real problem for that already strained community there. And so in my mind I was thinking like, oh, they’ll be able to bring in by air and get communications gear in, and get everything back up and running and avoid the descent into lawlessness.

And while I’m thinking this happily to myself, the person on the radio says, ‘Oh, and the airport has been almost destroyed. There’s one runway which is still actually open. One plane in or out every half an hour.’ And so I’m thinking like, ok, so that coping mechanism is not going to work. So then I thought, well, ok, there’s a highway that goes from the Dominican Republic into Haiti. They’ll be able to truck gear in. It’ll be slower, but they’ll be able to still get it there in that critical one- to three-day timeframe to maintain law and order and really help people. So while I’m thinking this to myself, the person on the radio says, ‘. . . and practically every road in Haiti has been destroyed by the earthquake.’

Francis: Wow.

Paul: And so then I’m kind of thinking, ok, so there’s a harbor in the Dominican Republic. They can load up a container ship. A couple of days cruising around into Port-au-Prince harbor, and it’ll be stretching the timeline a little bit, but they’ll be able to get piles of gear in and really make a big impact to start getting the place back together. And while I was thinking that, the person on the radio said, ‘. . . and the harbor in Port-au-Prince has collapsed!’

And I just remember the real emotional rawness of that, and realizing that things were going to go very badly for the Haitian people. And unfortunately as history shows, that was indeed the case. I remember hearing from doctors working in the area about militia roadblocks and rape gangs, and just all sorts of really nasty stuff that was going on. So I came to this position that this should never happen again, that whenever a disaster happens people shouldn’t be deprived of communications in a way that lets things descend in this kind of way. And it really led me on a journey over the next couple of months, until I realized that the mobile phones that people carry actually were the solution, that they already had them, that they were already in the disaster zone. They had battery backup. They just needed to be able to be programmed to talk to one another. And really by historical accident that hadn’t happened. And so that’s what we set about doing.

Francis: So you kind of realized the resilience of our world, of our civilization, just isn’t that great.

Paul: Yeah. And I guess really there hasn’t been the commercial imperative to do it, because we’ve been able to make such fantastic, centralized, large enterprise systems. And really I mean, when you look at history, particularly the second half of the 20th century I think will stand out as an aberration where centralized was by far the most efficient way to do things. I mean you look at 3D printing now, and suddenly manufacturing is starting to move away from ‘big is best’. And it will be interesting in 50 years’ time to see that shift.

Francis: So what does it do? How does the Serval Project work?

Paul: So basically you load an app, at the moment onto an Android phone, and it allows the phones to talk directly to one another. And we put a lot of thought into making it being really easy to use, so that all you need to tell our software is your phone number, and your name if you want people to be able to see your name on the network, and then use our app. And you can dial, you can send text messages, you can share files — you can really do all those kinds of things that you expect a mobile phone to do, but in the kind of places where normally only a CB Radio would work. So really it’s a fusion of smartphone and CB radio for the 21st century.

Francis: So there’s no base station or anything like that? It just works directly between the phones?

Paul: That’s right. So in its simplest and most ideal arrangement, that’s how it works. Unfortunately, to get the full function out of the phones, to talk directly to one another, you need to root the Android phones. And of course we know that that’s actually not that viable an option for most people, and so we’ve been working on what we call a ‘Mesh Extender’, which is basically a little, battery-powered, pocket wireless router with an extra-long-distance radio in it so that instead of just Wi-Fi range, we can get, in an urban area you might get a quarter of a mile through buildings; in open country you might get potentially a few miles between units. And the idea is that people will get these things, hopefully before a disaster strikes, but even if they can’t, they’re small and cheap enough that you could overfly an area and drop these in to people.

Francis: So without that extender, in what kind of range can the phones communicate to each other? This is with Wi-Fi, is it?

Paul: That’s right, yeah. So the usual Wi-Fi kind of range. So indoors is going to be 10-15 meters, outdoors is going to be 30-150 meters is what we find with mobile phones typically.

Francis: 150 meters?

Paul: Yeah, that’s the upper limit that we see with mobile phone Wi-Fi.

Francis: And it uses this thing called ‘mesh networking’ as well, does it? — or not? — to route things via other people’s phones?

Paul: Exactly. That is precisely what it does. So if I wanted to ring you and there were a couple of phones in between, the call would potentially be routing through those phones. And the people with those phones in between don’t have to think about that, they don’t have to do anything special. If the mesh software is running on their phone so that they can receive a phone call, then it will actually relay calls for other people as well.

Francis: Ok. So just to go back to the Haiti situation — how would the extra communication, if that kind of system had been available to people, how would that have helped them? How does it. . .

Paul: I think a really important thing is that it enables local communities to stay in contact and ultimately defend themselves. I mean if you think about the terrible example of the rape gangs that were going around, if local communities could alert one another and say, ‘Hey, there are these guys going around doing this,’ well, for a start you can bet that every wife, daughter, and granddaughter would be safely inside when they came by. But you can also bet then that you would have — the local community would be organizing to defend themselves against these groups and say, ‘Hey, get out of here. You’re not doing this!’ Or citizen arrest. Or there’s a variety of options.

In fact, I would even argue that a lot of these sorts of things probably wouldn’t start to begin with, because people would know that the local community could organize to defend themselves. And the same with the militia roadblocks — if everyone can tell everyone else where the roadblock is, you can avoid it. And so there are those sort of civil defense things. And then beyond that you have just the ability for people to say, ‘Hey, I’ve got water, I’ve got food,’ or ‘I need food and water.’ Or immediately after an earthquake you kind of go like, ok, I’ll get in contact with my family and friends who live nearby. Ok, they’re safe. Cool. I now know that I don’t have to go and find them. I can go and help other people to get rescued. And so really it’s a massive enabler for recovery.

Francis: So it’s a few years old now, the project, isn’t it?

Paul: Yeah.

Francis: Has it been used in any real situations, or are there any interesting stories about people making use of it?

Paul: So the one real deployment that we had was actually in Nigeria. There are some communities living in waterfront places along there that the government wanted to basically bulldoze and redevelop, despite the fact that those people had lived there all their life, for multiple generations. And so they trialled a very early version of our software. And it was really interesting. Despite the fact that we didn’t have mesh extenders at that time, so it really was limited to the Wi-Fi range, it turned out that the people there were able to communicate more often and spend less on their communications than they were before they had the mesh technology. So that was really tremendous. And there’s a report from that available on the Internet. And there are actually a couple of videos that they made of their protest against the proposed forced eviction, which was really pleasing to see — that our software had been used to help people to defend themselves, and to be safe, and to be able to keep living where they’d lived for generations.

Francis: Yeah! So that was literally because they didn’t have the money to afford full-on phone calls, or they didn’t have the connectivity, or?

Paul: So, yeah. So I think cost was a limitation because these are fairly poor communities, which of course is part of their vulnerability. And so I believe there was mobile phone coverage in a lot of the places, but their ability to afford to use it, and particularly things — to make a video of an interview with a resident and get that to Amnesty International whom they were working with, was a fairly expensive proposition. So suddenly they were able to film and have it go onto the mesh and get replicated and find its way back to Amnesty and other partners that way.

Francis: Yeah, this is quite exciting! It’s quite interesting, because some of the other people we’ve interviewed so far, the subject of resilience hasn’t really come up, but I think the reasons people seem to want to decentralize things seem to be either to do with privacy, or to do with resilience, or it has to do with fun; just like bringing back interesting things to the world. So how do you think, as we get better again at decentralizing things, and as maybe we create new ways of doing things — based on this, can you describe the world and how it might look in the future, what you think would be good?

Paul: So I think fundamentally it will be a more just world. So I mean, you look at the digital divide, and the tremendously effective big-infrastructure approach to mobile communications and Internet, for example, means that — you know, by global standards we’re both rich white men, and we are extremely privileged, and here we are talking between continents without even — we don’t have to think about whether we can afford to do that, or whether we can do it. And to see at least some of that capability start to bleed through into the people actually who need it most, for whom it can, you know, enable subsistence farmers to contact local markets and get better pricing for their food, and to not have to walk to two or three markets to get fair pricing. And you get into that whole interesting calorie budget problem for these guys. That simply means that they have more food on their table and more that they can sell, and so really from the ground up we see it helping communities leverage. And actually one of the interesting ironies is that it will help them get to the point where cellular service becomes affordable for them and feasible to provide.

So we really see it as a great complementary bootstrap for social and economic development, as well as a fantastic and affordable insurance policy for when things go wrong, as we’ve seen in Haiti, and even more recently in Japan with the earthquake there, and even in New Zealand with our partners New Zealand Red Cross, following Christchurch and even the scare they had in Wellington a week or two back. Actually they had two — in a fortnight they had a 5-point-something earthquake that was very shallow and nearby, and they also had what they call an Antarctic weather bomb, which is basically a, they call it an extratropical cyclone. So basically it was a really intense low-pressure system that moved up from Antarctica, and they had 200-kilometer-an-hour winds there without actually being in the tropical cyclone or hurricane belt.

So there are a whole variety of these things where I think it will help. And nomadic communities where it’s just never likely to be feasible to provide them with cellular coverage. So you know, Aboriginal people here in Australia; there are still nomadic tribes in the Arabian Peninsula and throughout Africa and areas of Asia as well, where suddenly these people can use mobile phones like CB radios but have the privacy, have the rich functionality that we’ve come to expect out of modern technology.

Francis: Yeah, it’s quite interesting really. So I’m quite interested in some of the other similar projects, like there’s a project called the Global Village Construction Set, where they’re making a whole set of open hardware things.

Paul: And Marcin [Jakubowski] there is fantastic with his vision and his persistence and energy in doing that. We certainly see that what we’re creating, I think in the longer term will be one of the pieces of that construction kit, or at least one of the options for communications in that. Really, I mean, what he’s designing is something that, if we were to colonize another planet, you could actually set up a civil society without having to ship everything from tractors to orange juice into the place.

Francis: It’s fascinating how that’s both useful now in countries which are already not in a stable state, and it’s also potentially useful if there is a disease, or a nuclear war, or any of the kinds of disasters that can affect all of us.

Paul: That’s right.

Francis: Wow. Ok, I’m going to step back out slightly, to be more techy for a minute. The mesh networking part — so there’s this part that via Wi-Fi, that at most goes maybe 100 meters, it can route your phone call via other people’s phones. So if lots of people in one area were running the software on a rooted Android phone or you had enough range extenders, it would somehow route the calls, like peer-to-peer, and chain them? So how does that work exactly, that algorithm? And how far can it go, like how flexible and powerful is it?

Paul: Sure. So we have two ways of doing it. One is for real-time communication, so specifically for phone calls, in fact. And at the moment it’s not used for a great deal else. And that can go probably five or ten hops.

Francis: Oh, wow!

Paul: So then you start saying, potentially it might be a few hundreds of meters, upwards to approaching a kilometer or a mile or so. We have the other one which we call Rhizome, which is designed for when the network can’t actually make a real-time link from end to end. And we really love the simplicity of it. Basically a phone says to its neighbors, ‘Hey, I’ve got these things,’ — which ultimately are files with, if you like, a cryptographic envelope around them — and so they compare their lists a bit like trading cards, and they go, ‘Ok, you’ve got that, I’ve got this. Let’s swap a copy of each so we now have both.’ And then they start doing that to their neighbors. And so a file can find its way across, effectively an unbounded distance, and an unbounded number of hops.

Francis: Does it copy itself to every hop, or is there some way it knows like, which direction to go in?

Paul: Yeah. So at the moment it copies itself to every hop, which of course has some scaling problems, but it’s amazing for a smaller community, the effectiveness. And even if you want to get information out to everyone in the community, then suddenly it actually is tremendously useful. We also have a mechanism where, once it’s been delivered, then you actually propagate out a deletion message across the network, and so it deletes itself. And we are looking at having, if you like, directed propagation and other things to improve the efficiency of it, but it turns out to be amazing in the way it works because you have infinite retry on delivery as well. So if the network is chopped up into little pieces because everyone’s too far away, or there’s too much interference, or maybe you’re in an area of unrest and the local militia are trying to jam all the frequencies available, the instant that that impediment to connectivity stops, suddenly the data actually starts getting delivered again, and quite quickly.

The most fun test we did of this, soon after we made it, was that we actually sent a text message from Magaliesburg in South Africa back to the lab here in Adelaide, where the means of transport was actually by carrying the phone home on an airplane that was already carrying the message. And then when we basically walked into the lab with that phone, the phone the message was addressed to just went, ‘Zzt zzt, you have a new message!’

Francis: Oh, wow!

Paul: It’s a slightly absurd example, but when you think about that with people walking around in rural areas, or in communities that have been damaged by earthquake, just that Brownian motion of people suddenly actually makes it a really effective way to get a lot of data. And because it caches using the SD card on the phone, you know, you can potentially have gigabytes of data being cached on each device. And when we start talking about text messages they’re pretty small. We’ve also actually created the best-of-breed open-source short-message text compressor to further leverage that as well.

Francis: Wow, that sounds actually useful on the train to London as well!

Paul: Yes!

Francis: So the — what was I going to ask — oh yes, so I was going to ask about, if there has been a disaster and you haven’t had the foresight to actually set up the software in advance, is there any mechanism for getting it out to people?

Paul: Yeah, so this was one of the first things that we thought about doing. In fact, in the Serval Mesh software (that people can download and try out the experimental version off Google Play; just search for ‘Serval Mesh’), you’ll find that it actually has a ‘Share’ option, and it can share itself via Bluetooth or Wi-Fi Direct or a variety of other ways, to other phones, so that if only one phone in a disaster zone had it, basically you could actually spread it onto all of the phones. It’s sort of like a positive virus, almost. But of course, it’s polite, it asks you for permission.

And with the mesh extenders, they will also offer the app for download directly as well. It’s really interesting because this solves what is normally an intractable problem. A disaster by definition is where the infrastructure and local capacity is insufficient to meet the demand caused by the event, and suddenly we can actually deploy software in the disaster zone. And we can actually update it. One of the first things we did on the first trial with New Zealand Red Cross — we were very rapidly developing the software while we were actually out with them in New Zealand. And it was just one of these aha moments — so we’re four days into the exercise, and I’m thinking, I’ve just pushed out the ninth update to our mesh software that has been automatically downloaded and installed by all of the phones on the network, which is amazing!

[Talking in background] Just hold on a moment, my son would like some oats.

Francis: Oh, ok!

Paul: Need some more oats, do you? Ok, let me get you so more oats. [Paul leaves.]

Francis: While he’s getting some oats — I’m going to have to ask him about the shoe phone. So I’ve heard about the shoe phone, and the Awesome Foundation grant that he got that started all this. I just think that if I didn’t ask about the shoe phone then I bet someone will tell me off or something. There’ll be like, retribution from somewhere! But we’ll see. And then, so after the shoe phone I’m then going to ask about the Indiegogo that he’s got going.

And what’s particularly fun about Google Hangouts on air, which we’re using to record — this is all I can do is end the broadcast; there’s no pause button. So we’re actually going to have to wait while he sorts the oats out. But that’s good. So yeah, Paul is actually in southern Australia and I’m in Liverpool, so it’s kind of quite fun. It’s morning for him but it’s the middle of the night for me. [Paul returns.] Brilliant! So I’ve just told everyone else what I’m going to ask you next. . .

Paul: Sure!

Francis: . . . which is quite funny! I’ve got to ask you about the shoe phone, because no one would let me not ask you about the shoe phone, so tell me about that part of how it all started, and the Awesome Foundation.

Paul: Yeah, so it was quite interesting. So I was on the committee for a local church camp, and they often have a movie theme that they tie in with the talks and activities on the camp. And so this particular year they decided that they would have a Get Smart-inspired theme. And so, at this meeting I still remember they like all pointed at me and said, ‘You’re an engineer. Make us a shoe phone, a cone of silence, and a phone box that people can talk out the bottom of!’ And I’m like, ‘Yeah, you know, actually I think we can do that!’

So I just looked on eBay and I thought, you know, surely someone’s made shoe phones and is happily selling them. But it turned out that no one had actually ever made a real working wearable shoe phone. The closest I could find was a basketballer — whose name I can’t remember [Shaquille O'Neal] — in the States, with like, size 83 million shoes that were large enough to basically just embed an entire large mobile phone into. But of course, when you need shoes the size of the planet you also are quite heavy, and so I believe the shoe broke when he tried to walk on it.

And so with the help of a local cobbler we modified a pair of shoes and embedded the hardware in the bottom, and used it with the kids and young adults on the camp, which was fantastic. I’ll see if I can actually just show you here [moves camera]. So the phone booth we did using a nineteen-inch rack with the door on it, and so — oh, blast, I’ve taken the door off and put it in the shed, [audio skips] got the telephone sign on the door of my rack. And the cone of silence it turns out actually works just as well as the one in the TV show! It is completely impossible to hear what the other person is saying inside, but everyone outside can hear you very well. I suspect, once I discovered that actually, that the gags around that in the TV show were actually based on their real experience of trying to use this thing. [Sound in background] Let me just go and diffuse a minor situation between our three small people.

Francis: Sure! [Paul leaves, then returns.] Ok, I think it’s. . .

Paul: Sorry about that!

Francis: It’s time for us to do the last part, which I — so one of the things that I’ve been thinking about with these interviews and redecentralization is — how can people get involved and help out? And one way is obviously to find the code of any of these projects and fork them and send patches, which I’m sure you would love.

Paul: Absolutely.

Francis: But the other way — I want to hear about the Red Cross and what’s happening with that, and the Indiegogo project that you’ve got on at the moment.

Paul: Sure, yeah. So what we’re looking to do is to raise about $300,000 to make the [audio skips] which are currently a fairly simplistic prototype, and to get that to the point where we can actually say, ‘Here is a design that could actually be manufactured in a more serious way,’ and get these things out into people’s hands. So really finalizing what [audio skips] in those and make something really interesting. So it would be fantastic for anyone to have a look at that and, if you’d like, to contribute, but also absolutely to spread the word as far and wide as you can. You know, the more eyeballs that we can get looking at this, then the easier it is for us to hit that target and make this technology available to those who need it.

Francis: And that’s Australian dollars, is it?

Paul: That’s US dollars, which — well, it’s quite funny. Three months ago they were about the same thing, and the Australian dollar has blessedly dropped in the meantime, which means it will actually be able to do even more for the same amount, which is fantastic.

Francis: And you’re doing this project with the Australian Red Cross — it’s like a specific version of the phone?

Paul: No, so it’s New Zealand Red Cross that we’re working with, and Flinders University where I’m based, and then actually the Serval Project has a not-for-profit incorporated association. And it’s actually that association which is doing the fund raising.

Francis: So if it’s funded they’ll actually manufacture some of these phones and use them in a real situation, or?

Paul: Yeah, so the idea is that we’ll be able to make a nice design and we can get some reasonable number of units manufactured, and we’ll get a whole bunch of those to our partners in New Zealand Red Cross. And they will start looking at using those in real deployments and making that part of their training; starting to be able to make use of the technology.

Francis: Yeah, that’s fantastic! So that’s called ‘Speak Freely’ on Indiegogo. I’ll put the links on the web. Great! Is there anything else you want to mention, or?

Paul: I think we’ve covered a fair bit here, which has been fabulous. I mean, certainly people can have a look at the source code up on GitHub, and we also have a developer wiki that has a whole pile of the technical information about how it works and what some of our plans are, and how some of the security and crypto works, because I know a lot of people find that quite interesting as well. So we can give you the links to put up for those as well.

Francis: Fantastic. Thank you Paul! So Irina and I are thinking about doing an actual podcast where we have little interviews with people with more conversation, so maybe we might ask you to talk again at some point on that.

Paul: That would be my pleasure.

Francis: And we’ll try and get a good enough audience to make that worth your while.

Paul: Sure.

Francis: Fantastic. Thank you very much. It’s a fantastic project. It’s really exciting, seeing it being useful in the world now, immediately. It’s very inspiring.

Paul: Thank you. And for us it’s always fabulous when people take an interest and get excited about what we’re doing, because when you’re in the middle of something there’s that sense of like, ‘Are we really doing the right thing? Is this really helping people?’ And then you get a whole bunch of people, like there was a fantastic reddit page on one of the bits of media we had recently, where we hit the front page of reddit. And you kind of go like, ‘You know, we are doing something that people care about!’ And it is really encouraging and fantastic.

Francis: Great! Have a fantastic day in Australia, and I’ll leave you. . .

Paul: No worries! You take care as well. See ya.

Francis: Ok, take care. Bye.