Connecting To The Outside World

One of the most interesting and hardest-to-describe things about local networks, as opposed to big-scale national ones, is exactly how they connect to the rest of the world. More than any other part of this project, the answer to these kinds of questions is “it depends!” In this post, I’m going to do my best to provide the long-form answer to this question and explain how things interconnect – in general, as well as our exact case in Bokondini.

Internet Access

The Internet was designed from the ground-up to be a decentralized “network of networks,” and this design makes it incredibly easy to add a network to the Internet: you pay an ISP for a connection, hook up a router to it, and you’re done! Though the scale’s a bit different, this process is pretty much identical, whether you’re hooking up a small home network or building a telecom company. This isn’t an exaggeration at all: small-scale regional ISPs literally just buy a super-fast Internet connection from a bigger ISP, the same way you do at home, and then divide and resell it to their local customer base.

In our case, the mountains around Bokondini are too rugged for traditional wired connections, so we’re using satellite Internet. Though it’ll cost us 300 USD/month for a 1Mbps connection (no joke!!!) and be pretty slow, it’s literally the only option we have.

Calls and Texts

So if hooking up to the Internet’s that easy, hooking up to the phone network’s probably similar, right? Wrong! It’s actually quite a process to register yourself as a phone network, and get assigned a block of phone numbers. This is the case for several overly-technical reasons, but the main one is that the Internet was designed to easily allow smaller networks to hook-in, whereas the telephone network (known as the PLMN, for Public-Land-Mobile-Network) grew up served by a much, much smaller number of operators. As a natural result, there’s no simple or easy system for a small network like ours to request a small set of numbers, mainly because that never happens. I’d love to learn more about this system and go through the process down the road, but for now, we currently do not have any plans to register our network in the PLMN.

Note that not having a PLMN hookup doesn’t actually affect our ability to support calling and texting, at least on a local scale. The nice thing about not being connected to the PLMN is that we can still assign our users whatever made up phone numbers they want, and we can still use these numbers for routing calls and texts within the network – we just can’t connect to the rest of the world.

Over-The-Top (OTT) Services

If hooking up to the Internet is easy, but hooking up to the phone network is hard… some of you might be thinking “why not just use the Internet for voice and text?” Great idea! In the cellular world, running voice and text over the data connection is known as “over the top” service provision, and it dramatically simplifies everything, for a number of reasons. Popular examples of over-the-top services include Skype, WhatsApp, Facebook Messenger, Google Voice, and many, many more.

There are tons of really good reasons to use OTT services: first off, they’re typically much cheaper than buying a phone plan. Second, all they need is Internet, so they work just fine over WiFi. Third, since the Internet was designed to be more decentralized than the PLMN, they don’t deal with issues like roaming.

Unfortunately, OTT services also have some drawbacks. First, whereas the telephone network is understood and supported worldwide, OTT services typically don’t interconnect with each other, which leads to complicated issues of fragmentation (i.e. you talk to John on WhatsApp, Matt on Skype, and Fred on Facebook) and a negative user experience. Second, in almost all OTT services, communication must be routed through the provider’s servers. This typically isn’t a huge problem in well-connected areas, but in our case, it means that a call from one user to another would go over that 1 Mbps satellite link twice for no good reason. Finally, and most importantly, since OTT services just do their own thing and don’t interconnect with the PLMN, they can’t support a lot of telephone-specific features such as emergency calls. This isn’t just unsafe, but actually has legal implications, and is why Skype has that big disclaimer saying “Skype is not a replacement for your phone and can’t be used for emergency calling.”

Bringing It All Together

Okay, so let’s recap: our network in Bokondini’s going to have a relatively slow and expensive satellite Internet connection. With this connection, our users can do anything they want: email, web surfing, etc. Additionally, our users can make phone calls and send texts to other users within the network, but can’t call anyone outside the network. Finally, our users can also use any Internet-based OTT services (WhatsApp, Skype, etc.) to call people outside the network.

Ordering SIM Cards

One of the coolest and most interesting parts of building CoLTE was the process of designing and purchasing our own SIM cards (1,000 of them, to be precise). Even though the SIM cards are a relatively small part of the whole technical system, they make a great visual, especially with our own design on them. Most importantly, as physical, tangible objects, they really drive home and solidify the idea of “DIY Telecom” like nothing else can.

Where The Heck Do You Get SIM Cards?

On, apparently! For those of you not in the know, Alibaba is a huge e-commerce site, based mainly in China and geared towards product manufacturers and other parts of the supply chain. It’s where you’d contract with a Chinese producer to buy (normally large) quantities of anything from custom-logo keychains to specialized electronics and machined parts. A quick search for “LTE SIMs” yielded dozens of results, which I eventually pared down to a supplier called GreenCard that appeared to be a small company based out of Shenzhen. GreenCard offered us everything we needed, all for 65 cents a SIM, minimum order size of 1,000 units.

I had never used Alibaba before, and was a little bit nervous about placing such a large order over such a far distance, getting something slightly wrong, and blowing a thousand bucks, but the process was remarkably straightforward and reassuring. The Alibaba website gives you plenty of support, and clearly spells out how to specify things correctly in your order to ensure order satisfaction and CYA in case of a dispute. The vendor I worked with was quite professional, helped me when needed, and delivered a bang-up job all the way through. All in all, the process went incredibly smoothly – the only hitch was communication took longer than I would have liked, but that’s just a time zone issue.

The Visual Design

Given that SIM cards are so tangible and user-facing (they’re literally the first and only object a user gets from their telco), we knew it was worth it to put a lot of thought and effort into the physical design, despite the risk of bikeshedding it. After a couple of brainstorming sessions oriented around themes like “empowerment,” “agency,” “local,” and more, I mocked up about ten different designs on Illustrator and passed them around. We ranked favorites, brainstorrmed some more, fine-tuned the designs, and finally settled on the final design below.

sim card 确认稿 20180228.jpg

The front side is a picture that Kurtis took in Bokondini several years ago, when they set up the initial Village BaseStation network. We wanted to keep the back simple, with the W centered so that even when punched out, our cards would be instantly recognized as the “Purple SIM” or the “W SIM”. Personally, I’m really happy with how these turned out, and am already thinking about a similar design when we expand into future communities: finding a relatively local picture that highlights and emphasizes the community itself, and using that as the main front design, because when it comes down to it, that’s really what this whole project’s about.

Technical Bits

Once we settled on a design, I thought I was almost done… right? Wrong! Turns out there were a ton of other details to take care of before we could place an order: everything from what our Public Land-Mobile Network (PLMN) ID should be to secret keys for security, PINs to lock and unlock the SIMs, even what phone numbers the SIMs should have. We were also able to set the “Network Name” that users see when they connect to our network: though it was tempting to take a page out of WiFi and come up with our best cellular-related puns, we eventually settled on a simple “Bokondini”.


Introducing CoLTE!

Hi everyone! My name’s Spencer, I’m a postdoc at the University of Washington, and I’m super excited to introduce the Community LTE Project (CoLTE, pronounced colt-ee). I’ll be writing about CoLTE a lot over the coming years, with two main goals. First, we want to keep everyone informed of the project development, major milestones, and community engagement efforts. Second, we’re going to be sharing the story of building and deploying a rural community-run LTE network in Bokondini, Indonesia.

The writing here should be pretty casual – our target audience is everyone, regardless of technical savvyness. We want everyone who’s interested in this project to understand and follow what’s going on: whether you’re a friend, web developer, network technician, VC, or community member, if you’re interested in CoLTE, we want you to know about it! In this first post, I’m going to tell you all about this project, what led to it, where it’s at now, and where we’re headed.

What is CoLTE?

The Community LTE Project is a new approach to high-speed, LTE cellular networking that is decentralized, democratic, and community-based. The goal of CoLTE is to empower anyone (and we mean anyone!) to start their own small-scale cellular network, for fun, profit, or both. Install CoLTE, hook it up to a radio, distribute SIM cards, turn the network on, and watch phones connect. Our goal is for the whole process to be as straightforward as hooking up a WiFi router: a little bit technical, but totally doable for a regular person, and hopefully easier than setting the clock on your VCR.

From a more-technical perspective, CoLTE is a set of many different software packages (all open source!) that we’ve integrated into a single platform that works nicely together. Everything from the “guts” of the network (LTE attach, security, mobility, etc.) to network management and billing are all taken care of, and each component can be turned on, turned off, or replaced without disrupting the other parts. We’ll have a more detailed technical breakdown of this very soon, so stay tuned.

What’s the point?

The core goal of the CoLTE project is to empower individuals and communities to connect themselves to each other and the global Internet. Right now, approximately four billion people (slightly over half the world’s population!) don’t have any Internet access at all. This population is primarily located in rural and remote areas, where infrastructure is more difficult and expensive to install and maintain. Combine these increased costs with decreased revenue due to lower population density, and it’s no surprise that large telecom companies and ISPs don’t want to invest.

To make headway on this problem, we have to ensure that these incentives are taken into account. Our stance is that if a community wants Internet access, the best way for them to get it is for them to literally form a local telecom company and connect themselves. This ensures that incentives are aligned, keeps value within the local community, and empowers local actors to develop a relationship with their network infrastructure. To this effect, the best thing that we can do to support this effort is to build tools that help, enable, and empower these communities as they connect themselves.

How are things going right now?

Right now, we’re hard at work on three main thrusts: the network core, the peripheral web services, and the production environment.We’ve still got plenty more work to do, but the main parts are all operational, and we finally got everything working last week. Additionally, we’ve been coordinating with some folks on the ground in Bokondini, and are making plans to stand up our network there in Summer 2018. Over the next couple of weeks, I’ll be writing in greater detail about each of these efforts, and chronicling our day-by-day process as we start bringing extra services online and getting everything ready for production.

How can I get involved?

Rural Communities: We’re currently looking to partner with rural communities that are underserved by traditional Internet providers. Does your community not have Internet access? Does it want Internet access? We want to hear from you! Use our contact page to get in touch.

Network Operators and Developers: Want to play around with our stack? Curious about our network architecture, or think it might be a good fit for you? Want to contribute to the project or get involved? Contact me, explore our site to learn more, and check out our Github here.

Friends: Drop us a line, spread the word, and stay tuned for more adventures. Thanks for reading, and until next time, here’s a couple of pictures of our day-to-day work.

Our basic testbed on my desk, much cleaner than usual
One of the commercial-grade base stations we’re taking to Bokondini
Our custom-made SIM cards!
IMG_20180523_142511916 (1).jpg
The mega creepy soundproof and RF-shielded chamber I run radio tests in


Spectrum Usage “Under The Radar”

This post is also published on the UW ICTD Blog here.

When it comes to frequency regulation, the FCC has to be strict — but its coarse-grained approach to licensing harms the small-scale hobbyist community. It could stand to learn a thing or two from the FAA on this issue.

Lately, I’ve been building a small-scale LTE network in our lab for testing. This network doesn’t use very much power, has a range smaller than my office, and is definitely stopped by any walls. This network is far too small, and too quiet, to bother any telecom operator — and yet, because LTE runs on a regulated frequency, it’s illegal for me to do this without an FCC license!

As we went through the licensing process, I started to notice a lot of interesting parallels between how the FCC regulates frequency and how the FAA regulates airspace: they’re both invisible (meaning you can commit a violation without realizing it), strictly regulated, and come with stiff penalties. Where they diverge, however, is the relationship with their DIY communities. The FAA heavily supports the general aviation community, and carves out airspace exemptions where possible to ensure that smaller airports and aircraft can continue to operate in an increasingly airline-dominated sky. Meanwhile, the FCC seems to have no such systems in place: frequencies are either completely licensed or unlicensed in an area, with no middle ground in between. Add this to expensive and complicated application forms, and it’s a perfect recipe for discouraging hobbyists and DIY experimentation.

How The FAA Handles Things:

A lot of people are familiar with the phrase “flying under the radar:” in common use, it means to do something discreetly, so as not to attract unwanted attention. What most people don’t know, however, is that the phrase also refers to a perfectly legal process designed for small aircraft and encouraged by the FAA itself.

When large airliners approach a major airport (like Sea-Tac, for example), the airspace they’re flying through is heavily controlled and monitored by Air Traffic Control (ATC). Starting about thirty miles away, the airliner must make radio contact and request permission to enter the airspace, at which point its movements are strictly dictated over the radio by Approach Control, who’re busy making sure that everything runs smoothly. While this high level of control is great for safety and helps keep planes on time, it also poses a serious challenge for smaller airports or airstrips that might be within that thirty mile radius. These small airstrips, which can be untowered or unpaved, are home to a wide range of small, local aircraft that have no interest in going anywhere near Sea-Tac. Meanwhile, the controllers at Sea-Tac already have their hands full with the airliners, and don’t want to talk to any airplanes they don’t have to.

The FAA’s solution to this problem involves a series of shelves. As airspace stretches farther away from the airport, the floor of the controlled airspace rises, leaving uncontrolled airspace below — literally “under the radar.” For an example, consider the University of Washington, roughly ten miles North of Sea-Tac. Over UW, Seattle Approach controls the airspace from ten thousand feet down to three thousand feet, but ignores anything below. This leaves plenty of room for small Cessnas to tour the city, fly from one airport to another, or just go out for flight lessons, without ever having to talk to Seattle Approach — and they’re perfectly legal, just as long as they stay under three thousand feet.

How The FCC Could Support Hobbyists:

The FCC application process isn’t the worst thing in the world, but it’s neither straightforward nor particularly easy to fill out. To be sure, they could probably make the process better, but there’s still a strong need for exemptions that would allow hobbyists, kids, and radio enthusiasts to play around without application forms or fear of breaking the law. Here’s a list of suggestions or ways that the FCC could better support the DIY community without disrupting bigger-scale ventures.

Idea #1: Waive regulation below certain signal strengths. Whether it’s AM/FM or cellular, a lot of hobbyists just want to make something work across the room, and then say “hey, we did it!” This kind of usage doesn’t interfere with anyone else, specifically because the signal’s at too low of a power to get very far. Granted, when transmitting something at such a low power, and over a relatively short distance, you’re not very likely to get caught — but it’s still explicitly illegal, and that’s a bit besides the point here.

Idea #2: Relax regulation on private property. The sum total signal strength of a radio signal degrades relatively quickly, at a rate of 1/r², with r being the distance to the antenna. If we also take walls and obstacles into account, this means that someone could broadcast a moderately strong radio signal inside their house without it bothering anyone outside. It’s easy to imagine an approach to frequency regulation that takes this into account, saying something like “you need to have a license to broadcast in public space, but anything you do in your own house is fine, just as long as you keep it below a certain signal-strength outside.”

Idea #3: Use a permissive, rather than restrictive, approach to licensing. Right now, we have an FCC experimental use application pending approval. The application is incredibly humble: we want to setup an LTE network inside a company’s headquarters, for a single day, on a single frequency band, to connect to a single cell phone, with a maximum radius of 15 feet.

Even though I know that this network isn’t going to bother anyone, we still had to apply, and are currently awaiting FCC approval. This process could be better served by a more permissive approach, wherein we could get immediate approval for our application, or a default answer of “yes unless we tell you no” as long as our application meets some certain criteria (e.g. kept indoors, or below a certain radius or signal strength). This way, the FCC is aware of the network, and still retains the option to say no, but operators can still move on with their lives in the meantime.

Idea #4: Create a basic, all-purpose “hobbyist” license. The suggestions above all stem from a core idea that if you’re careful, it’s not hard to stay safe and avoid disrupting other users. However, if the operator doesn’t know what they’re doing, it’s easy to screw up and broadcast on the wrong frequency, or at a much higher signal strength than desired. Even worse, since radio waves are invisible, there’s no clear or obvious indication that this is happening!

The FCC could reduce these risks by providing a basic operator or hobbyist licensing program. Rather than focusing on a specific type of radio protocol (e.g. AM/FM or LTE), the program could simply ensure that the operator understands basic radio principles, best operating practices, and how to stay in compliance. This idea could be paired up with the previous three, granting license-free permissions only to people who have demonstrated an understanding of and ability to comply with the restrictions placed on operation.

This system, if taken seriously and built out properly, could end up looking a lot like the FAA’s: pilots still need licenses to operate aircraft, but are allowed to go wherever they want as long as they know how to stay away from restricted areas. This approach to regulation has been vital to the general aviation community, which includes amateur pilots, flight instructors, students, and charter pilots. Similarly, I have no doubt that relaxing frequency regulations within a clear set of boundaries would be equally beneficial to the DIY radio community.