Spacex & Starlink: Whether a Satellite Internet is a Good Idea

 In June 2020, I went outside in the middleof the night to spot a train of Starlink satellites in the sky. They are designed to offer Internet accessanywhere on Earth. And that made wonder: how does satellite Internetwork and why isn't it more common? 

Let's start with the basics of satellite Internet. Traditionally, your home is connected to your Internet service provider via a phone line, coaxial cable, or fiber connection. Your ISP, in turn, is then connected to the broader Internet, which is what allows you to watch this video. Now for satellite Internet, the cable at your end is replaced by a wireless connection to a satellite in orbit around Earth. But this satellite isn't connected to the Internet, so it beams the signal back down to a ground station which is. This brings us to the most obvious downside of satellite Internet: delay or latency. Most communication satellites are orbitingin geostationary orbits at an altitude of over 35,000km and that massive distance delay seach Internet request by 400 to 600ms (roundtrip). The signal needs to travel from you, all the way up to the satellite, down to the ground station, and then all the way back to you again. Now you'd think this wouldn't be much of a problem because these signals travel at the speed of light. But if you divide the distance between you and satellite by the speed of light, you get 116 milliseconds. And this distance has to be traveled at least 4 times. But all in all, satellite Internet is fine for browsing the web, but anything that requires real-time interaction like gaming or video conferencing will be painful. 

In terms of speed, it depends on where you're located. Viasat, for instance, promises 12mbps in mostareas, with some getting up to 100. Not too bad, except that they limit your monthlyusage to somewhere between 12 and 150GB, depending on your plan. And that means you can blow through the limit of the cheapest plan by downloading at max speed for just 2 hours and 13 minutes. Okay, so how does SpaceX plan to tackle these issues with Starlink? 

Well, for starters, they want to launch 12,000 satellites into low-earth orbit, at an altitude of about 550km. That's 60 times closer than traditional communication satellites and reduces the latency to just 25 to 35ms. Not as fast as fiber connections, but comparable to our current cable and cellular networks. One side effect of this proximity is thatthe satellites don't stay in one place. They constantly circle Earth, which has implications for the antenna that you will need to use Starlink. But more on that later. 

As far as speed goes, we don't really know yet what SpaceX will offer. But the US Air Force Research Laboratory has  already tested Starlink aboard an aircraft and achieved 610 mbps. Not bad for an in-flight test! And in fact, this is 6 times faster than my current connection at home. To put this in perspective, this kind of speed would allow you to download a modern game of say 50GB in just over 11 minutes. My current connection, which is no slough, needs an additional hour! SpaceX's COO said that you'd be far happierwith Starlink than with your current service as she promised more bandwidth for the sameprice or lower prices for the same amount of bandwidth. 

Let's move on to hardware. Launching 12,000 satellites seem like a big challenge, and to cope with that, Starlink satellites are small, light, and cheap. They weight just 260kg (560lb) and are launchedin batches of 60 on top of a reused Falcon 9 rocket. Do they really need that many satellites? Well, SpaceX only needs 24 launches to have global coverage, so that's 1440 satellites. Every additional satellite improves the coverageand increases capacity. They already completed 10 launches, so they're making good progress. But by far the coolest piece of hardware on Starlink satellites are the lasers. As said before, the satellites are connected to the Internet via ground stations. But they're also connected to each other withlasers, forming a giant mesh network. And that comes in handy for long-distance communications.  

With Starlink, your request can be beamedup to the nearest satellite, which can use its lasers to forward it to other satellites that are closer to a ground station near California. A nice trick to reduce latency, but it also increases resiliency. If a satellite cannot reach a ground station, because it's offline, it can simply relay the traffic to satellites further away that can access a ground station. But why would you need to boost your satellite? Or why, de-orbit? Well, this is done to prevent creating morespace debris, also referred to as space junk. NASA is currently tracking over 500,000 pieces of debris in orbit around Earth. They vary in size, going from small pieces of paint all the way up to dead and controllable satellites. But the problem is that they travel at speedsof up to 28,000km/h (17,500mph), almost 10 times faster than a bullet. At these speeds, even tiny pieces of paintcan cause severe damage to other satellites in orbit. A damaged satellite can then break apart, make more debris, and can cause more collisions. This endless cascade is known as Kessler syndrome, and if it happens, it will make space activities really hard or even impossible, severely impactingus. No more satellites for GPS, weather predictions,TV broadcasts, and no more International Space Station. SpaceX is doing a lot of work to prevent this. The reason why Starlink satellites are soclose to Earth is not only to reduce latency but also because, at these altitudes, the satellites experience atmospheric drag. If they don't periodically boost themselves up, they will burn up in the atmosphere after a couple of years. Cleaning up after themselves. So this way, SpaceX does not have to worry about an unresponsive satellite. 

Who is Starlink intended for anyway? I mean, most people in western countries havedecent Internet connections, right? Well, I was shocked to find an FCC report that said 21 million Americans or 6% of the population don't have a broadband internet connection. Another study put that number at 42 million. And Microsoft suggested that 157 million Americans don't have Internet that qualifies as 'broadband.' And that's in a country we consider 'well-developed.' If we zoom out and look at the rest of the world, things become even worse. UNESCO estimates that only 55% of house holdshave an internet connection. That's due to bad coverage or expensive service. In sub-Saharan Africa, for instance, a single GB of data costs nearly 40% of the average monthly wage. So that's who Starlink is intended for. Clearly, there is a market for low-cost, high-speed internet access anywhere on Earth. And that market could become really big. Morgan Stanley estimated that the market for satellite internet would be worth over $400 billion dollars by 2040.

Okay, so what do these future customers needto use Starlink? Well, based on leaks, you'll need a relatively small satellite dish outside your house. It's similar to regular satellite dishes buthas a much sleeker design. Using it is dead simple: plug-in & point atthe sky. That's quite different from say TV dishes, which have to be pointed at a specific place in the sky. That's because TV broadcast satellites are in geostationary orbit, they follow the Earth's rotation and are always in the same spot in the sky. Starlink satellites aren't. They constantly float around. To fix this, the Starlink terminal has two features. First, it has built-in motors that can automatically adjust the angle of the dish. Once in that position, it uses a phased array antenna to electronically steer the signal towards a Starlink satellite. It can do this without physically moving the device and can switch from satellite to satellite as they pass over you. 

That brings us to the last part of this essay: has nobody attempted to do this before? Well, mainly because it's very expensive to design satellites and launch them into space. A launch on top of the Ariane 5 or Atlas Vrockets costs over $165 million dollars. That's quite a lot in comparison to SpaceX's Falcon 9, which only costs $62 million dollars or less than $30 million if you choose a Falcon 9 that has flown before. This gives SpaceX a big advantage. Every-time they launch a Falcon 9 for a customer, they can make a profit, AND they get to keep the landed rocket. So essentially, each SpaceX customer brings in more money that can be invested in Starlink. But aside from SpaceX, there is also Amazon and OneWeb who are giving this a shot. Amazon's plans are called Project Kuiper and include over 3,000 satellites, also in low Earth orbit. But currently, they haven't launched any yet, and it's unclear who will launch them. OneWeb meanwhile has plans to launch 650 satellites and so far managed to launch 34. In 2020, however, the company faced financial issues and was saved by the UK government, which took a 45% stake in it. But again, of all these 3 companies, only SpaceX can launch their own satellites. But Amazon might follow in the future when Blue Origin gets up and running. Now personally can't wait to see what SpaceX will do with Starlink and how the market of satellite Internet will evolve.