91av

AISHTAN SHAKARIAN knew there was money to be made from the internet. So she took a spade into the woods near where she lived, about 50 kilometres outside the Georgian capital of Tbilisi. The 75-year-old hoped to dig up copper wiring to sell for scrap. Instead, she cut through a fibre-optic cable – worthless to her, but priceless to the millions of people in neighbouring Armenia left staring at blank screens for 12 hours. She had cut off the country’s internet.

The 2011 incident shows how easily this can happen with underground cables, and those under the sea are even more vulnerable. Every few days, an earthquake, anchor or boat damages one of the roughly 430 sea-floor cables. Tonga went offline for nearly two weeks in January after an underwater cable was cut.

In some ways, as an isolated island nation, Tonga is lucky to have this connection. The cost of laying cables to remote places means only about 10 per cent of the planet’s surface has terrestrial communication links. According to the UN, nearly half the world’s population has never been online.

To reach them, and ensure everyone has a reliable connection, billionaires like Elon Musk, Jeff Bezos and Richard Branson plan to reinvent the internet – to free it from its Earthly roots and build a wireless web above our heads. Balloons in the stratosphere, constellations of satellites, cruising drones – there is no shortage of ideas. Pull this off and humanity’s greatest information repository would find a dazzlingly futuristic home. To make it work, we just need to deploy some old technology, albeit in a highly unusual way.

The internet is a gigantic network of computers. When you type an address into a browser, you are instructing it to connect with another machine that could be anywhere in the world. Most of those connections are via cables. Even smartphones only use radiowaves to connect the last few hundred metres to a cabled cell tower. But longer stretches are possible. Satellite internet today often uses relay stations far enough from Earth that they remain in a steady “geostationary” orbit: as seen from the ground, they are in a fixed position. Pinned 35,000 kilometres above the equator, they can serve a wide swathe of land. But the 70,000-kilometre round trip adds a lag of half a second or more to signals, an annoyance that disrupts voice calls and makes multiplayer online gaming or high-speed financial trading impossible. On top of that, download speeds are slower than modern cable connections and subscriptions are pricier. The set-up also requires a large dish and a clear view of the sky.

One alternative that tech companies have recently considered is the stratosphere. From around 10 to 50 kilometres up, this layer of the atmosphere is high enough for a transmitter there to serve a city-sized area below, yet low enough that a phone could communicate with it without the need for a receiver dish.

Better still, putting things in the stratosphere is easy compared with space.

Hungry for extra customers, Facebook and Google built prototype solar-powered drones that could loiter about 20 kilometres up for weeks, beaming down the internet. But these projects are now on hold following crashes and damage when landing the feather-light aircraft. Other companies, including Boeing and Airbus, are working on similar drones, but the technology is far from proven.

Google’s next idea sounds kookier still: a train of gracefully floating balloons delivering data to those below. The balloons, which provided connectivity following natural disasters in Peru and Puerto Rico, have been spun out into a company called Loon that plans to launch a commercial 4G service in Kenya soon. However, balloons are unlikely to do much for most of the offline billions as many places lack reliable winds or convenient launching places. If you are a tech billionaire with a global vision, only a truly worldwide service will do.

Luckily, there is a middle way. What if we put transmitters lower than the lonely geostationary orbits with their unpleasant lag, but higher than the stratosphere with its fickle weather?

There is a price to be paid for using this low Earth orbit (LEO). From the surface, satellites in LEO appear to zip from horizon to horizon in about 10 minutes. For continuous service, then, you need multiple satellites flying in quick succession and enveloping the globe, all sending signals from one to another in a smooth relay. The firms Iridium, Globalstar and Orbcomm each have a few dozen LEO satellites that already offer basic, slow internet services. Get the tech right and turning these limited services into something transformative becomes a numbers game. And if there is one thing tech billionaires think they understand, it is scale.

“Only about 10 per cent of the planet has terrestrial communication links”

This occurred to Bill Gates years ago. In the 1990s, he backed a start-up called Teledesic that envisioned a mega-constellation of 840 LEO satellites relaying radio signals from one part of Earth to another. The company had big plans: to deliver affordable broadband to 95 per cent of the world’s surface.

It never happened. Teledesic folded in 2002 without launching a working satellite, having neither developed the technology, nor raised the billions of dollars required. But a fresh generation of billionaires is ready to try again, buoyed by cheaper launch costs and new, more powerful technologies.

Take OneWeb, a company backed by Airbus, computer chip-maker Qualcomm and entrepreneur Richard Branson. It put its first six satellites – costing $1 million apiece – into orbit in February. The firm says 600 satellites connecting users to 40 or so ground stations should be in place and providing a service by 2021. Yet OneWeb’s satellites only plug gaps in cable internet. The last leg of each data packet’s journey to the user may be through space, but at other times it travels through the same wires as the rest of the internet.

Other companies are trying something more ambitious (see “Traffic in the sky”,). Elon Musk’s SpaceX, Canadian company Telesat and Luxembourg-based LeoSat all plan to create an end-to-end, space-based internet system in the early 2020s. Each of their fast-moving satellites must be able to communicate with others in their respective networks, relaying data in hops from one side of the world to the other. All three companies have the same idea of how to do it: lasers.

Theoretically, lasers are a smarter way to communicate in space than radio waves, says astronautics engineer . They are more energy efficient, so transmitters and receivers don’t need to be so large. “Physical size can be smaller with a laser system,” she says. Lasers also sidestep the issue of working within the increasingly crowded radio spectrum. Radio transmissions are highly regulated and the frequencies that would be most practical for small receivers are already taken. In contrast, laser beams, with their tight focus, almost never cause interference with other services and are largely unregulated.

“SpaceX is aiming to have 40 million subscribers to its space internet by 2025”

They are also fast. Imagine a financial trader in London wanting to access the latest from the New York Stock Exchange. If her connection were routed through SpaceX’s planned constellation of nearly 12,000 Starlink satellites, data might reach her in 45 milliseconds, according to . That is half the time it currently takes via fibre-optic cable, an advantage that could be worth millions. expects its laser internet to attract 40 million subscribers and $30 billion in revenue by 2025, with Starlink eventually helping to fund Musk’s other dream – of colonising Mars.

But lasers have their own problems. The main one is successfully pointing a beam with the thickness of a human hair at another satellite speeding past at thousands of kilometres an hour.

Still, this isn’t totally virgin territory. As early as 2001, the European Space Agency established the first inter-satellite laser link, between large satellites in low Earth and geostationary orbits. And in 2013, NASA used lasers to beam a picture of the Mona Lisa from Earth to a spacecraft circling the moon.

Cahoy is now working on a mission called that will pack lasers capable of high-speed communication over hundreds of kilometres into . These satellites will use wide-angle lasers as beacons to locate each other, then activate narrower beams for their high-bandwidth link. NASA is due to launch two of them in 2020 to test the idea and Musk could be using similar technology for Starlink. SpaceX already has two prototype laser satellites in orbit – though each is the size of a car. It plans to start deploying commercial satellites this month, the first of hundreds of launches in the years ahead.

High-speed dance

Linking two satellites with lasers is one thing. Maintaining connections between thousands, 24 hours a day, is another entirely. “Talking to the satellite in front of or behind you in an orbital plane is actually pretty easy,” says Cahoy. “When you try to go from your plane across to another plane, that gets a little bit more challenging.”

SpaceX’s planet-spanning constellation will feature criss-crossing orbital planes with satellites whizzing past each other at high speed. Choreographing this dance entails knowing exactly where each spacecraft is at any moment, and where it is heading. In an agreement signed last year, the company asked NASA for technical support in choosing GPS hardware and in running the agency’s orbital calculation software on Starlink’s on-board processors.

The company also plans to have more than 7500 of its satellites orbiting at an altitude of just 340 kilometres, far lower than any existing communications satellites. That will keep transmission delays to a bare minimum and reduce power needs, but it will also subject the satellites to increased drag from wisps of the atmosphere. Without regularly firing their thrusters, the satellites would be dragged down and burn up within weeks. Even so, each Starlink is expected to last just six years or so before exhausting its propellant and heading Earthwards (see “What goes up, must come down”). To sustain its fleet, SpaceX will have to commit to the task of continually replacing all the satellites that re-enter the atmosphere – as many as five a day after the first six years.

An even bigger issue is how consumers will access the LEO satellites from the ground. Although lasers are likely to operate at eye-safe infrared frequencies, the beams would be blocked by even mild cloud cover. So all the laser satellite projects plan to use radio waves to communicate with the ground from orbit, which means using pricey steerable antennas (see “Cutting the cord”).

This skewers any claim that space internet systems are about getting the poorest people online, says media scholar . Even the pizza-sized receiver that Musk says SpaceX is developing will cost $100 to $300. This will be “prohibitive for unconnected users in much of the world”, she says. “While connecting the unconnected sounds good from a marketing perspective, there is little clear evidence that such a model can deliver profits to companies and investors.”

There is also scepticism that many users in richer nations have an appetite for additional gadgets and subscription plans. “By the time any of these constellations is fielded and ready to supply a paying service to real customers, we’re going to have 5G infrastructure in Western countries, and Africa is going to be covered in 4G towers,” says aeronautics engineer .

However, some think that wireless companies will welcome mega-constellations as partners rather than rivals, sharing satellite signals so that customers can access space internet from their phones, without having to buy an expensive antenna.

Another potential benefit for those living in repressive countries is that satellite internet could theoretically skirt censorship measures such as China’s Great Firewall. It is unclear how that would play out, however. When Musk raised this possibility in 2015, he suggested that SpaceX wouldn’t serve users in such countries. “Any country could say it’s illegal to have a ground link,” he . “We could conceivably continue to operate, but then they have a choice of, do they shoot our satellites down, or not? China can do that.”

Elsewhere, a space-based internet could be simpler and more reliable than accident-prone cables. Invisible beams of light are largely immune to hacking or tapping, and if one or two satellites fail, replacements could be quickly launched.

But betting on space lasers puts an awful lot of eggs into one basket. A powerful solar storm or a cascade of satellite collisions, as depicted in the film Gravity, might damage or disable entire constellations in a blink. However, in a world already dominated by a small number of tech firms, perhaps the most disturbing aspect of tomorrow’s mega-constellations would be having power over such critical infrastructure concentrated in just a few hands. If you thought the story of Aishtan Shakarian’s shovel was worrying, imagine having a few powerful individuals in control of the space internet’s off switch.