It's not often that one thinks of Google as being a heavy engineering firm. But right now it’s laying cable. And it’s some cable: at 6,437km long, the Dunant cable, which is set to come online this year, is expected to be able to transmit a record-breaking 250 terabits of data per second, the equivalent of the entire digitised Library of Congress three times a second. But what’s striking is that this cable will run under the Atlantic, all the way from Virginia Beach in the US to the French Atlantic coast. Google, in fact, already has some 14 cables running through oceans across the globe, whizzing your emails, updates and cat videos to all and sundry, streaming your box sets, feeding your news.
Indeed, look beyond Google’s efforts and the planet is criss-crossed with cables, each not much thicker than a garden hose but carrying multiple pairs of glass fibre lines about the width of a human hair. These fibre optic strands are wrapped in urethane and copper, then again in urethane, and sometimes stainless steel or Kevlar. The cables are laid at a torturously slow pace—around 10km/hr—by specially commissioned ships using a kind of plough that digs a small trench on the ocean floor before lowering the cable, the trench then being covered over with sand by the sway of the sea. Closer to shore, specially constructed concrete channels might be built on the seabed. And, for all that you might think you live in a seemingly wireless world, it’s these fragile, physical arteries that carry 99 percent of all international communications.
“The fact that these cables even exist is just so counter- intuitive,” says Nicole Starosielski, associate professor at New York University and author of The Undersea Network. “When we think of media now we think of the likes of teleportation, of it being this futuristic thing, when in fact it’s a long-distance engineering structure. It just doesn’t resonate with our notions of the digital revolution, of the Internet as being this ethereal thing. People struggle to grasp that it all goes down a tube.”
That seemingly old-fashioned idea of sending our messages through a cable is, for sure, rather old-fashioned. The first undersea telegraphic cable was laid 170 years ago, across the Channel between England and France. Just eight years later, two battleships met in the middle of the Atlantic, where they spliced together the two ends of a 4,023km-long, 1.27cm- diameter cable, connecting two continents by telegraph for the first time.
It wasn’t the most sophisticated by today’s standards, though it did make use of a number of recent technological advances: the development of iron rope, the introduction of gutta-percha —a rubber-like tree sap found only within the British Empire —and the first deep-sea soundings of the Atlantic by the US Navy. It wasn’t all that reliable either. The 1858 cable, which took three attempts to lay, failed after just a few weeks. And it certainly wasn’t the fastest. Queen Victoria sent the first short message to US President Buchanan by Morse code. It took 17 hours to get there.
But the idea of connecting the world through what were then strands of copper would be revolutionary. It only took one notable success—the British government being able to send a message to its army in Canada countermanding an order to return to England, saving a small fortune in the process—for the commercial benefits of fast communications to become evident, inspiring a boom in cable-laying. Today there are some 400 cables in operation, stretching over 1.2 million kilometres, enabling communication at speeds around 16 million times faster than your typical home Internet connection. And there are many more to come, each more sophisticated than the last.
Indeed, for all that the basic premise remains the same—a pipe on the seafloor—what goes into that pipe is advancing all the time. New types of fibres and more sophisticated modulation have been among developments introduced since the 1980s, as have improved optical amplifiers—undersea cables need these travel trunk-sized devices every 60km or so to boost the signal.
There are now experiments looking at the development of fibres with multiple cores to allow multiple optical signals; or fibres big enough to allow those signals to follow different paths through the same fibre; or using fibres that work by opening additional wavelengths for transmission, such as the Oregon-to-Japan Faster Cable built by a consortium comprising Google and five Asian telecom companies. Last year even the International Space Station took delivery of hardware, devised to make a kind of optical fibre, massively improved in its transmission quality—the only problem being that it has to be made in low or zero gravity.
But in the medium term it all looks to be about trying to pack more fibre pairs into a cable, though even the very latest
cable designs can only handle 16 at most. You can’t just make the cable thicker because it needs a lot of power to function, which can only be drawn from either end. Improved electronics at the landing stations—where the cables touch terra firma at either end—may boost speeds. And in fact it’s improvements here, rather than in the cables, that have driven costs—calculated per gigabit per second transmitted—down.
“After all,” says Geoff Bennett, director of solutions and technology for cable manufacturer Infinera, “for some companies ad revenue is being lost for every millisecond you’re waiting for your video to load. That’s why the longer you can keep a cable operational and competitive, the better. But the glass in fibre optics ages and every repair adds a small loss [to efficiency]. Ultimately any cable will have an economic performance and if the data it carries falls below a certain level, then it’s no longer economical given its running costs.”
That may not be so much the case if you have bottomless pockets. Until recently most undersea cables were laid by telecommunications companies. They formed consortia—often between companies you’d imagine would be competitors— to share costs and risks. They aimed to use the cables to drive their own businesses. Then came private companies, which saw the opportunity to sell cable capacity to telecom companies and others. Now it’s the technology giants that are getting in on the act to connect their data centres around the world. In 2018, Google, Amazon, Facebook and Microsoft together owned or leased over half of all undersea bandwidth, only adding to concerns about the growing power of Big Tech and who controls the Internet. Google alone is said to need to double its capacity every year just to sustain the appearance of seamless cloud computing.
"In this business often the real money is made not in the cables but in all the equipment around them… in the past a company that needed to move goods around might have built a railway. They just want to move their goods—data—as cheaply as possible."
“[Big Tech’s involvement] may bring all sorts of questions with it,” says Starosielski. “They have so much money that it gives them so much say over the network. But that means that, say, running cables to small islands is less likely, or that other cable networks may become obsolete. In many respects it’s bad for regional sovereignty. There’s an argument that we need more diversity of network ownership, not less.”
But then building undersea cables is a risky business, not least because it’s still hugely expensive: an undersea cable can easily cost between USD350 million and USD500 million, and take three years to actually lay, not counting the years of planning required ahead of that. During the dotcom bubble, phone companies spent some USD20 billion laying undersea cables, only for the expected explosion in Internet traffic not to come, forcing them to offload capacity at bargain prices. Now there’s another investment boom—80 percent of it from Big Tech—such that more cable was laid in 2018 than in the previous two decades combined. Facebook, Amazon and China Mobile are, for example, working on an ambitious cable to connect Hong Kong and Singapore to San Francisco. Climate change is, for the first time, even encouraging some to consider laying cable under the newly accessible Arctic, all just to shave off some of those milliseconds.
“It’s a tough, challenging business in today’s environment, a bit hit and miss,” concedes Byron Clatterbuck, CEO of Seacom, a cable builder focused on demand in Africa, where it developed the continent’s first private land cable. “None of the investors in our cable had a background in telecoms but they shared the vision we had because 10 years ago there was, effectively, no Internet connection in South Africa. But undoubtedly it was risky for them. There’s a huge need for capital upfront, you’re not going to see a return soon, there might be delays and over that time the market might change. In Africa there are risks of war or governments being inconsistent with regulations.
“In this business often the real money is made not in the cables but in all the equipment around them,” he adds. “It’s not surprising that the cable industry has had these massive boom- and-bust cycles that have seen lots of companies go bankrupt. Or that most of the cables being built now are by the likes of Facebook or Google. Building a cable network is just part of their operating costs now, much as in the past a company that needed to move goods around might have built a railway. They just want to move their goods—data—as cheaply as possible.”
That riskiness is only exacerbated by the fact that cables, being physical things, have a shelf life of around 25 years and are easily damaged—at least one of the world’s undersea cables is said to be damaged every couple of days.
“These cables are wonders of engineering, laid at the bottom of the ocean, under all that pressure and yet proving as reliable as they are—such that if one goes down, as [an Internet user] you’d hardly even know it,” says Alan Mauldin, research director for telecom market research company Telegeography.
“Yet look at the cable map of the world and in a way it’s surprising that there are so few of them. It seems a lot but cables differ in capacity. They get old and need replacing.
And, yes, they do get broken.”
While tales of sharks biting into cables are apocryphal—though they may well be attracted to the cables because of the magnetic currents they generate—events do conspire to sever these essential communications links. Undersea rock slides, shifts in the tectonic plates, a geologically active seabed—as is the case across Southeast Asia—and storms can all cause cables to break. An earthquake off the coast of Taiwan in 2006 knocked out eight cables, causing outages across the Far East. Hurricane Sandy, in 2012, knocked out several exchanges where cables linked the US and Europe.
It even brought about a change of thinking: with so many cables running into the New York region, Microsoft and Facebook decided that their jointly funded Marea cable— which came online in 2018—should come in further down the coast, in Virginia. It’s a wise move, seeing as so many cables tend to be laid to meet in busy hubs on land, only consolidating risk. Starosielski’s research suggests that if, for example, the cables off Egypt were broken it could take down a third of the global Internet.
But most cable faults—around two-thirds of them—are the product of deep-sea trawlers’ fishing nets or ships’ anchors messing with them. In 2007 fishermen managed to drag up kilometres of cable off the coast of Vietnam, disrupting communications in the region for months. The following year it was an abandoned anchor that damaged a cable and blacked out most of North Africa and the Persian Gulf.
Small wonder then that there’s talk too of the potential for these undersea cables to be tampered with maliciously, either tapped for the purposes of espionage or clandestinely cut to prevent communications in a way that would slow an economy. These cables, after all, run for thousands of isolated kilometres, making them especially vulnerable.
Trying to hack cables isn’t a new idea either. During World War I the British severed Germany’s international telegraphic cables, forcing it to ask the then still neutral US to deliver a transatlantic message via its series of diplomatic cables from Berlin to Mexico. But these cables passed through a relay station on the westernmost tip of England, where all traffic was intercepted by British Intelligence. The relevance of the telegram? It proposed a military alliance between Germany and Mexico, thus helping to bring the US into the war. Likewise during the Cold War US submarines tapped Soviet cables in the Sea of Okhotsk.
But these were pre-fibre optic times and to tap the lines underwater—as opposed to at the point where a cable comes into land—is another matter now. Forgetting for a moment that some 10,000 volts runs through these cables, the precise point of the tap—or the fault as it might appear—would be instantly revealed through a drop in signal power. A ship would be despatched to haul up the cable to effect a repair. Then the sheer amount of seriously encrypted data traffic would make collection and analysis in itself a gargantuan task.
"We’re already at speeds by which you could download, say, the entirety of Game of Thrones in 1/100th of a second, yet there’s always demand for faster still."
“I’m not saying it can’t happen and the NSA [National Security Agency] probably has a fleet of mini-subs that could do it. But it’s not as easy as it might be shown in a James Bond movie,” laughs Infinera’s Bennett.
Not that this stops nations thinking, or worrying, about its feasibility. The US is said to have a submarine, the USS Jimmy Carter, outfitted for the job. And, in 2017, Australia blocked a plan for Huawei to install an undersea cable linking Sydney with the Solomon Islands, effectively on the grounds of national defence. Instead Australia funded most of the Coral Sea Cable System itself. Beijing’s interest in constructing undersea cables—and the perceived influence that comes with it—is a growing sore point in international relations.
And yet, despite all this, it looks as though cables are here to stay. No technology has yet come along to pose a challenge. When radio was developed in the 1920s, claims were made that cable had become obsolete, yet the lack of security on wireless transmissions helped keep the cable industry going. When communications satellites were first launched, they said the same. Cable traffic did see a dip, but the time delay in round-trip satellite transmission again had most communications going back down the wire. That hasn’t stopped the likes of Elon Musk and Jeff Bezos proposing new satellite swarms to beam the Internet around the world. Perhaps Bezos doesn’t know that, by one estimate, his proposed low-Earth orbit ring of satellites would together offer less capacity than a single fibre optic pair on the Marea cable.
Speed, of course, isn’t everything. Slower cables get laid around the Cape of Good Hope to avoid the bottleneck of the Suez Canal, for example, where there are already a dozen cables. But it’s seeking to avoid delay, or latency as they call it in the business, that is likely to keep demand for what has become the virtually instantaneous communications by cable to the fore.
The Internet is said to have a second digital revolution in the offing, as more and more unconnected parts of the world get online for the first time; as more and more technology becomes connected—the so-called Internet of things; as the likes of driverless cars become a reality; as the Internet enables the likes of remote surgery by robot. But the fact is that all of these advances depend not so much on the Internet as on the speed of connection—and, as yet, nothing trumps cable.
“Data traffic is increasing all the time and while that pressure grows there’s a need to maintain the positive experience that consumers have [in using the Internet] to meet that expectation that it will work at a near instantaneous pace. We’re already at speeds by which you could download, say, the entirety of Game of Thrones in 1/100th of a second, yet there’s always demand for faster still,” says Jen Robertson, president of field operations for AT&T.
“We’re only going to see a need for more capacity in these cables because I don’t see anything coming along to replace them soon,” she adds. “Of course, we’ll also increasingly talk about wirelessness and accessing the Internet through the devices we hold in our hands, wherever we may be. Most of us will probably keep thinking that the whole network is wireless. And yet so many of the things the Internet will allow, which we’re yet to enjoy, will depend on cables. They may be the unsexy parts. But they make the whole thing work.”