The Underwater Internet

A wireless world relies on more than cloud technology. Half-century-old cable stations are critical nodes where messages can still be delayed, censored, or intercepted.

Peter Leahy / Shutterstock
In 1962, during a period of technological and political transition in the undersea-cable industry, the Keawaula cable station was built on Oahu’s west shore for the landing of the Commonwealth Pacific Cable (COMPAC). As new telephone cables were being strung across the oceans, the geopolitical balance of telecommunications power was shifting from the British to the Americans.
While the British “All-Red Line,” a globe-spanning telegraph network, had landed only on colonial territory, the Commonwealth system came aground in the United States. Keawaula interconnected British and American transpacific cables for the first time.
As geopolitics were shifting, so was network security. Building one’s cable stations on national territory was no longer a sufficient form of protection. With the threat of nuclear war, Keawaula was better secured by its burial underground in a fallout shelter. And despite being run by a Canadian company, the station was forced to comply with U.S. military specifications, with walls between 18 and 24 inches thick, showers where employees could wash off radioactive material in case of a nuclear attack, and kitchens stocked with enough food for 30 days. These strategies of insulation were designed to shelter signal traffic from the imagined turbulence of the Cold War.
Half a century later, this cable station supports not only phone calls, but a sizeable amount of transpacific Internet traffic. Streaming video, corporate communications, military intelligence, and news from The Atlantic all pulse under Keawaula’s shores.
* * *
Visiting the station in 2009, it seemed far from the bunker-like building of the past. An elegant set of steps led up to the door, the front wall composed of square glass blocks, and a long ramp on the right-hand side makes it wheelchair-accessible. Designed by the Honolulu-based firm Martin & Chock, Keawaula’s aboveground expansion in the early 1980s won several architecture and design awards. One cable engineer described the rebuilt station as a “palace.”
Inside the front door, there is a couch for visitors, a receptionist’s desk, and locally inspired artwork. A caption on one tapestry claims that it reveals “the feeling of Hawaiian water … a vital vehicle for communication for the ancient Hawaiians just as it is today.” The elevated station has unobstructed ocean views.
The manager recalls that the expansion was influenced by a visit from the CEO of Teleglobe Canada, the company that operated Keawaula at the end of COMPAC’s life. With little space for gathering inside, the 20 staff members held their meeting in a small and cramped lunchroom. Afterward, the CEO came outside and saw the station workers joking around and hanging out in the trees. They must have made an impression on the CEO, the manager commented, and— partly inspired by the cable community—Teleglobe built a new, staff-friendly environment that would support local labor through innovative architecture. The redesigned station grounded the communications hub in the connective capacity of the environment of Hawaii yet left its Cold War foundations intact.
Far from the bustling community of the early telegraph days in Honolulu or the telephone networks of the 1960s, however, Keawaula is almost empty—half of its technicians were cut in the downsizing of the early 1990s. There was no receptionist in the lobby to greet me when I arrived, though I was being watched on cameras scattered throughout the complex. Only two people had clocked-in at that point, and the station manager himself had to meet me at the front door. Much of the labor had been outsourced to network operations centers or delegated to computers. The station’s large conference room, complete with a long center table and several high-backed chairs, appeared as if it had not been used in years. AT&T was considering changing it into a storage room. After descending three flights of metal steps in a large concrete stairwell, I was shown the still-functioning battery plants that power the equipment and the multiple backup generators. Everything was redundant—except for the workers. The manager pointed out that a dwindling labor force might be the weakest point of cable networks today.
* * *
Cable stations such as Keawaula are gateways to the undersea networks that currently support 99 percent of transoceanic Internet traffic. Here, transmitted signals are routed from international cables to national backbones and regional networks. Like seaports and airports, cable stations are critical nodes, strategic geopolitical locations where messages can be delayed, censored, or intercepted. One of the revelations of Edward Snowden’s leaks was that intelligence agencies, including the National Security Agency and Government Communications Headquarters, had been monitoring traffic at cable stations. This is no surprise to those familiar with cable history. Since the telegraph era, governments have kept an eye on traffic passing their shores at the undersea cable station.
As a result, these are places where security is absolutely critical. Power can be shut off, bugs can be introduced into the system, and traffic can be intentionally re-routed. We tend to think of cybersecurity in terms of data’s vulnerability to hackers, but data is also vulnerable along the physical route lines of cable networks. While this hasn’t always been the government’s concern—President Barack Obama’s 2013 executive order on cybersecurity made no specific mention of the undersea-cable industry—telecommunications companies have invested heavily in strategies that protect the currents coursing through the cable station.
The modes of network protection have ebbed and flowed with the changes in the surrounding social and political environment. From the 1850s to the mid-20th century, telegraph stations were vibrant and visible hubs of our global networks. In contrast to the secrecy that surrounds them today, stations were depicted on cable envelopes and commemorative postage stamps. Some were central structures for the surrounding area. The multistory cable station in Bamfield, Canada, with its sweeping hilltop view of the Barkley Sound, was compared by one cableman to a French chateau. It had close to 50 rooms, including not only telegraph offices but also accommodations for servants and cablemen, a billiard room, a music room, and a library with books about the British Empire (jokingly referred to as the “All-Red Room”). It was a center of community where events were hosted, and at times, tourists were even accommodated.
The vulnerabilities of the “cable colonies,” as they were known, were markedly different from today. The critical point to be protected in the network was not the station itself, but the cableman. Undersea-telegraph operators literally comprised part of the circuit. One recollects: “During our solitary watches of six hours, we were, for the time, the only links completing communication between Home and the Colonies … We were keeping up that important connection single-handed … Until relieved at the end of a duty, intercourse between Australia and the rest of the world depended entirely upon the man at the circuit.” This was an intense job. Cable transmission required unwavering concentration and physical stamina, no matter the time of day or night.
Yet the boundaries of the telegraph station were porous. Men and women came in and out, and cultures mixed on the empire’s frontier. Social boundaries could also become unclear. Protecting the undersea network entailed regulating the body of the cableman, ensuring that he would uphold strict standards of performance, both in and out of the station. This was achieved by encouraging men to see themselves as part of a tight-knit “cable family,” which reinforced their loyalty to fellow operators and the network itself.
During the telephone cable era, concerns about network protection shifted away from the cableman’s body to the walls of the station. As in the case of Keawaula, many Cold War stations were “hardened,” regulated, and closed off to outsiders. One author remarked: “Growing up in Heart’s Content in the 1950s, unless your father worked at the cable station you probably never got to see inside the door.” Entry was highly selective and carefully monitored.
In this period, a wave of “superstations,” as some transmission engineers described them, were built around the world. These housed an enormous amount of equipment and were designed to endure immense force. Keawaula was required to withstand overpressures of at least 50 pounds per square inch, equivalent to a wind of over 900 miles per hour. In other places, the outside of the station was disguised to look like part of the landscape in case of enemy attack. While access to the interior was limited, cable stations were still promoted as a point of national or local pride. At the commemoration of one cable station, on Guam, the governor announced publicly: “The station is a most welcome addition to our island, both functional and beautiful in its thin-shell, barrel-vault roof design—the first of its kind on Guam.”
* * *
The development of fiber-optic cable stations, which support today’s Internet traffic, yet again reflect the contemporary landscape of security. Instead of a fear of bombs from above, stations are protected from tactical terrorist attacks on the ground. Instead of hardening station walls, companies secure the environment around the station, termed “the buffer zone.” Thinking about the security of one station, a cable entrepreneur narrates his thought process: “Can it take a light plane crash? It’s got a really heavy-duty double-skinned roof. Can it take an 80-kilometer-an-hour, 20-ton truck? Yes, it can because of the way it’s been constructed. What if someone decided to take you out? Could they?”
At some stations in the United States, the Department of Homeland Security conducts site assessments and suggests increased security measures. On the list of changes they would like owners to make, for example, is clearing out the trees between the station and the coastal landing point in order to enable better visual surveillance of the cable route. At another station, the department recommended moving the road back from the building, making it more difficult to access from the street. Today’s cable stations often remain unmarked and without signs. Some do not list their addresses. Others are indicated in policy documents only by coordinates. These strategies—monitoring security cameras, withholding information, and regulating the buffer zone—manipulate a station’s field of vision, keeping the network secure by amassing information about its activities and letting very little of this information out.
Although some companies continue to use the model of “security through obscurity,” others have begun to publicize cable infrastructure in order to appeal to potential customers. Pipe International developed a blog for its PPC-1 cable that not only followed the cable laying, but also broadcast images of equipment and workers in the cable station. The hope is that publicity will ensure trust and investment in their system. After all, information about these networks is already available to the public. Given that there are more operators, from more diverse backgrounds, involved in running the cable network than ever before, the challenge is as much in protecting information as it is in coordinating between the many organizations involved. Cable protection is not simply about regulating behavior or hardening walls. It is about correctly managing the mass of data about networks, their operations, and their surrounding environments.
Although network advertising would have us believe that all of our data is hovering up in the air, zigzagging between light and fluffy light cumulous clouds, in fact, most of it transits underground and underwater. The intense investments in installations like the cable station set imprints for future developments. Like a river sculpts out a path in a landscape, they make it much more likely that subsequent flows will follow the same path. Fiber-optic Internet cables are routed through Cold War-era installations and along colonial cable lines. Earlier regimes of security are latent in our current networks, sometimes literally sitting beneath today’s cable stations.
At times this is good: The “cable family” is still strong today, and despite downsizing, the closed nature of the industry keeps sensitive information at bay. Yet other times, strategies of protection from a former era endanger future cables. For example, the choice to bury stations underground now makes some susceptible to a rising ocean. The strategy of locating telegraph stations near cities in order to keep cablemen connected now makes these lines vulnerable to the anchors of ships and the shovels of local developers. No matter which path we take, what secures the network today may compromise it in the future.

This article has been excerpted from Nicole Starosielski's book, The Undersea Network.