In our "wireless" world it is easy to take the importance of the undersea cable systems for granted, but the stakes of their successful operation are huge, as they are responsible for carrying almost all transoceanic Internet traffic. In The Undersea Network Nicole Starosielski follows these cables from the ocean depths to their landing zones on the sandy beaches of the South Pacific, bringing them to the surface of media scholarship and making visible the materiality of the wired network. In doing so, she charts the cable network's cultural, historical, geographic and environmental dimensions. Starosielski argues that the environments the cables occupy are historical and political realms, where the network and the connections it enables are made possible by the deliberate negotiation and manipulation of technology, culture, politics and geography. Accompanying the book is an interactive digital mapping project, where readers can trace cable routes, view photographs and archival materials, and read stories about the island cable hubs.
About the Author
Nicole Starosielski is Assistant Professor of Media, Culture, and Communication at New York University.
Read an Excerpt
The Undersea Network
By Nicole Starosielski
Duke University PressCopyright © 2015 Duke University Press
All rights reserved.
Circuitous Routes: From Topology to Topography
At first glance, the Arctic Fibre cable system—the 15,600-kilometer network that would link England and Japan via the Arctic Ocean—makes little sense. With plenty of terrestrial networks spanning North America, why pioneer a circuitous route through one of the least populated areas of the world and an expanse of inhospitable environments, where the cable would be covered by several feet of ice during much of the year? Why spend a projected $640 million to connect Arctic communities with a high-capacity cable when they will likely never use a significant percentage of its capabilities? Given the extensive number of data connections and surplus capacity between London and Tokyo, the cable's two endpoints, why does the world need another cable? Similar questions could be posed to other cable projects seeking to spend millions to link already connected endpoints, from the Hawaiki cable, a proposed system between New Zealand and the United States, to the South Pacific Island Network (SPIN), which was designed to hop from island to island across the Pacific.
One answer to these questions is that these cables offer diversity, a redundant and geographically disparate pathway that would make signal traffic more reliable. If established, these systems (or so their projectors argue) would better insulate our information flows from people, environments, and other forms of interference encountered by networks. Arctic Fibre avoids pressure points on existing routes between Europe and Asia, including the earthquakes of the Luzon Strait, South China Sea, and western Pacific, as well as social and political unrest that make the Suez Canal and Mediterranean particularly turbulent environments. It satisfies a need for routes between Europe, the Middle East, and Asia that skip the United States, an option desired by Asian and Middle Eastern carriers who want to secure their traffic from potential monitoring on U.S. territory. It does not hurt, of course, that Arctic Fibre would also be the quickest route from London to Japan, at 168 milliseconds for a transmission, enabling the company to leverage the market of high-frequency traders who seek to capitalize on the path with the shortest transit time.
The diversity of our global networks, long a concern of telecommunications companies, has recently come to the forefront of policy makers' attention. Prompted in part by a series of very public cable failures, representatives of the telecommunications industry, members of the financial sector, technical experts, and policy agents came together in 2009 at the Reliability of Global Undersea Communications Cable Infrastructure Summit (ROGUCCI) to discuss how to make global network infrastructure more resilient. The summit's participants concluded, among other points, that even though individual systems were highly reliable, on a "global level, the overall interconnectivity of the continents violates a fundamental reliability design principle—avoid single points of failure." Participants pointed out several "geopolitical chokepoints" where cable paths were funneled together and warned that a disaster in one of these areas "could cause catastrophic loss of regional and global connectivity." Our global network, with no overarching body to ensure its robustness, has been constructed in an ad hoc way by companies with a range of different interests, creating an infrastructural reality that runs counter to the common imagination of the global Internet as a distributed mesh, able to easily circumvent any attack in a specific geographic region and route traffic around any disruption.
This chapter argues that in order to understand the geography of signal traffic, we must move beyond network topology, the observation of the geometric or mathematical distribution of nodes and links, to consider network topography, the way that infrastructures are embedded into existing natural and cultural environments. Paying attention to network topography counters several widespread assumptions about the determinants of cable routing. First, there is a terrestrial assumption: people believe that undersea cables are used only when a terrestrial cable route is not available. Researcher Linda Main argues that since running cables over land is less expensive, "undersea cables have traditionally been used only between continental landmasses, where terrestrial links are not feasible." Cables are assumed to follow the shortest geographic route between continents in order to minimize underwater segments. When cables do not land at the closest terrestrial points between landmasses, their geography is often explained by an urban logic that assumes that cables directly connect the centers of major cities (reflecting a broader focus on the city as the critical unit of analysis for telecommunications infrastructure). A third assumption is that cable routes are driven by demand: they are extended to places where there is a surplus of signals and not enough cables to carry them. This is an anthropocentric and equalizing view of the cablescape: we assume that, like people, cables must be easier to sustain in accessible and central environments and are simply extended to users who lack connection.
This chapter shows that even though terrestrial, urban, or demand-driven logics may at times play a role in determining cable routes, our global infrastructure has been constructed in relation to historically specific social and environmental imaginations. Security, in the broadest sense, has always played a critical role: companies route cables in ways that insulate them from potential interference in their surrounding environments, ranging from natural disasters to anticipated geopolitical friction. Counter to any presumed terrestrial logic, a secure network has most often been an underwater one—territorial politics have made laying routes on land incredibly difficult. Douglas Cunningham of Arctic Fibre argues that the ocean and its ice will serve as a layer of insulation. He commented in one presentation that "60% of cable breaks are from ship anchors and trawls" and observed that "this route will actually be safer than the other routes because we have, on 34% of the length, ice cover for seven months of the year. And that is protection." The deep sea has been the safest zone for a cable because it is farthest from people and subsea engineering is more reliable than terrestrial engineering (components function for decades underwater). The closer a cable gets to shore, the more heavily armored it must be. Although it is true that many cables do connect urban centers (which remain significant endpoints for data flow), landing points are rarely established in the heart of cities. Rural and suburban environments, like the ocean, protect cables from the potential turbulence of human traffic in densely populated areas. The Arctic Fibre cable, though described as going from London to Tokyo, actually lands far outside these cities.
Complete insulation of a cable network is never entirely possible (or profitable). Such insulation would make it impossible for cables to interconnect with the sources of traffic necessary for operation: other undersea cables; domestic fiber-optic systems; and rail, road, and air transport networks. As Stephen Graham and Simon Marvin argue, for this reason new infrastructures are often layered over existing transport and resource systems and rarely develop in isolation. Cables have often been laid to locations where there is already enough capacity as a way of generating competition, stimulating increased signal traffic, and lowering prices. Although many people assume that cable networks are demand driven, advocates of interconnection often make the economic argument that the networks drive demand. Rather than viewing cable laying as a struggle to overcome the ocean or link urban nodes, we might see it as a process of securing routes from turbulent environments and interconnecting them with existing cultural and technological circuits.
Moving from the techniques of the telegraph era, through those of the coaxial period, and to those of today's fiber-optic networks, this chapter describes how investments in cabled environments have shaped the topographies of three generations of transpacific cables and provides a context for the situated geographic analysis of the following chapters. The global telegraph network, constructed in the second half of the nineteenth century, drew support from colonial networks and pioneered the use of the ocean as a layer of insulation. The reinvention of cable technologies after World War II involved negotiations between existing routes of empire, emerging forces of infrastructural decentralization, and a new club system of cable laying (described below). Today, the lines along which the Internet flows evidence a similar push and pull: deregulation and privatization have helped pioneer a new cable geography, which nonetheless is layered into a geopolitical matrix of preexisting colonial and national routes. This genealogy will be familiar to some readers, as historians have analyzed the complex histories of telecommunications systems in detail elsewhere and have well documented the importance of economic and political influences on cable development.
The aim of this chapter—with its focus on security, insulation, and interconnection, especially in the construction of new routes—is to recast our understanding of cable networks to better account for their contingent and material qualities, projecting them as "fragile achievements" that reflect both the ideals of the past and the geographic contours of the earth. I suggest that we view cables not simply as technical systems, but as long-lasting contours in the environment—places where capital, labor, and knowledge have been sunk into the earth's surface, "accumulated imprints" of investment in particular spaces. As a riverbed shapes the flow of water through it, such contours affect—though they do not determine—the direction and force of subsequent circulations. In pursuing the most secure routes, individual cable companies have solidified the network along relatively few lines. As a result, in many places our cable network remains circuitous rather than direct, underwater instead of on land, rural rather than urban, and connects places that are already connected. Perhaps ironically, this tends to perpetuate unequal topographies of global exchange and has left us with a relatively concentrated, semicentralized, and precarious geography that is now proving expensive and difficult—as the case of Arctic Fibre shows—to diversify.
Copper Cable Colonialism
The geography of telegraph routes in the late nineteenth century followed transportation and trade routes, many of which had been pioneered by British colonial investment and served to support existing networks of global business. Cables often landed at the same sites as ships, not only to interconnect with marine transport (the shipping industry was a significant user of cable systems), but also because these sites were often geologically appropriate for both forms of traffic (each required a smooth transition between land and sea). Cables were strewn between ports in the Pacific, from the major hubs of Hong Kong and Singapore to outposts such as Port Darwin. It is no coincidence that Australian cables were brought ashore ten kilometers south of Sydney at Botany Bay, where Captain James Cook claimed the continent for England in 1770 and where the French explorer Jean François de Galaup, comte de Lapérouse, landed in 1788 (figure 1.1). Even at the network's more remote landing points, from Bolinao in the Philippines to Banjoewangi in Indonesia, there was almost always an existing set of colonial infrastructures, however limited, that could support the new stations.
The selection of network routes during this period rarely embodied a completely terrestrial or urban approach. Instead, it represented a balance between the need to interconnect with existing populations and infrastructures and the affordances of an area's natural and social topography. For example, in 1876 the Eastern Telegraph Company laid a trans-Tasman cable from Botany Bay to Nelson, where the company sought to connect with economic trade that was then centralized in New Zealand's South Island. Nelson was neither the closest geographic point to Australia, which was determined to be too remote to support even the cable station, nor a major commercial center—Christchurch was the endpoint for much of this traffic but was deemed too far away. Rather, this route, like many others, was a compromise: Nelson gave the cablemen sufficient access to existing infrastructure but did not needlessly waste expensive undersea cable. Landing a cable in Nelson, as was true for other locations such as Botany Bay and San Francisco, entailed laying a longer marine route between landmasses than was absolutely necessary. Although arguments were certainly made about the need to directly connect urban hubs and to minimize undersea routes, it was not uncommon for such statements to be made by individuals with specific commercial interests or geographic pretensions.
More than any single technological justification, security and insulation were important to the negotiation of cable routes. When undersea cables were first laid in the 1850s, the most significant challenge of installation was protecting them from water. Terrestrial wires could remain without physical insulation since air, a nonconductive medium, kept signals from diffusing, but insulation was of paramount importance for an undersea cable. Without it, the signal would easily dissipate into a conductive ocean. In his history of cables, Willoughby Smith dates the origins of undersea cabling not from the successful conduction of signals but from adequate insulation: he considers the 1850 cable from England to France the first pioneer of undersea telegraphy, despite the fact that it conveyed indecipherable messages, because it proved that a circuit of power could be insulated from water. British companies were able to dominate the cable business throughout the late 1800s due not only to their ability to master cable-laying technology and monopolize technical-support infrastructure, but also to their development of an effective form of insulation using gutta-percha, the rubber-like gum from Malaysian trees, and their control over its extraction and distribution.
As the network expanded, aquatic environments once perceived as a threat became the most significant form of cable protection. Counterintuitively, the weakest parts of cable networks were the terrestrial links and coastal segments; in contrast, the deep ocean was the safest zone. Overland cables could become targets during popular revolts, as happened during the Boxer Rebellion in China, when telegraph lines were sabotaged, cutting off communication between local diplomats and the British government. Australia's transcontinental line was often disconnected by Aborigines and the occasional lost cowboy who wandered across the landscape, cutting the line in order to be rescued by cable repairmen. Most natural breaks occurred in shallow coastal waters where tides dragged the cable across rocks. In contrast, very few breaks occurred in the deep ocean, where cables functioned for decades without interruption. To establish political support for the transpacific cables, for example, the argument was made again and again that "in the depths of the Pacific Ocean, the cable would be absolutely safe from interference." In addition to keeping the network safe from perceived physical disruption, running cables through extraterritorial space enabled users to circumvent transit taxes that might have been levied by countries as messages passed through their territory.
Excerpted from The Undersea Network by Nicole Starosielski. Copyright © 2015 Duke University Press. Excerpted by permission of Duke University Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
Preface. Edges ix Acknowledgments xv Introduction. Against Flow 1 1. Circuitous Routes. From Topology to Topography 26 2. Short-Circuiting Discursive Infrastructure: From Connection to Transmission 64 3. Gateway: From Cable Colony to Network Operations Center 94 4. Pressure Point: Turbulent Ecologies of the Cable Landing 138 5. A Network of Islands: Interconnecting the Pacific 170 6. Cabled Depths: The Aquatic Afterlives of Signal Traffic 198 Conclusion. Surfacing 225 Notes 235 Bibliography 263 Index 281
What People are Saying About This
"Nicole Starosielski's The Undersea Network is as expansive as its subject, revealing the networks that make global communication possible as vital worlds unto themselves. In most stories of new media, infrastructure fades into the background. But Starosielski flips the script, making infrastructure the star, vividly describing the places, the people, the institutions, and the politics that constantly work to make global communication possible. In the process, The Undersea Network offers new insights into globalization and digitization. It also teaches us how to study large and largely invisible technical and cultural institutions. Coupled with its groundbreaking digital companion (www.surfacing.in), The Undersea Network will transform our understanding of the networks that make modern media possible."