The End of the Internet?

The World Wide Web celebrated its 25th birthday recently. Today the global network serves almost 3 billion people, and hundreds of thousands more join each day. If the Internet were a country, its economy would be among the five largest in the world.

In 2011, according to the World Economic Forum, growth in the digital economy created 6 million new jobs. The McKinsey Global Institute estimates that transborder online traffic grew 18-fold between 2005 and 2012 and that the global flow of goods, services, and investments—which reached $26 trillion in 2012—could more than triple by 2025. Facebook has launched a major initiative, in partnership with tech giants including Samsung and Qualcomm, dedicated to making the Internet available to the approximately two-thirds of the world’s population not yet connected. Cisco forecasts that between 2013 and 2022, the so-called Internet of Things will generate $14.4 trillion in value for global enterprises.

Yet all of this growth and increasing connectedness, which can seem both effortless and unstoppable, is now creating enormous friction, as yet largely invisible to the average surfer. It might not remain that way for much longer. Fierce and rising geopolitical conflict over control of the global network threatens to create a balkanized system—what some technorati, including Google’s executive chairman, Eric Schmidt, have called “the splinternet.” “I’m the most optimistic person I know on almost every topic,” the Internet entrepreneur Marc Andreessen recently said in a public interview, and “I’m incredibly concerned.” Andreessen said it is an “open question” whether the Internet five years from now “will still work the way that it does today.”

If the long history of international commerce tells us anything, it is this: free trade is neither a natural nor an inevitable condition. Typically, trade has flourished when a single, dominant country has provided the security and will to sustain it. In the absence of a strong liberal ethos, promoted and enforced by a global leader, states seem drawn, as if by some spell, toward a variety of machinations (tariffs, quotas, arcane product requirements) that provide immediate advantages to a few domestic companies or industries—and that lead to collective immiseration over time.

The U.S. has played a special role in the development of the Internet. The Department of Defense fostered ARPANET, the precursor to the Internet. As the network evolved, American companies were quick to exploit its growth, gaining a first-mover advantage that has in many cases grown into global dominance. A vast proportion of the world’s Web traffic passes through American servers.

Laura DeNardis, a scholar of Internet governance at American University, argues that the Internet’s character is inherently commercial and private today. “The Internet is a collection of independent systems,” she writes, “operated by mostly private companies,” including large telecommunications providers like AT&T and giant content companies such as Google and Facebook. All of these players make the Internet function through private economic agreements governing the transmission of data among their respective networks. While the U.S. government plays a role—the world’s central repository for domain names, for instance, is a private nonprofit organization created at the United States’ urging in 1998, and operating under a contract administered by the Department of Commerce—it has applied a light touch. And why wouldn’t it? The Web’s growth has been broadly congenial to American interests, and a large boon to the American economy.

That brings us to Edward Snowden and the U.S. National Security Agency. Snowden’s disclosures of the NSA’s surveillance of international Web traffic have provoked worldwide outrage and a growing counterreaction. Brazil and the European Union recently announced plans to lay a $185 million undersea fiber-optic communications cable between them to thwart U.S. surveillance. In February, German Chancellor Angela Merkel called for the European Union to create its own regional Internet, walled off from the United States. “We’ll talk to France about how we can maintain a high level of data protection,” Merkel said. “Above all, we’ll talk about European providers that offer security for our citizens, so that one shouldn’t have to send e-mails and other information across the Atlantic.”

Merkel’s exploration of a closed, pan-European cloud-computing network is simply the latest example of what the analyst Daniel Castro of the Information Technology and Innovation Foundation calls “data nationalism,” a phenomenon gathering momentum whereby countries require that certain types of information be stored on servers within a state’s physical borders. The nations that have already implemented a patchwork of data-localization requirements range from Australia, France, South Korea, and India to Indonesia, Kazakhstan, Malaysia, and Vietnam, according to Anupam Chander and Uyen P. Le, two legal scholars at the University of California at Davis. “Anxieties over surveillance … are justifying governmental measures that break apart the World Wide Web,” they wrote in a recent white paper. As a result, “the era of a global Internet may be passing.”

Security concerns have catalyzed data-nationalization efforts, yet Castro, Chander, and Le all question the benefits, arguing that the security of data depends not on their location but on the sophistication of the defenses built around them. Another motive appears to be in play: the Web’s fragmentation would enable local Internet businesses in France or Malaysia to carve out roles for themselves, at the expense of globally dominant companies, based disproportionately in the United States. Castro estimates that the U.S. cloud-computing industry alone could lose $22 billion to $35 billion in revenue by 2016.

The Snowden affair has brought to a boil geopolitical tensions that were already simmering. Autocracies, of course, have long regulated the flow of Internet data, with China being the most famous example. But today such states are being joined by countries across Asia, the Middle East, and Europe in calling for dramatic changes in the way the Web operates, even beyond the question of where data are stored.

At a December 2012 United Nations treaty conference in Dubai, the first in history to consider the global control and regulation of the Internet, 89 countries approved a new telecommunications accord that contains a resolution calling for the International Telecommunication Union (ITU), a specialized UN agency, to play an enlarged, as yet unspecified role in Internet governance. The United States and 54 other nations, including the countries of the European Union, refused to sign on (I was a member of the American delegation to the conference). But in April, post-Snowden, Neelie Kroes of the European Commission stated: “The Internet is now a global resource demanding global governance.” This fall, the ITU will meet to define its mandate, including its potential authority to recommend Internet regulations.

Laura DeNardis, of American University, speculates that if the Internet becomes subject to direct governmental regulation, a cascade of destabilizing consequences could follow. In a complex gambit arranged by a consortium of European telecommunications providers, a bloc of African states tried in 2012 to use the ITU’s machinery to enact a new payment model dubbed “sending party pays.” In essence, the regulation would have required any content providers that transmit data between countries to pay an additional fee for the use of the service network in the destination country—a tax on international data transmission little different from a tariff wall on foreign goods. “If adopted, the proposal would have completely undermined the economic model of the Internet,” writes Vint Cerf, a senior executive at Google, in a paper he co-authored with two colleagues. The measure was never brought to a vote. But as the idea of a more heavily regulated Internet gains legitimacy, and as national governments and regional or international bodies begin to govern the Web, efforts to bend regulation to the advantage of business interests are sure to multiply.

Fragmentation could also reduce the reliability and security of the Web for ordinary users. Two years ago, a coalition of states including Russia, China, Egypt, Saudi Arabia, and Sudan sought to remove the domain-naming function from U.S. legal oversight, and instead place it under an authority like the ITU. In March, the Department of Commerce announced vague plans to relinquish control, but only under the condition that no government-led or intergovernmental organization take its place. If no compromise is reached, blocs of countries could theoretically go their own way, giving rise to competing or duplicative domain-name systems. If that happens, Internet users would be at high risk of being routed to wrong addresses, including fraudulent sites that perfectly mirror legitimate ones.

Some experts anticipate a future with a Brazilian Internet, a European Internet, an Iranian Internet, an Egyptian Internet—all with different content regulations and trade rules, and perhaps with contrasting standards and operational protocols. Eli Noam, a professor of economics and finance at Columbia Business School, believes that such a progressive fracturing of the global Internet is inevitable. “We must get used to the idea that the standardised internet is the past but not the future,” he wrote last fall. “And that the future is a federated internet, not a uniform one.” Noam thinks that can be managed, in part through the development of new intermediary technologies that would essentially allow the different Internets to talk to each other, and allow users to navigate the different legal and regulatory environments.

Perhaps. But at a minimum, this patchwork solution would be disruptive to American companies like Google, Facebook, Amazon, and eBay, which would see their global reach diminished. And it would make international communications and commerce more costly. The U.S. government is resisting this transformation. But the Internet is simply too consequential—socially, politically, and economically—for states to readily forgo control of it, and America, as Marc Andreessen observes, has lost “the moral high ground” in the debate. Perhaps it was never realistic to expect the World Wide Web to last.