On March 14, the U.S. government announced that it would seek to relinquish a privileged role in the management of Internet names and numbers. An organization called ICANN—the non-profit Internet Corporation for Assigned Names and Numbers—is to continue doing what it’s doing without maintaining an ongoing contract with the Department of Commerce to do it. And what does ICANN do? It helps keep IP addresses in order, ensuring that each address—used to let parties on the Internet identify one another—is not assigned more than once. And it facilitates the addition of “top level domains,” those suffixes like .com, .org, .uk, and more recently, .clothing, which, with a concatenation of names to their left, become the names for nearly all online destinations, including newrepublic.com. A receding role for the U.S. government has been anticipated for over a decade, and the move is both wise and of little impact. Some reaction has been surprisingly alarmist.
A Wall Street Journal columnist described it as “America’s Internet surrender.” Said one member of Congress: “Giving up control of ICANN will allow countries like China and Russia, that don’t place the same value in freedom of speech, to better define how the internet looks and operates.”
From a former Bush administration official in the Daily Caller: “This is the Obama equivalent of Carter’s decision to give away the Panama Canal—only with possibly much worse consequences.” (Namely, to “endanger the security of both the Internet and the U.S.—and open the door to a global tax on Web use.”) And Newt Gingrich: “Every American should worry about Obama giving up control of the internet to an undefined group. This is very, very dangerous.”
The venerable information technology publication The Register summed it up this way: “US govt: You, ICANN. YOU can run the internet. We quit.” And from the National Journal: “When U.S. Steps Back, Will Russia and China Control the Internet?” As Betteridge’s Law of Headlines suggests, the answer is no. Indeed, the truth is much less salacious—and far more interesting—than any of the reactions above.
To understand why, we need to talk about the difference between owned and unowned technologies. Owned technologies are easy to grasp, because they’re so prevalent. They’re technologies that are developed and shaped by a defined group, usually someone selling it. The original AT&T phone system was an owned technology—and so are its descendants like Verizon landline services and mobile phone networks. TV broadcasting is owned, in the sense that governments around the world have asserted power over the airwaves that permeate their territories, deciding who can use what bandwidth and why—and those with licenses then, with exceptions determined by regulators, decide what to broadcast. If you’re reading this article on a digital device, chances are good that its hardware is owned—Apple or HP or Lenovo designed and built it—and so, too, is its operating system, whether iOS, Android, or Windows.
If any of these technologies were to break, we’d turn to their vendors for an explanation and a fix. If a government wanted to affect how they work, it would seek to pressure or outright require certain changes—the way that, for example, the U.S. Communications Assistance to Law Enforcement Act requires AT&T and Verizon to design their telephone networks to be responsive to lawful wiretap orders.
But owned isn’t the whole story. Occasionally unowned configurations emerge. In 1983, we might have assumed that walled gardens such as CompuServe and America Online would keep being the way we communicate with one another—classic, owned information services for which we paid for access by the minute. But something odd happened: An experimental network, subsidized by the U.S. National Science Foundation, shaped by researchers at universities and corporate think tanks, came about. This Internet was meant to provide compatibility among any number of smaller networks, and unlike CompuServe and its siblings, it had no CEO, business plan, or budget. CompuServe and AOL become mere ways to access the Internet, rather than their original incarnations as globe-spanning one-stop information shops whose subscribers could be reached only by special arrangement with them.
Access CompuServe and you’d be asked for an ID and password to prove you’re a paying subscriber, and then you’d be shown a main menu of content and activities selected by CompuServe to appeal to you. By comparison, when you access the Internet, there’s nothing: no main menu, no meter from the overarching network, and no persistent identity upon it. (That’s why various Web sites fight one another to be the home page of your browser.) The Internet was less a particular set of hardware and more a set of protocols. You are simply assigned a number—which is not really meant to identify you personally—that lets you reach out to any other number on the network with bits.
But it turns out that for the Internet to work, certain functions benefit greatly from a little centralized record keeping. To surf the Net with your unique number, it helps greatly if that number isn’t already assigned to someone else—assign it twice and bits can get confused as they wend their way towards you (or is it towards your doppelganger?). Same problem with phone numbers: they shouldn’t belong to more than one pizza shop at a time. So someone had to maintain a master list of IP addresses for the simple purpose of not handing them out twice. That someone was Jon Postel, a computer scientist who in essence drew the short straw to have to keep track. As the Internet’s protocols were written up it seemed a little informal to say with a technical document, “Well, a guy named Jon performs this function,” so Jon was labeled to be something much more official-sounding: The Internet Assigned Numbers Authority, or IANA. No official paperwork was filed; nothing was incorporated. Still no CEO. Just IANA. Jon also helped get domain names going, so we could visit mnemonic-sounding addresses like www.newrepublic.com instead of 126.96.36.199. (Both work, though – try it!)
Jon asked various colleagues to manage lists like all those names ending in .com, and others ending in .org, etc. He maintained a list, called the root, of those names. It was Jon who agreed to create .uk for those interested in United Kingdom-themed domain names. When he realized that domain names were taking on real meaning to people, he looked for other ways to create names rather than just deciding on his own. (For countries, he found a list of country names maintained by the International Standards Organization and stuck to it—creating names for lots of peoples whose governments hadn’t formally asked for them.) And when disputes came up, he looked for consensus to settle them, such as when there was objection over the person originally entrusted to maintain names under .pn, for Pitcairn Island, population 50. (The objection was lodged by the entire adult population of Pitcairn Island, with the exception of the trustee and his wife.) It took years to settle the issue.
By 1997 it was clear that having Jon simply run numbering and names, however fairly, was tricky. Entrepreneurs stood to gain millions should they be entrusted to register names in a new domain like .web or .chef. How to decide how many more to create and who would get them? The U.S. government, in the form of the Department of Commerce, began a process to create a “new IANA,” one supported by the Internet as a whole. This was seen as consistent with the idea of privatizing the Internet. But note that this process skated over the U.S. government’s authority to choose a new shepherd to begin with. Why was a new leader something for the U.S. government to designate? It wasn’t as if there were local airwaves over which the government claimed power. The only real hook was that American funding had subsidized the creation of Internet protocols—but these were grants, not fee-for-service; subsidies, not permission. And the small fees the government awarded to companies that maintained the domain name databases—after Jon found it boring to maintain on his own—were ones that the companies would be delighted to simply waive, charging Internet users directly instead.
In reality, the U.S. government got to choose for two reasons, neither of which had to do with any legal authority. First, the government made sure that there wasn’t really a choice for anyone to make: the ultimate ballot it was to cast had only one entry on it. When three entities stepped forward to be the new IANA, the Department of Commerce persuaded them to negotiate with one another until only one proposal was left. Choosing something from a list of one is not controversial. Second, nearly everyone concerned about the future of the Internet wanted certainty and stability. So the U.S. government’s “decision” to recognize ICANN as the new IANA in 1998 was welcomed as a rallying cry to get a move on with the allocation of numbers and names. ICANN, a California non-profit, had a Byzantine set of by-laws to make Madison proud: Board members were appointed from various constituencies (“stakeholders” in governance parlance) and from various regions of the world. But it is not run by the governments of the world. Many had seen the way the UN operated and had little interest in replicating it. And in 1998, the U.S. government’s recognition of ICANN took the form of a cooperative agreement between ICANN and the Commerce Department spelling out certain minimal responsibilities–and a nominal way for the U.S. government to pull the plug if something went terribly wrong.
With this background in mind, we can process the news that the U.S. government is letting that agreement ultimately lapse.
First, the U.S. government control so far has had minimal impact on how ICANN has operated. For example, there was some consternation within the U.S. Congress about the creation of a .xxx domain, which was within ICANN’s purview to create. This likely delayed .xxx, but it didn’t stop it. And that accords with the government’s role in ICANN’s creation: Had it tried to be more heavy-handed, it’s not clear that it could have pulled off the move to a new IANA. Whoever newly contracts with ICANN for these IANA functions—yes, once again the U.S. government has vaguely called for a new organization to step up—will be similarly constrained. So there’s no obvious place for Russia or China to take control.
Second, the plausible ways in which ICANN could trample free speech are narrow. ICANN does not itself hand out domain names—it only designates who runs each list of names. ICANN does not directly “shut down” names or otherwise deal in individual decisions. So far it has established procedures in some domains for trademark-like disputes to play out. But these are only over domain names themselves—not over claims of behavior taking place more generally on a Web site. Register gap.clothing and be prepared to justify your action through an ICANN-approved process; sell fake Gap clothing on your website goodclothes.clothing and that process won’t have anything to say about it. Any attempt to impose broad-based censorship through domain name assignments would be met with stiff resistance by the operators of domain name registries, and ultimately by the Internet Service Providers who choose to consult those registries for information about what destination each name represents. Anyone trying to tighten the screws too much will simply strip them.
Taxes? ICANN takes a cut of fees generated by registering and renewing names for many domains like .com and many new ones in the process of being unveiled—so much so that ICANN enjoys tens of millions of dollars in income each year—but it doesn’t and can’t otherwise impose a “tax on Internet use.”
Thus last week’s news is simply about symbolism. Having the U.S. nominally, but not really, controlling the modest functions of top-level numbering and name assignments provided ammunition to those who think the Internet should be utterly stateless—some of whom, oddly enough, might favor turning over ICANN’s functions to the International Telecommunications Union, which is an arm of the United Nations and has states as its members. To eliminate this symbolic U.S. involvement, an action envisioned from the moment of ICANN’s creation, helps address that complaint, while costing nothing. As ICANN’s own Q-and-A on the topic put it:
How does this announcement affect the individual Internet user?
This announcement does not affect Internet users and their use of the Internet. However, all Internet users have a stake in how the Internet is run, and it is therefore important to get involved.
Confusing: Nothing to see here, but Internet governance matters, so go on and get involved. Such are the puzzles of unowned technologies. They can become incalculably powerful even with no one at the helm—or perhaps precisely because of it. Numbering and naming is a tiny part of the Internet, and governing it is of interest mostly because it’s one of the few things we can point to where decisions can be made. But these decisions happen by consensus, and are implemented one ISP and router at a time, rather than some kind of fiat. You may be reading this article at newrepublic.com, and if you are, you’re here because your ISP, your operating system vendor, your browser maker and you are agreeing to map that name to this online place. Any could change it, notwithstanding actions of governments and institutions like ICANN. Internet protocols at large aren’t implemented through anyone’s fiat; they are generated through open processes channeled through unincorporated organizations like the Internet Engineering Task Force (motto: “We reject kings, presidents, and voting: we believe in rough consensus and running code”), and then implemented through the actions of hardware and software makers.
The Internet is a collective hallucination, one of the best humanity has ever generated. To be sure, it is delicate in many ways, with its unowned character threatened from many quarters. But rest easy that ICANN isn’t one of them.
Jonathan Zittrain is a professor of law and professor of computer science at Harvard University, and author of The Future of the Internet – And How to Stop It. He is a former trustee of the Internet Society, which facilitates the work of the Internet Engineering Task Force. He served on a membership advisory committee to ICANN in 1998, testified to Congress about it in 2000, and was among the authors of the Berkman Center’s 2010 report done at the request of ICANN, “Accountability and Transparency at ICANN: An Independent Review.”
Image via Shutterstock.