by David Post
Webposted on 06 June 1999
In a column last Fall I suggested that the pending reorganization of the Internet's domain name system (DNS) had the potential to become cyberspace's own "constitutional moment," a profound and thorough transformation of the institutions and processes responsible for law-making and regulation on the global electronic network. [See Cyberspace's Constitutional Moment] Over the last several months, the shadowy outlines of a new kind of constitutional structure for cyberspace have indeed begun to emerge. The consequences of these developments for the Internet's future could not be more profound, and the picture that is emerging is not always a pretty one. Not many people are paying close attention to these developments; they should. Not to be an alarmist, but to boil a frog alive, the parable tells us, you need just to turn up the heat by increments so small that the frog never notices -- never pays attention to -- the rising temperature until it is too late to do anything about it (like jump out of the pot). Well, the temperature on the Internet is starting to rise.
(A bit of) Background
[Those of you who already know how the DNS works, and what a "root server" is, might want to skip ahead . . . ]
For the Internet to exist as a single coherent network, there must be a way to be sure that a message sent from any point on the network to email@example.com, or a request to view the webpage at www.school.edu, is routed to the "right" machine; that is, each of those addresses (xyz.com or www.school.edu) must be associated unambiguously with a particular machine if message traffic is to move in a predictable way.
To make an extremely long story short, the global network we call "the Internet" manages this by, first, requiring that each machine on the network have a unique numerical address [e.g., 220.127.116.11]; indeed, to be "on the Internet" means to have (or to be connected to) a machine to which such an address has been assigned. For your message to www.school.edu to be routed correctly, your computer must somehow be able to find the numerical address corresponding to the machine named www.school.edu. This is accomplished, on the Internet, by means of what Tony Rutkowski calls a "magical mystery tour."
When you send off a message requesting a copy of the www.school.edu home page ("http://www.school.edu"), the message first stops off at a machine known as a "DNS [Domain Name System] server." It is the job of the DNS server, which is usually operated by your Internet Service Provider, to find the correct numerical address for your message. The DNS server reads, in effect, from right to left; seeing that this is a message destined for some machine in the EDU domain, it needs to find out where addresses in the EDU domain are stored. It does this by asking a different machine (known as the "root server") that very question: "Who is responsible for the EDU domain?" The root server replies with the numerical address of a different machine (known as the "EDU domain server"). Your DNS server then asks the EDU domain server the question: "Who is responsible for "school.edu?" The EDU domain server replies with another numerical address [or with "Not Found" if it cannot find an entry for "school.edu" in its database of names and addresses]. Your DNS server then asks this machine: "Who is responsible for www.school.edu?" School.edu replies with yet another numerical address, and now your DNS server has completed its task; once it receives the address for www.school.edu, it places that address into your message and sends it on its way.
How does this all work as smoothly as it does? Who is in charge of the root server? How does the operator of the root server decide which machines are the "authorized" domain servers for EDU, or COM, or ORG, or any of the other top-level domains? Who controls those machines (and the database of names and addresses contained in them)? And how is this whole scheme enforced? That is, what makes the root server "the" root server? Why do the many thousands of Internet Service Providers, operating the many thousands of DNS servers worldwide, all use the same root server?
In the early days of the Internet, of course -- through, say, the early 90s -- no one outside the small cadre of engineers that was putting the system together cared very much about the answers to these questions. There were, of course, answers to them all. The United States government had long operated the root server (a holdover from the days that the Internet was a Defense Department project), and had worked with something known as the Internet Assigned Numbering Authority (IANA) -- a group of engineers led by the late Jon Postel -- to organize the necessary data and to see that the various domain servers were being properly managed.
As long as it all seemed to be working smoothly enough; who cared what was going on behind the Wizard's curtain? And who noticed when, in 1992, as the extraordinary growth of the network began to outstrip the management capacity of this (largely volunteer) operation, the U.S. government engaged a private firm, Network Solutions, Inc. (NSI), to manage and maintain the databases and domain servers for the COM, ORG, and NET domains?
But slowly, as more and more people began to realize that the Internet was a Really Big Deal (and that these funny "domain name" things might actually be of real value), more and more people started to pay attention to all of this, and this arrangement began to come under increasing fire from many quarters. The government and NSI found themselves increasingly under attack from within and without the Net community by those challenging NSI's apparent monopoly control over these increasingly valuable top-level domains, by trademark owners concerned about domain names that appeared to infringe upon their valuable trademark rights, and others.
As the expiration date of the government's contract with NSI approached last June, the Commerce Department announced that the government wanted to get out of the DNS management business entirely. Citing "widespread dissatisfaction about the absence of competition in domain name registration," and the need for a "more formal and robust management structure," the government proposed transferring responsibility for management and operation of the DNS to a private non-profit corporation. This new corporation, to be formed by "Internet stakeholders" on a global basis, would take over responsibility for overseeing the operation of the authoritative Internet root server, and would be charged with introducing competitive, market mechanisms into the allocation of Internet names and addresses.
Um, What Does This Have To Do With "Internet Governance"?
This new corporation -- ICANN, the Internet Corporation for Assigned Names and Numbers -- has, over the past several months, set up shop and gotten to work. It's been a busy time. It has begun to establish "Supporting Organizations," new coalitions comprising various Internet constituencies (e.,g., domain name registrars, trademark owners) who will be responsible for electing members of the ICANN Board of Directors and for formulating aspects of domain name policy. It has taken the first steps towards introducing competition into the domain name system, accrediting five companies (America Online, CORE (Internet Council of Registrars), France Telecom/Oléane, Melbourne IT, and register.com) to begin issuing registrations in the COM, ORG, and NET, domains during a two-month test period (along with twenty-nine other entities who can begin accepting registrations in these domains once the test phase is completed). It commissioned, and recently adopted (in part), a report from the World Intellectual Property Organization (WIPO) outlining the procedures to be used in cases involving "cybersquatting" (the intentional "warehousing" of domain names for later sale).
My goal here is not to discuss any of these specific actions; there is much here to digest and debate, pro and con, and I will have a great deal more to say about the specifics of ICANN's activities over the next several months. [As, I hope, will others; I particularly recommend Michael Froomkin's commentary on the WIPO Report referenced in the preceding paragraph, posted at at his website.]
Rather, my goal here is just to suggest that notwithstanding the government's (and ICANN's) protestations to the contrary, this is about nothing less than Internet governance writ large. The Commerce Department took pains to characterize it in other terms; this new corporation would be responsible only for "technical management of the DNS," the "narrow issues of management and administration of Internet names and numbers on an ongoing basis" -- sort of what the International Telecommunications Union does with respect to managing interconnections on the international telephone network. This new framework for managing the DNS ". . . does not set out a system of Internet "governance." Existing human rights and free speech protections will not be disturbed and, therefore, need not be specifically included in the core principles for DNS management. In addition, this policy is not intended to displace other legal regimes (international law, competition law, tax law and principles of international taxation, intellectual property law, etc.) that may already apply. The continued applicability of these systems as well as the principle of representation should ensure that DNS management proceeds in the interest of the Internet community as a whole."
It is all well and good to say that this new institution will not be engaged in Internet governance -- but words will not make it so. Any entity exercising control over the DNS will be subject to immense pressure to do more than mere "technical management," because, bizarre as it may seem at first glance, the root server, and the various domain servers to which it points, constitute the very heart of the Internet, the Archimedean point on which this vast global network balances.
To appreciate that, imagine for the moment that you had control over operation of the root server. You alone get to decide which machines are "authoritative" domain servers for the COM, NET, ORG, EDU, and the other top-level domains, the machines to which all Internet users worldwide will be directed when they try to send any message to any address in those domains. You have the power, therefore, to determine who gets an address in those domains -- who gets a passport without which passage across the border into cyberspace is impossible. You can say "From now on, we will use the data in machine X as the authoritative list of COM names and addresses, but only so long as the operator of that machine complies with the following conditions," and then you can list -- well, just about anything you'd like, I suppose. It's your root server, after all. You can require that all domain server operators pay you a certain fee, or provide you with particular kinds of information about the people to whom they have handed out specific names and addresses, or only allow transmission of files is a specified format, or abide by a particular set of laws or rules or regulations. And you can demand that they "flow through" these conditions (or others) to anyone whom they list in their authoritative databases, that they revoke any name given to anyone who does not pay the required fee, or provide the required information, or use the specified file format, or comply with the specified rules and regulations.
This is quite literally a kind of life-or-death power over the global network itself, because presence in (or absence from) this chain of interlocking servers and databases is a matter of [network] life or death: If your name and address cannot be found on the "authoritative" server, you simply do not exist -- at least, not on the Internet. Eliminate the entry for xyz.com from the COM domain server and xyz.com vanishes entirely from cyberspace; designate as the new COM domain server a machine that does not have an entry for xyz.com in its database, and you have imposed the electronic equivalent of the death penalty on xyz.com.
Anyone interested in controlling the rules under which activities on the Internet take place -- and many commercial interests, who now realize the huge economic stake they have in this medium, and many governments, who have spent the last few years worrying about how they would ever get back their taxing and regulatory authority over Internet transactions, find that they are indeed quite interested in that now -- is likely to find the existing of a single controlling point awfully tempting. Anyone with a vision of how the Internet can be made a "better" place " by making it safer for the exploitation of intellectual property rights, say, or by eliminating the capability to engage in anonymous transactions, or by making it more difficult for children to get access to indecent material -- is likely to view control over the root server as the means to impose its particular vision on Internet users worldwide. After all the talk over the past few years about how difficult it will be to regulate conduct on the Internet, the domain name system looks like the Holy Grail, the one place where enforceable Internet policy can be promulgated without any of the messy enforcement and jurisdictional problems that bedevil ordinary law-making exercises on the Net.
And that is why these are governance questions, why any reorganization of this system, far from being an arcane technical detail of Internet engineering, is inherently of constitutional significance. Power corrupts, absolute power corrupts absolutely -- on the Internet as elsewhere. Questions about constraining any form of absolute power are constitutional questions of the highest order, and "governance" means nothing more (and nothing less) than the search for mechanisms to insure that absolute power is not exercised in an unjust or oppressive manner. How can we be assured that ICANN will be able to resist pressures to stray beyond this limited "technical" mandate? Where are the checks on the new corporation's exercise of its powers?
You think, perhaps, that I exaggerate the significance of these developments, and perhaps I do. But let me point to a few dark clouds on the horizon that make me very, very nervous about what ICANN is up to. Remember all those things you could do if you were in control of the root? Like " require that domain server operators pay you a certain fee"? Well, ICANN has imposed the requirement that each accredited registrar pay ICANN a fee of $1 for each new domain name they hand out -- can anyone say "taxation without representation"? Or "provide you with particular kinds of information about the people to whom they have handed out specific names and addresses"? ICANN, having now adopted the WIPO Report referenced earlier, is about to impose a requirement on all domain name registrars that they collect and make available "accurate and reliable contact details of domain name holders," and that they agree to "cancel the domain name registrations" wherever those contact details are shown to be "inaccurate and unreliable." -- a move with grave consequences for the continued viability of anonymous communications on the Internet. Or "abide by a particular set of laws or rules or regulations"? The WIPO Report, again, envisions that all claims by trademark holders that a domain name registrant registered an infringing name "in bad faith" be submitted to a single, uniform, worldwide dispute resolution process for adjudication.
Now, some, or even all, of these may be good ideas. But this is already way beyond the realm of technical "standards-setting," and we really must ask whether we really want or need this kind of global Internet policy and whether this is the way it should be put together. When did the affected constituency -- all Internet users worldwide -- decide that they want a global policy-making organ of this kind? Who decided that the bottom-up, decentralized, consensus-based governance structures under which the Internet grew and flourished are incompatible with its continued growth and development? When are we going to get a chance to ratify these new arrangements?
There are hard questions here, but one thing is clear; we need to disabuse ourselves of the notion that this is somehow not about Internet governance if we are going to make any serious headway on them. We know something about how institutions that possess life and death power can be constrained, about constitutions and constitutionalism, about the fragmentation of power and the need for checks on the exercise of power, and we better start thinking about this problem in these terms before too much more time elapses.
Oh and about James Madison? Madison not only thought more clearly and more insightfully about these questions than anyone before or since, he understood the necessity for public discussion and debate about issues of this kind; the Federalist Papers, in which he and Alexander Hamilton (and a somewhat recalcitrant John Jay) laid out the arguments for (and against) the constitutional structure put together in Philadelphia in 1787, began life, let us not forget, as newspaper columns appearing weekly in the New York press. We could do worse than to start thinking about updating that for the new cyber world we are building now.