CONTENTS ABSTRACT INTRODUCTION I. THE INTERNET IS BORN A. The Early Days B. Continuing to Build By Design C. Standards and Protocols D. The Result II. THE INTERNET IS ALIVE A. I Come from Cyberspace, New Home of the Mind B. I[nternet] Fought the Law and the Law Won C. But Some Things Are Still As They Were III. "Good GRIEF, MAN, How COULD You PUT A HOLE IN OUR INTERNET?" A. Breaking the Domain Name System: Blacklists and Redirects B. ISP in the Middle C. Rent Seeking in the Network Transport Layer D. Putting the Pieces Together: A Puzzle or a Maze IV. ONE WAY FORWARD: AN INTERNATIONAL RESPONSE TO AN INTERNATIONAL THREAT A. A Global Network Demands a Global Solution B. Existing Global Structures, Existing International Opportunities C. An Impractical but Perhaps Ideal Solution V. CONCLUSION: VIRTUAL REALITY AND GLOBAL REALITY It was on a dreary night of November, that I beheld the accomplishments of my toils. With an anxiety that almost amounted to agony, I collected the instruments of life around me, that I might infuse a spark of life into this lifeless thing that lay at my feet.
Mary Shelly (1)
Some things evolve. (2) Once created, they morph, slowly changing, waiting for that push in one direction or another. Systems, as well as ideas, can evolve. (3) Moving from person to person, or through time, they change slightly or significantly, and in the end are different. But not everything and not everyone evolves, and not at all levels. Frankenstein, for example, did not evolve into existence. Once in existence, he did evolve. But his initial existence was not a mutation that followed from millennia of incremental change in human development. He was created. He was put together. He was designed. And once he was planned and executed, he was constrained by that initial planning. He could do certain things, such as walk, and he could not do other things, such as fly. He could evolve in many ways, but he could not go beyond what was enabled and disabled in the process of his creation, and that creation followed a particular plan or design.
In a similar way, the underlying and most critical technologies of the Internet were not the result of an evolutionary process following from the mutation of prior technologies. (4) In addition, once in existence, these technologies did not mutate further, with competing innovations providing disparate options from which users or even Internet applications could choose. The development of the Internet's transport level did not involve a Sony Betamax versus VHS style war of possibilities played out in the market. (5)
Instead, the base--or transport--layer of the Internet was planned. After its initial, designed creation, the transport level's development continued to change primarily by design. This is not true at all levels of the Internet; it is certainly not true at the higher "application" levels, where innovation has flourished and users are enabled to morph functioning and outcomes to meet their needs and desires. (6) But at the deeper levels, the primarily standards-based levels on which the rest of the Interact relies, it is true: the Internet exists because of a conscious choice to bring it into existence, and a choice in doing so to give it a particular shape and structure.
Ever since it was developed, however, the Internet has been under stress. With multiple "moving" and critical parts, private parties, corporations and governments have all taken aim at its various layers of functioning in an attempt to gain or protect rights, to scratch back perceived losses of control. Some of these attempts have been aimed at the very parts of the Internet with which we are concerned here: the various hardware- and software-based elements that are, at their core, concerned with the efficient transport of data across the network. While this layer of the Internet was planned, and additions and changes carefully checked against the functioning of the system as a whole, these pressures are individual, isolated, and focused in nature. And in seeking solutions to problems, there is little incentive to consider the overall effects of a particular choice in relation to the Internet as a whole. In other words, when implementing specific solutions, there is little consideration given to whether a specific solution might "break" the Internet. And while one change may not do so, many disparate changes attacking a variety of perceived problems, and implemented from diverse areas, very well may. And therein lies the rub.
My thesis here attacks this problem, and argues that the Internet is and has been, at its core, a planned system. Before we (7) alter the fundamental technological tenants of the Internet's transport layer through national and international actions (both public and private) we should agree to do so, that is, we should plan to do so. We should not simply let individual state and market interests take steps themselves that alter the fundamental technological nature of the Internet. The most legitimate way to alter an international network is by international action, and an international treaty--its terms drawn primarily from technologists and civil society--is the solution proposed here.
I begin by next considering the technology "behind" or "underneath" the Internet. (8) This may seem a natural starting point, but it is important to note initially that the technology is not itself important, nor am I particularly concerned with what the technology is. Rather, I am concerned with how the technology came to be the technology. The process of its development, implementation, and ongoing augmentation is the concern here, and this is the focus of section I of this paper. Section II follows this discussion of technological development with one of social and legal development as applied to the Internet. Section III then takes up the threats to the design of the Internet at the core transport level, outlining how diverse requirements aimed at the Internet's transport layer each "break" the Internet slightly, and how together they have the potential to significantly alter how the planned elements of the Internet function as a complex system of hardware, software, and standards-based protocols. Section IV then investigates one way out of this morass, proposing that existing international engineering and civil society groups drive forward a process intended to bring about an international convention addressing regulation of the Internet's transport-level elements.
THE INTERNET IS BORN
Many myths exist about the beginnings of the Internet; Al Gore creating it comes immediately to mind, (9) for example. Others include the development of the Internet as a response to the potential for a nuclear attack to wipe out communications infrastructure in the United States (a story disputed by those who actually designed and implemented the framework on which the Internet was initially built (10)). These apparent complications aside, the Internet's creation story is more straightforward. (11)
The Early Days
In 1969, the Defense Advanced Research Projects Agency (DARPA), issued a call for proposals to develop a new kind of network, one that would allow scientific researchers to share data and computer resources at the same time as it built on new digital technologies to allow for greater efficiency and robustness in network design (the network was to become known as the ARPANET). (12) Prior to this time, communication occurred primarily through circuit-switched networks, in which the ends being connected were just that: connected. (13) With dumb ends consisting of telephones and smart middles made up of the many and varied circuits necessary to connect phone to phone, circuit-switched networks were prone to failures that had the potential to cause large-scale disruption where one of the switches failed. (14) In such a case, no one who relied on that particular switch could communicate over the network. (15)
ARPANET was to be different. It was to be a packet-switched network, one in which messages or content were broken into pieces, sent independently across the network, and put back together at their destination. (16) One piece of the message or content might not follow the same path across the network that other pieces had followed, but in the end the users--senders and receivers--would neither be aware that the message had been sent this way, nor would they need that knowledge. Instead, the underlying technologies would be invisible to the end users, who would only be aware that their message had been sent and received. (17)
Critical choices were made as the ARPANET project began. They were focused on a number of priorities, but critical among them was the notion that the network itself should be largely unconcerned with what was passing through it. It was to be content-agnostic. Unlike the telephone system, which had been put in place to carry voice communications from one human to another--and thus, for example, allowed for a fairly high level of noise on the lines themselves because human beings can often discern the message through the noise--the new network was called on to efficiently pass on all kinds of traffic as quickly as it could. The design did not favor voice, or pictures, or video, (18) but rather accepted all content as equal, and passed it all along.
To bring the design into existence, a multilayer architecture was developed. (19) There are a number of ways to conceive of the layers that are necessary. One conception involves seven layers, as follows (from bottom to top): the physical layer, the datalink layer, the network layer, the transport layer, the session layer, the presentation layer, and the application layer. (20) Each layer sits "on top of" the layers below it, and utilizes the lower layers to perform its necessary function, with the ultimate goal being that information at the application layer is usable by the receiver in the manner intended...