by François Bar, Walter Baer, Shahram Ghandeharizadeh,
and Fernando Ordonez

More than just a new medium, the Internet is fast becoming our primary communication infrastructure, progressively supplanting older radio, telephone, and cable television networks. In the Net’s early days, during the 1970s and 1980s, it mainly provided support for text exchanges through e-mail, broadcast listservs, and text-based precursors to the Web such as Gopher servers. The widespread traffic of images came next, with easier media attachments to e-mail and the introduction of Mosaic, the first popular graphical Web browser, in 1993. Today, with broadband access widely available, the Internet is commonly used to transmit audio, including a growing share of our phone conversations through VoIP, streamed radio programs, podcasts, and recorded music. Video, too, is being broadcast over the Net, from traditional television programs and films to video blogs, home movies, and creations made by a multitude of emerging amateur producers. Thus, a single communication infrastructure has progressively absorbed a multitude of media streams that once each required specialized networks.

This is not to say that old networks are simply discarded. Throughout this book, we have observed how transformations in place, culture, and politics build on existing historical conditions. The case of infrastructure is no different. Existing telephone copper wires, cable television’s coaxial lines, long-haul optical fibers, and satellite and microwave-radio links are being folded into the Internet as telecommunications companies and cable carriers convert their respective networks to IP (Internet Protocol)—the set of data transmission conventions that allow communications across the various parts of the Internet. In fact, rather than a separate physical infrastructure, the Internet is primarily a virtual network—the assemblage of a multitude of transmission and routing facilities tied together by the IP’s common software “glue.” As a result, the Internet is perhaps best understood in its original, unabbreviated sense, as an Internetwork, or an agglomeration of separate networks that agree to connect to each other and exchange traffic through gateways where they speak the IP lingua franca. As an Internetwork, the Internet differs from traditional telephone and cable networks in two fundamental respects—its decentralized governance and its E2E architecture—and yet, relying on telephone and cable networks for the last mile of connectivity, the Internet is subject to economic and political pressure from established communications companies.

Decentralized governance means that no single organization is in charge of managing the Internet, which is in stark contrast with traditional telephone or television networks. In particular, individual networks can become part of the Internet as soon as they find an existing member of the Net that agrees to connect with them through a gateway to exchange traffic. Connected to one node, they are connected to the entire Net. Historically, this allowed the Internet’s spectacular growth, as more and more network operators chose to join.

A brief glance at the history of the Internet illuminates the exponential nature of growth that this decentralized system allowed. Initially, the IP architecture was worked out in the ARPANET (Advanced Research Projects Agency Network), a military-sponsored experiment, during the early 1970s. It was notably expanded in the late 1970s and 1980s to support computer-intensive research through the government-sponsored NSFNet (National Science Foundation Network), an IP-based network linking universities. By the end of the decade, a number of corporations began to use the same networking approach to build their internal corporate networks, and a multitude of private Internet Service Providers (ISPs) started to offer dial-up Internet access over telephone companies’ lines. During the 1990s the Internet became a mass medium, propelled by easy-to-use, multimedia content available through the World Wide Web and the absorption of consumers from earlier, largely self-contained networks such as Compuserve, Prodigy, and America Online. Throughout, the Internet’s expansion did not require the blessing of centrally-controlled telephone or television networks, but instead proceeded in a decentralized fashion as increasing numbers of private and public operators adopted the new networking model and peered with existing participants to join the Internetwork.[1]

Another unique feature of the Internet is its E2E architecture, which further fueled this success because it enabled the deployment of a communication infrastructure that did not predetermine how it would be used, thus opening the Internet to wide-ranging experimentation and innovation. The E2E model calls for processing information in the devices connected to the ends of the network whenever possible, while the Internet itself remains a dumb network, simply transporting bits of information from end to end between intelligent terminals.[2] This results in a network architecture inverse from that of traditional telephone or cable networks, in which an intelligent network processes information that passes between dumb terminals. In effect, this compounds the Net’s decentralized governance, resulting in an infrastructure where the capacity for controlling information flows and inventing new communication services resides in the ends, under the control of network users rather than network providers.

Decentralized governance and E2E architecture together have thus played a key role in the spectacular success of the Internet since its emergence in the late 1960s. Indeed, these principles created an infrastructure fundamentally open to the creative contributions of a multitude of innovators, be they hardware designers, network operators, application creators, or content authors. Openness enabled flexibility, supporting wide-ranging experimentation with Internet possibilities by established companies, start-ups, and end users.[3] As a result, the Internet has been the source of many innovations, which have given rise to entirely new-media companies such as Amazon.com, Google, and eBay, as well as to communication applications such as VoIP, BitTorrent, and other P2P services.

The decentralized, E2E Internet emerged in parallel with the gradual introduction of competition within the nation’s telecommunication network during the past half century. Starting with the landmark 1956 Hush-a-Phone and 1968 Carterfone decisions, AT&T had to let others connect their equipment to its network. During the 1970s, another series of legal decisions, in particular the Execunet rulings that made MCI viable, forced AT&T to interconnect with competing long-distance providers. Finally during the 1970s and 1980s, the FCC’s (Federal Communications Commission) Computer Inquiries regulations prevented AT&T (and the post-breakup “Baby Bells”) from blocking the provision of enhanced data services on their networks.[4]

Each of these decisions was the result of intense battles in which AT&T, the dominant phone company throughout this period, wanted to decide how others could use its network. With each decision, policy makers chipped away at AT&T’s control and pushed the phone network toward greater neutrality, gradually allowing other equipment makers, service providers, and users to gain greater freedom in how they could use the network. The regulatory environment that emerged was essential to the success of ISPs such as AOL, CompuServe, Prodigy, and many others who offered Internet access over modems connected to the telephone network. They thrived in the late 1980s and early 1990s because the phone network they depended upon could not discriminate against them. As the Internet enters the broadband era, however, that long-standing battle for network control has returned in full force. Alongside AT&T, it now pits the small group of large network owners (phone and cable companies) against the multitude of network users, large and small (ranging from Google and Microsoft to end users and start-ups).

Local access in the United States is the de facto exclusive domain of the telephone and cable network operators (TCNOs). While the goal of the 1996 Telecommunications Act was to stimulate competition within the local access market, it remains effectively dominated by the corporate grandchildren of the Bell System, primarily the new AT&T (formerly SBC), Verizon, and Qwest. These carriers emerged from successive corporate reorganizations among the Regional Bell Operating Companies (RBOCs) that were created from the Bell System when it was broken up in 1984. Each continues to enjoy a largely dominant position in the area that used to be the monopoly territory of its RBOC precursors. For their part, the cable TV operators function as local monopolies as well, under franchises granted by municipal governments. Two large companies, Comcast and Time Warner Cable, control well over 50% of the U.S. cable market and, instead of competing, cooperate so as to consolidate their regional monopolies. Thus, the recent evolution of both telco and cable industries has been marked by a trend toward consolidation.

Within this context, the battle for network control was revived in large part by the substantial investments required to transition from narrowband to broadband. The corresponding network upgrades had to take place within the network itself, once again putting network owners in full control. As they tried to navigate the broadband transition, ISPs who did not own a network, like AOL or Covad, were brushed aside by network owners. Today, infrastructure owners argue that in order to invest in building broadband networks, they need greater revenues from the resulting business, which they believe they will get if they are able to discriminate among applications.

Thus, the Internet openness we have come to associate with decentralized governance and E2E architecture is again coming under strain. Content and application providers worry that this could leave them at the mercy of infrastructure owners, forcing them into business arrangements that would restrict their options. Users would certainly like greater and more reliable transmission speeds but don’t want this to come at the cost of freedom of access or future innovation. As in many complex situations, there is no clear-cut best approach to set the Internet on a course that will continue to leverage broad-based innovation and create new ways to communicate over a ubiquitous broadband infrastructure. Rather, there are trade-offs between the interests of today’s infrastructure owners, application providers, content creators, and users. This chapter explores these trade-offs and the impact they could have on the evolution of our communication infrastructure and the activities it will support.

Speed Bumps on the Road Toward Ubiquitous Broadband

Some of the tensions underlying the transition toward the next generation Internet infrastructure are universal. Network owners in all countries, whether telecommunications carriers, cellular providers, or cable systems, are moving toward the deployment of integrated broadband IP infrastructures for all communication services (voice, data, and video) to the home. In doing so, they need to secure funds to finance the upgrades to infrastructure and must set up sustainable business models for operation in light of falling costs for traditional cash generators such as long-distance tolls.

Nevertheless, the American context is unique and explains the specifics of the domestic policy debate. Three features of this context stand out: the eroding U.S. position in broadband worldwide, the enduring structure of the U.S. local access market as a duopoly, and the lack of a national broadband policy.

The local access network, what the phone companies used to call the last-mile connection between long-distance networks and customers premises, be they residences or corporate campuses, is the critical bottleneck for broadband. Indeed, today there is abundant bandwidth available in the Internet’s backbone, in part as a result of exuberant investment in cross-country optical-fiber links during the dot-com boom. Some of these optical fibers are still dark, awaiting the installation of lasers to light them up so they can carry information, but the potential is there. Similarly within customer premises, business and residential alike, high-speed networks are commonplace, either in the form of Ethernet’s ubiquitous blue Cat 5 cables or wireless Wi-Fi networks. The missing link for true broadband deployment is the connection between the two networks.

The United States led the development of Internet technology and its early deployment into a widely used infrastructure. In the current transition to universal broadband, however, other countries have taken the lead. OECD data (figure 4.1) shows the United States ranking fifteenth among OECD countries in broadband subscribers per one hundred people as of December 2006.[5]

[Figure 4.1]

Analysts attribute the broadband lag in the United States to a combination of demographic, economic, institutional, and policy factors. Broadband penetration is positively correlated with population density. The United States has lower population density than most other OECD countries, which increases the average length and cost of American broadband access lines. Dial-up Internet access may also be a better substitute for broadband in this country, where local calls are generally free, than in other countries where local calls are metered.[6]

Moreover, unlike Japan, Korea, and some European countries, the United States lacks a national policy to spur broadband uptake. Compounding this, the United States also has a telco/cable duopoly with relatively weak broadband competition from other providers.[7] Studies within and among OECD countries generally conclude that more competition results in lower prices and greater broadband penetration.[8]

 [Table 4.1]

Government policy to encourage and subsidize investment in fiber-optic networks has spurred big broadband—above 20 megabits per second (Mbps)—development in Japan and South Korea. In 1995, the South Korean government sponsored construction of a nationwide, high-capacity-fiber broadband network that could be used by any telecom carrier. The Korean government also subsidized loans to broadband providers as well as user purchases of personal computers. Moreover, government unbundling policies that require incumbents to allow competitors to use their local access lines have enabled big broadband competition from companies such as Yahoo! BB in Japan and Hanaro Telecom in South Korea.[9] This international situation provides the backdrop for the current U.S. policy discussion about the future of the Net. All sides agree that access to an advanced communication infrastructure is essential to the future social and economic well-being of the country. There is substantial disagreement, however, about how best to achieve this goal.

In the United States, each local market is at best served by a duopoly: customers who want broadband Internet access can choose between telco-provided DSL (Digital Subscriber Line) service or cable-modem service from their local cable operator. In large portions of the country, particularly in rural areas, customers only have one broadband option, or even none. Where other alternatives exist, they come mostly from companies that provide Internet service over telephone lines they lease from the incumbent telco. For all practical purposes, the two members of this TCNO duopoly control the current availability and future evolution of the local distribution network, key to the connection between high-capacity Internet backbone and customer premises. So long as that situation persists, they jointly hold the key to the future of the country’s broadband Internet infrastructure.

For big broadband to become available throughout the United States, and for the country to reverse its downward slide in the broadband rankings, there are two fundamental options. The first is for the TCNOs to make significant investments to upgrade their networks; the second is for an alternative broadband infrastructure to be deployed that can challenge the existing duopoly.

The TCNOs claim that making the investments needed to offer ubiquitous big broadband as opposed to current DSL and cable mini-broadband (5 Mbps or less downstream; much less upstream) would only make financial sense if they were allowed greater control over the applications and content delivered over these networks. In particular, the telcos point to obstacles inherited from their history as common carriers that require them to transmit all messages and data alike (e-mail data bits just like high-definition video frames, traffic from partners just like traffic from competitors). According to them, unleashing adequate investment for the broadband infrastructure requires a break from the Internet’s tradition of decentralized control and E2E architecture. (See box 1.)

The second option, the deployment of an alternative broadband infrastructure, has been the elusive goal of U.S. telecommunication policy over the past two decades. This was, in particular, the underlying goal of the Telecommunications Act of 1996, which structured a set of incentives for the telecommunications companies to gain greater regulatory freedom as they allowed competitors to enter their local markets. These efforts have largely failed, and to date no infrastructure builder has emerged that can credibly compete with the TCNOs. Over the past few years, however, a new candidate has come forward with the advent of wireless data networks. Promoted in particular by the newfound availability of unlicensed radio spectrum, technologies such as Wi-Fi and WiMAX suggest that new infrastructure models could provide the basis for an entirely new last-mile broadband infrastructure. This could come in the form of wireless networks run by for-profit firms or local government operators. More dramatically, some also foresee the advent of ad hoc networks, where wireless-enabled devices connect to one another when they come within radio range, creating a decentralized mesh without operators. We review these options in more detail later on.

The Debate Over Network Neutrality

As a result, there is considerable debate about whether the Internet’s upgrade to ubiquitous big broadband can proceed within the traditional decentralized, E2E framework. In particular, the TCNOs who will have to invest in this infrastructure upgrade claim they can only justify this next round of investment if they are able to capture a portion of the revenues associated with the new broadband applications—in particular video distribution. In order to do this, they seek to shape traffic over their network so they can deliver better performance to those (content/application suppliers or consumers) willing to pay for it. TCNOs argue that unless they can manage traffic and applications in such a manner that they can reap economic benefits, they will have no incentives to invest in network upgrades and, thus, have little reason to expedite the move toward ubiquitous and affordable broadband for all end users. Of course, network owners already manage traffic to some extent, for example, when they allocate available bandwidth among neighbors. But the kind of management they yearn for would allow them to differentiate traffic on the basis of business relationships with content providers and consumers. For example, they might make YouTube videos to flow more smoothly if YouTube gave them a cut of the related advertising revenues, or provide lower latency (faster reaction time) to World of Warcraft players in exchange for a share of the game subscription fees.

In response, a coalition of content and application providers (including Amazon.com, eBay, Google, Intel, Microsoft, and Yahoo!) and consumer protection organizations have argued that by unduly favoring TCNO-owned content, this would be the end of the Internet as we know it. They have proposed network neutrality principles to guarantee that network owners do not treat different traffic flows differently, whether on the basis of fees paid by those exchanging traffic or according to what application they use. They see network neutrality as a way to ensure overall economic and social benefits in light of increased TCNO market power. (See box 2.) Network neutrality advocates argue two fundamental points. First, information networks should preserve the E2E Internet architecture and be as neutral as possible among competing content, applications, and services. Second, if and when it is necessary, government should intervene to promote or preserve the neutrality of these networks.[10] The FCC did, in fact, adopt a broadband policy statement in August 2005 that consumers are entitled to their choice of lawful Internet content, applications, services, and devices (see box 3); but the FCC has not adopted rules to enforce these principles.

This underlying tension about how much control infrastructure owners should be allowed to wield over the communication activities carried by their infrastructure is not new.[11] As with Carterfone or the FCC’s Computer Inquiries in previous incarnations of this debate, there are two categories of issues. The first is about gatekeeping; static, immediate, and highly visible, it captures much of the attention. The second is about future evolution; dynamic, longer term, and unknown, it is much harder to measure its impact.

The gatekeeping debate revolves around the fact that TCNOs want to charge a premium to those who want to transmit their content over reliable broadband streams (broadcasters, game servers, etc.). In order to deliver high-quality broadband streams to their highest-paying customers, TCNOs claim they need the ability to shape traffic—to decide which packets have priority, and which communication flows get reserved bandwidth. With the ability to shape traffic, they also get control over content flows, ranging from small tweaks in how quickly Web pages load (for example, making it faster for content providers willing to pay more) to outright censorship by blocking certain categories of content or certain applications (for example, redirecting Internet users toward online stores with whom they have a partnership or preventing them from using the VoIP services of their competitors).

Other competitive issues can arise if the TCNOs configure their network for a specific kind of communication activity and as a result hinder other kinds of traffic. For example, a network optimized for video broadcasting downstream, with very little upstream capacity, would inhibit VoIP services. Network operators could also privilege the traffic of their content partners and hinder other flows. In particular, this is most likely to affect amateur or small-scale content and application providers, who are less likely able to set up partnerships with large TCNOs. One strong countervailing force is that carriers want to deliver what they call the “triple play,” to hold on to their customers by delivering all their video, voice, and data traffic.[12] Decisions in this regard are likely to come down to their assessment of whether they stand to gain more from carrying more traffic or from prioritizing certain traffic.

Whatever the case, such gatekeeping threats are easily detectable: one can imagine consumer advocates benchmarking performance on different sites. At the extreme, outright blocking of certain content or applications is unlikely to go unnoticed. In a rare case where its service was blocked by Madison River Communication, VoIP provider Vonage complained and the FCC fined the telco and forced it to stop the practice.[13] So although there is a threat at the gatekeeping level, there are also existing checks against it.

The second kind of danger will emerge as the network evolves and, we believe, is more insidious and potentially more consequential. When infrastructure owners design their network to favor specific applications and prioritize certain kinds of traffic, they inevitably limit—or outright prevent—experimentation with alternatives. For example, if the network owners decide to optimize their infrastructure to best deliver asymmetric video flows, the resulting network will inevitably be less suited to experimentation with new kinds of symmetric applications. One could argue that the innovation process leading to the emergence of a participatory, collaborative, media-rich communication platform, what many observers refer to as Web 2.0, is greatly encouraged by the availability of a neutral, E2E Internet infrastructure.[14] If the TCNOs decide to optimize the next generation infrastructure for video distribution, what are the paths of innovation that will remain open for Web 3.0? Proponents of network neutrality argue that to continue to have the level of innovation that has made the Internet a success, the network infrastructure must preserve its E2E architecture.

By contrast with gatekeeping, the consequences of evolution are much harder to monitor and assess. With respect to future evolution, the opportunity cost of not requiring network neutrality lies in avenues not explored. It is therefore impossible to know what we might miss. Network neutrality opponents would also argue that different kinds of innovation will be encouraged by allowing infrastructure owners to shape their networks in favor of certain categories of applications. For example, an Internet optimized for broadcast would probably lead to greater innovation in technologies and services supporting asymmetric, one-to-many communication patterns, perhaps even forcing a return to the kind of mass-media culture that existed prior to the spread of Internet. In the end, this may come to a trade-off between favoring innovation along a preselected path (or a few paths) versus encouraging exploration of a broad variety of evolutionary paths.

Overall, whether we consider the static impact of gatekeeping or the dynamic implications of favoring alternative evolutionary paths, there may not be a clear-cut best choice for the future of our communication infrastructure. Rather, there are a number of trade-offs to be considered, and the decision will have to be political rather than technical. In order to explore some of these trade-offs, we developed three scenarios about possible futures for the U.S. network infrastructure. These scenarios, which follow this chapter and are available on the Web at http://networkedpublics.org/conference/infrastructure_videos, were prepared for the April 2006 Networked Publics Conference to stimulate discussion about possible network futures. They were not meant to describe the entire range of possibilities but rather to help tease out some of the important dimensions of this debate about the network’s future and explore ways to resolve the associated trade-offs. Our three scenarios differ along three dimensions in particular: the extent to which the infrastructure is specialized or generic; the locus and extent of control over patterns of communication and content; and the forms and extent of privacy and security. Each relates to basic principles about the future of the communication infrastructure, and the choices made will powerfully shape how networked publics emerge, as explored in the other chapters in this book.

The first scenario, Neutral Net, assumes passage of comprehensive network neutrality rules, forcing any infrastructure owner to let others use the network on a nondiscriminatory basis. In this future, carriers would likely adjust their business models to focus on bit carriage, multiple organizations would provide competing communication services over that open infrastructure, and end users would engage in wide-ranging experimentation with applications and content. The result would likely be an infrastructure able to support a large variety of communication patterns and applications, though perhaps less directly optimized to the needs of any single one of them. Infrastructure owners would not be able to favor certain content or applications over others. Privacy and security concerns would likely have to be addressed by legislation, along a model that could be an extension of traditional common carrier principles.

By contrast, the second scenario, TCNOtopia, assumes that lawmakers reject any form of network neutrality and give TCNOs free reign to control and shape traffic as they see fit. In that future, network owners would be able to optimize their network architecture to deliver high performance for the applications and content that generate revenues for them or their business partners, without having to provide comparable performance or access to their competitors’ applications and content. They would be in charge of the shape and evolution of the communication infrastructure, including implementation of privacy and security features that enable better control of outside threats to users, such as spyware, spam, and viruses. On the other hand, network owners would have greater freedom to exploit their detailed knowledge about customers’ communication patterns for commercial gain.

The third scenario, AutoMata, explores a future in which an alternative broadband network emerges separately from the existing TCNO infrastructures. This new wireless mesh network would develop from the spontaneous agglomeration of devices, primarily in cars and other vehicles that are able to communicate as soon as they find themselves within radio range of each other. There would be no communication infrastructure per se, but rather organic, ad-hoc networks of radio devices that create multi-hop pathways for exchanging information among users. Control would be widely distributed in this decentralized network, and experimentation with new forms of content and applications is broadly empowered. However, this might result in a network highly vulnerable to privacy and security threats, a wireless Wild West of sorts, where individuals must shoulder the responsibility to protect themselves from harm and abuse.

In the next section, we examine some of the technological factors that influence these possible futures, as well as alternative possibilities for the emergence of a third broadband infrastructure that can effectively compete with the existing telephone and cable networks.

Technical Factors and Trade-offs Driving Broadband Evolution

Discussions of broadband deployment tend to focus on bandwidth (used here as synonymous with data speed) alone. Over time, users and application/service providers will demand faster broadband connections, especially for music and video applications. With audio and video clips, higher bandwidths are essential to stream data in support of a hiccup-free display. High-definition video in particular will require big broadband speeds greater than 20 Mbps.

However, other technical characteristics of the future communication infrastructure can be equally important, especially as the Internet moves beyond the current dominance of Web-based applications. In addition to bandwidth, there are four other key dimensions that can be particularly critical to the communication practices (such as multiplayer games, pervasive media, or user-produced content) highlighted in the other chapters of this book: low latency, symmetry, ubiquity and affordability, and mobility.

Latency on the Internet or in other telecommunications networks is the amount of time, generally measured in milliseconds, it takes to get a response to a request.[15] Latency depends not only on the bandwidth, or speed, of the communications lines between sender and receiver, but also on the transit time over the network and the computer processing time within the network and at both ends. For example, latency between two points on a satellite link will be one to two seconds longer than on a terrestrial link, even if the link bandwidths are the same. And latencies may differ among similar packets that take different routes on the Internet, especially if some packets go through lower speed lines or significantly more routers than others.

Low latency is critical for synchronous communication (e.g., live voice or videoconferences and live video broadcasts) and for real-time collaboration (e.g., multiplayer games or concurrent engineering). This is why satellite Internet is a poor choice for multiplayer games or videoconferences. It is also one reason why telco and cable operators want to manage Internet traffic actively within their networks, so they can prioritize delivery of packets that require low latency (and perhaps other packets for which they receive higher reimbursement). An E2E alternative is to accelerate deployment of big broadband at speeds greater than 20 Mbps in which latency becomes less of a problem.

For asynchronous communication, latency is much less critical and can be mitigated using intelligent traffic management techniques such as pre-staging of content.[16] This can be done either within the core network using carrier-controlled equipment or with user-owned servers and other equipment at the network “edges” that are not under direct carrier control.

Where latency describes the responsiveness of one’s connection, symmetry refers to the ratio of bandwidth for uploading as compared to bandwidth for downloading. Thus far, DSL and cable broadband (and so far, incumbent telco broadband over fiber in the United States) have been designed to be asymmetric, with speeds typically eight to ten times faster for downloading than uploading (see table 4.1). While such asymmetry works well for Web browsing and similar applications where users receive many more bits than they generate, applications that involve real-time video streaming from multiple sources (such as videoconferencing as well as online-based applications) can require more symmetric download and upload capabilities.

The growth of amateur, DIY video content could also affect the need for symmetric distribution bandwidth. Most DIY production is supported well by today’s asymmetrical networks. Amateurs upload their content at relatively low speed to well-connected servers from which others can download at higher speed. That is the concept of the “Google grid” and similar approaches that allow access to a grid within which users can store and publish. On the other hand, using P2P distribution (e.g., BitTorrent or P2PTV) to share content files demands more symmetric network capabilities.

Moreover, the Internet is not simply a production/publishing platform, but also a communication and collaboration channel (person-to-person, person-to-computer, computer-to-computer, thing-to-thing).[17] It is difficult today to estimate whether this traffic will be mostly symmetrical or asymmetrical in the future.

What of ubiquity? The examples of Japan and South Korea indicate that making big broadband widely available at affordable prices over fiber-optic access networks is technically quite feasible. But such connectivity throughout the United States would require new investment that would likely approach $100 billion. While the TCNOs continue to invest in extending their DSL and cable broadband networks, by 2010 fewer than 15 percent of U.S. households will have big broadband with the capabilities that are already available in Asia.[18]

Such new investment could be funded by telco and cable network operators (as in our TCNOtopia scenario), by local governments, by application/service providers (e.g., Google’s partnership with municipal Wi-Fi deployment in the TCNOtopia scenario), or by end users (e.g., in a wireless mesh scenario such as AutoMata). However, the TCNOs say they would have little incentive to invest if government enforces network neutrality. Some on Wall Street also argue that end-user affordability will depend on network operators having multiple income streams, such as charging content providers for fast, low-latency distribution. 

The arguments and trade-offs between accelerating investment in broadband networks and sustaining user innovation in a nondiscriminatory E2E environment are at the core of the current network neutrality debate.

Nor is it enough to offer ubiquitous Internet in homes and offices. As lifestyles and habits change, Internet users are demanding broadband connectivity in different locations or while they are moving from place to place. Many users now carry a portable device—such as a laptop, smart phone, or PDA—for movable access. With this device they connect to the Internet from school or work, different rooms at home, a hotel, an Internet café, an airport, or other fixed locations. The broadband access technology (DSL, cable, fiber, or wireless) will differ from place to place, although most users expect performance comparable to what they get at their primary access location. Users currently expect less from mobile Internet access while on the go, since it is available today primarily from mobile phone networks with lower performance and higher cost than other broadband options. However, expectations for mobile broadband will undoubtedly increase as new devices and networks become available.

There are two technical ways to pursue each of these issues: brute force (e.g., by deploying massive bandwidth) or clever network engineering (e.g., virtual circuits (IPv6) and quality of service (QoS) features). The first is compatible with network neutrality, but the second is generally not, as it involves making particular, low-level modifications to the way the network operates.[19] In fact, the choice also demands decisions about whether to address these technical features within the core network or outside it with user-controlled devices at its edges. This involves important trade-offs such as: Who pays for the incurred costs? Who controls access to these features? Who can capture the revenues and the benefits resulting from their use?

In the short term, these trade-offs reflect bounded choices about how best to provide infrastructure for applications we know and understand. The longer-term dynamics associated with these trade-offs are much more complex because then the choices are among different innovation trajectories and different predictions about what kinds of innovation we want to encourage for future applications we cannot yet imagine nor understand.

To a large extent, this is a replay of debates from the 1970s and 1980s about telco plans for an intelligent network versus the E2E principles that guided the designers of the early Internet.[20] Of course, there is no way to tell what innovative opportunities were lost from not pursuing the intelligent network route. But as the wealth of media on the Internet today proves, there is plenty of evidence that E2E has yielded spectacular results.

Is There a Viable Third Infrastructure?

The current discussion about the future of broadband infrastructure plays out against the background of the current last-mile duopoly of TCNOs. Both telephone and cable network operators have largely similar visions of where they are headed and, as a result, the fundamental policy debate is whether they should be left alone, or whether government should impose some rules on how they build and operate their networks.

This debate would be dramatically different if there were a third viable broadband infrastructure platform beyond the control of the incumbent TCNOs. Many of the concerns about loss of user innovation and other disadvantages to users in a TCNO-dominated future would disappear if greater competition existed for broadband Internet access. Prospective candidates for competitive access include several approaches to broadband wireless, municipal or other government-owned fiber-optic networks, and user-owned access links. From a technical and economic perspective, we are less sanguine about the prospects for broadband access over electric power lines or satellite, although satellite access can be important in rural areas where other alternatives are not available.

Broadband wireless access.

Wireless Internet access is available in several forms, including mobile phones, local Wi-Fi networks, direct point-to-point links using fixed-wireless technologies such as WiMAX, and satellite. Each has some advantages, but each faces significant problems in scaling up to a widely available and affordable broadband infrastructure. Finding a successful evolutionary path from today’s limited range and data capacities to broadly interconnected networks of wireless devices is, from our viewpoint, the key to wireless becoming a third national broadband infrastructure.

Mobile phone networks serve more than two hundred million customers in the United States and are aggressively working to expand beyond narrowband voice and text services to deliver music, images, video, and high-speed data. Building on the huge infrastructure they already have in place, U.S. cellular carriers have selectively introduced 3G services that offer Internet download speeds of 300–700 Kbps (kilobits per second). Cellular 3G offers mobile Internet access at near-broadband speeds where coverage is available, but that coverage is limited and generally costs more than DSL and cable alternatives. Not surprisingly, demand up to now appears limited primarily to businesses and high-income consumers.

Moreover, from the perspective of this chapter, current 3G and other cellular services are not compatible with network neutrality. Cellular firms operate proprietary, closed systems with active network management and full control over the services offered. For example, mobile phones, PDAs, or other devices used for Internet access on one cellular network generally cannot be used on other networks. Cellular carriers, in fact, often disable device features that would enable users to receive music, videos, or interactive services from other than their own affiliated sources. It seems unlikely that these proprietary business models will shift significantly in the next several years, particularly since Verizon, AT&T, and other incumbent wire-line carriers dominate the U.S. cellular industry. Consequently, it is unclear when, or whether, cellular networks will become more than niche competitors for broadband Internet access.

Wi-Fi networks, based on the IEEE 802.11 technical standards developed in the 1990s and using unlicensed spectrum, have grown phenomenally over the past decade, both within homes and as hot spots in commercial and public spaces. Wi-Fi hubs currently have a limited range of, at most, a few hundred feet and offer symmetric data speeds on the order of tens of megabits, which are shared by all devices on the network. Technically, however, local Wi-Fi networks do not scale very well,[21] and they are subject to interference problems.[22] Moreover, as presently implemented, they generally rely on DSL or cable connections to the Internet. Wi-Fi usage thus is covered by DSL or cable terms of service, which generally forbid unrestricted sharing as well as limited upstream data rate provisions.

Unlike 3G, however, it is possible to imagine that today’s Wi-Fi could evolve to a third national broadband infrastructure. Perhaps the most straightforward path would follow successful development of citywide Wi-Fi networks such as those under way in Philadelphia, San Francisco, and other U.S. municipalities (see our Neutral Net scenario in Appendix A). This path might well include one or more large commercial firms (e.g., Microsoft, Google, or Yahoo!) that would provide the backbone fiber-network-linking municipal Wi-Fi as well as manage some of the local systems. A third infrastructure based on municipally-owned Wi-Fi would be likely, in our view, to embrace net neutrality and open-access concepts, although that is not an inevitable outcome. Moreover, incumbent telco and cable firms strongly oppose government-owned Wi-Fi or other broadband systems and have lobbied heavily, and often successfully, to restrict their development (see our TCNOtopia scenario in Appendix A).

A second path for Wi-Fi development would be to create a ubiquitous, open access, decentralized wireless network from the voluntary interconnection of hundreds of thousands, or millions, of existing home, commercial, and public Wi-Fi hot spots. FON, a Spanish company, has an ambitious plan to provide inexpensive wireless routers to individuals and organizations who agree to share their Wi-Fi connections in return for free roaming on other Wi-Fi systems.[23] FON’s business plan relies on the introduction of new 3G mobile phones such as the Nokia E series that automatically switch to Wi-Fi whenever an open Wi-Fi signal becomes available. So far FON has focused its initial efforts largely on countries outside the United States where broadband Internet providers allow customers to share their Wi-Fi connections. While U.S. DSL and cable broadband providers generally forbid such practices, in 2007 Time Warner Cable signed a sharing agreement with FON, and AT&T began offering wireless service with the Apple iPhone that can obtain data (but, as of this writing, not voice) through available Wi-Fi networks. When, or whether, other U.S. cellular and broadband providers will enable interconnection with local Wi-Fi remains to be seen.

A third possibility, albeit futuristic, is for decentralized wireless-mesh networks to evolve from technological advances, such as those depicted in our AutoMata scenario in which cars and other vehicles serve as mobile wireless hubs. This mobile network could then interconnect with local Wi-Fi systems to spur the evolution of open, ubiquitous wireless networks in the United States and other countries.

Each of these evolutionary paths could include the use of fixed wireless technologies, such as WiMAX, to extend the range and capabilities of local wireless networks. In our view, they also would require congressional action to allocate additional unlicensed spectrum for expanded Wi-Fi and other wireless services, as has been suggested for currently unused portions of the broadcast television spectrum.[24] This could at least double the roughly 110 MHz of spectrum available for Wi-Fi in the most successful unlicensed bands below 3 GHz and go a long way toward spurring development of an alternative U.S. broadband infrastructure.

Municipally-owned Fiber Networks

Beyond Wi-Fi networks, some local governments see a municipal role in building high-capacity optical fiber networks to carry Internet traffic. The most advanced such network today is the Utah Telecommunication Open Infrastructure Agency (UTOPIA) serving 14 cities in northeastern Utah.[25] The first phase of the UTOPIA fiber network, completed in February 2006, has been funded though sales of $85 million of municipal bonds.

ISPs such as Mstar now offer phone, television, and Internet services over the UTOPIA fiber network. Residential subscribers pay $44 per month for symmetrical 15 Mbps broadband, and business customers can purchase 30 Mbps symmetrical service for $150 per month. These data rates are ten times what most asymmetrical DSL services in the United States provide at substantially higher cost per Mbps. Technically, UTOPIA’s active-Ethernet network makes it easier to offer symmetric bandwidth to customers and simpler interfaces to ISPs and content providers than is possible with the passive optical networks (PONs) used by Verizon’s FiOS and AT&T’s Lightwave services. Although initial capital costs are somewhat higher for active-Ethernet networks than for PONs, UTOPIA’s financing with twenty-year municipal bonds allows it to offer higher-speed services at more favorable rates.

UTOPIA faced many political challenges in its early years. In 2000, AT&T (then in the cable business) sought state legislation to keep local governments from competing with private networks. A last-minute amendment to the Utah Municipal Cable Television and Public Telecommunications Services Act of 2001 exempted municipal networks that did not sell retail services to customers. UTOPIA thus offers network capacity only on a wholesale basis. Mstar is currently the sole provider of voice, video, and Internet services; but UTOPIA expects to attract other providers as the system expands. Salt Lake City, Utah’s largest municipality, and three other founding cities withdrew from UTOPIA in 2003–2004 after strong lobbying by Qwest, the incumbent telephone carrier.

UTOPIA shows what is technically and economically feasible today with municipally owned broadband, and its wholesale-only model provides a clear path toward competitive broadband services that can benefit its customers in northeastern Utah. Whether other U.S. cities will follow UTOPIA’s example will depend more on politics than on technical or economic factors.

User-owned access

There are now numerous examples in Canada, and a few in Europe and the United States, of universities, schools, businesses, and public agencies financing direct links (usually fiber) from their internal broadband networks to points of presence (POPs) on the Internet where they can receive competitive services. This approach effectively bypasses local access monopolies or duopolies and inherently supports net neutrality.[26]

Besides paying for the initial capital cost, users must also arrange for normal operation and maintenance of their access links. This has proven feasible in Canada, where the concept of user-owned access is well advanced. At present, users must be sufficiently large and sophisticated to negotiate favorable arrangements for initial construction and subsequent operations. Institutions that already operate their own local area networks are thus the primary initial customers for user-owned access; but the concept can be extended to apartments, co-ops and condominiums, housing developments, and (eventually) individual homes.[27]

Opposition from incumbent telco and cable operators can be expected; but again, Canada provides successful examples of how such opposition can be overcome. In fact, incumbents would be likely to win the majority of user-financed construction and maintenance contracts if they embraced the concept. Firms that have extensive high-speed fiber networks, such as Level 3 and Google, as well as other large Internet content providers, might be interested in encouraging user-owned access as an alternative to the present duopoly. What is needed now, however, are some successful demonstrations of user-owned access at sufficient scale to show that the concept is viable for big broadband expansion in the United States.

Conclusion

Although net neutrality is currently at the center of debate over the future evolution of broadband communications in the United States, legislators and other policy makers need to consider it within a broader perspective. Like other countries, the United States is in the midst of transition from mini broadband, represented by DSL and cable modem, to big broadband that provides much higher speed, lower latency, and more symmetric bandwidth primarily over fiber optic and wireless links.[28] The precise path to big broadband is still unclear and may involve a number of twists and turns, but it is important that both public- and private-sector stakeholders keep this goal in mind in making near-term as well as long-term decisions.

Current U.S. policy favors broadband competition between telephone and cable companies, each of which owns and operates its own infrastructure, and such facilities-based competition is generally good for broadband customers and providers of online content and applications. At present, however, both telco and cable network operators have substantial market power, which permits them to limit the consumer benefits from TCNO competition. As a consequence, we advocate policies that encourage investment in independent broadband infrastructure facilities or remove barriers to their development. One example would be to remove restrictions on local government ownership of broadband infrastructure (wired or wireless), while encouraging them to provide competitive service offerings on these facilities (as exemplified by the UTOPIA model in Utah). A second example would be to eliminate government or TCNO-imposed barriers to user-owned broadband access facilities and, perhaps, to subsidize a few pilot projects to test scalability. Another important step would be to allocate more usable spectrum for Wi-Fi, WiMAX, and other broadband wireless applications. In each case, the overall benefits would come not just from the availability of an additional broadband alternative, but from the salutary effects this would have on telco and cable service offerings and pricing.[29]

Calls for net neutrality are a response to growing market power of the TCNO duopoly during the transition to big broadband. But because broadband is a moving target—20 Mbps may well be considered mini broadband within a very short time—it is difficult to legislate or adopt regulatory rules that will both be enforceable and remain relevant.[30] An approach set out in the March 2006 “Annenberg Center Principles for Network Neutrality” (see box 4.4) focuses on “light touch” regulation, emphasizing general principles of competition policy where network operators have significant market power, rather than detailed rules for all broadband providers. In any event, network neutrality may be only the first topic in an ongoing public debate over how to make U.S. broadband competitive, ubiquitous and affordable, as well as a continuing source of technical and social innovation.

Appendix: Three scenarios for U.S. broadband access evolution

1. Neutral Net

It’s 2017. The U.S. government runs the national communication GRID (Government-Run Information Distributor), comprised of the country’s fiber optics, cables, and radio links. Access to the GRID is open to all, on an equal basis, for any application and any content. Most of the population now creates and shares media of all kinds—what their productions lack in polish and sophistication, they make up in imagination.

Thanks to the government-run GRID, there no longer is a divide between urban and rural areas. The open access GRID has ushered in the era of micropolitics: every conceivable constituency can propose any initiative at any time and set up a virtual debate space and e-voting mechanisms.

Neutral Net was set in motion in 1983, when the FCC forced the local phone companies to let all enhanced service providers use their wires for free. Within a few years, thousands of ISPs jumped at the chance to offer new services without the need to invest in costly networks. With the release of the Mosaic Internet browser in 1993, a new mass medium was born. Soon after, in 1995, DSL and cable modems turned the old phone and cable television networks into broadband always-on information networks.

During the next ten years, a multitude of innovators built upon the open Internet to offer new communication services that radically transformed people’s ability to create, share and access information.

In our scenario, in September 2010 the U.S. Congress decided it essential to preserve the Internet’s openness. Strict rules forbid all network owners, telephone, cellular, and cable alike, to discriminate among users. They are not allowed to favor any traffic, nor to charge different fees for different users or different applications.

Anybody can now provide any communication service over the carriers’ networks. Wal-Mart introduces low-cost “WAL-Media”: their branded combination of wired and wireless Internet access, voice and text communication, and film and video distribution.

In the next few years, amateur production of content explodes. YouTube and MySpace garner audiences that far surpass those of traditional television channels. Blogs have now replaced newspapers as most people’s primary source of news. The Net supports a vibrant public sphere in which all constituencies find a voice, a virtual town hall, and viral tools to mobilize voters and make their voices heard.

To sort through this massive amount of news, debates, games, music, video, and films, users rely on each other. Social filters, recommendation engines, and distributed online marketplaces allow them to find, discover, rank, and select materials that match their passions.

Every device on the network is a server, whether in homes, public places, small businesses, or civic organizations. They support P2P communication tools; distribute user-produced stories, songs, and videos; and host collaborative spaces that bring together families, workgroups, clubs, churches, or citizens.

A growing number of cities build their own Wi-Fi and fiber networks to foster greater civic Internet use. However, funding for professionally-produced premium content starts to decline, partly because it is impossible to guarantee the network performance that would allow optimum delivery of that content and partly because P2P distribution of pirated content proliferates (it is hard to maintain control over intellectual property now that a multitude of service providers operate over the networks).

By 2012, network owners are unable to raise funds to upgrade their networks. Verizon discontinues FiOS, and AT&T abandons project Lightspeed. Cellular networks never fully upgrade to 3G. The network owners decide to become pure bit carriers, scale down their production and programming operations, and concentrate instead on cutting their costs down to a minimum, retaining only skeleton maintenance crews.

Meanwhile, although content from millions of amateur sources is now available, Hollywood loses its preeminence as the world’s main center of content production. Instead, big-budget entertainment is now produced in countries like China, France, and India, where the network owners keep tight control over who distributes what and can thus guarantee protection of their intellectual property.

By 2014, investment in the U.S. network infrastructure has fallen so low that its derelict state resembles that of the nation’s bridges and roads. To ward off catastrophic failure, the U.S. government takes over all communication networks, consolidating them into the GRID. A new tax on advertising is created to fund the GRID.

By 2017, the GRID provides uniform Net access throughout the U.S. territory. The nation ranks a weak twenty-ninth in the OECD’s assessment of broadband performance, but a dynamic community of users constantly invents new ways to squeeze extra bits out of the country’s infrastructure.

U.S. elites are dissatisfied with the poor performance of the national GRID. They live in teleparks, the new gated communities, which tend to congregate in border cities and ports, where they get easy access to foreign network head-ends and submarine high-capacity fiber.

2. TCNOtopia

In the year 2017, two huge TCNOs control broadband Internet access throughout the United States. Each TCNO has its own content affiliates who provide online entertainment, sports, games, and information to the consuming public. Their operational motto is “we create, you enjoy.”

The path to TCNOtopia began in 1969, when the first bits sped across a new computer network funded by the U.S. Department of Defense. Soon the elements of what would become the Internet were in place: an open architecture where users innovate at the edges of the network and E2E communications with no gatekeeper inside the network core, all of it riding on top of the nation’s phone network, providing little compensation to the telcos who had built that infrastructure. In fact, the Internet stands in sharp contrast to telephone and cable visions, which place intelligence, control, and innovation inside the network.

By 2007, the TCNOs provide more than 96% of residential broadband connections. But most of the real profits are made by firms who use the TCNO networks, such as Microsoft, Amazon.com, Google, Yahoo!, eBay, and Disney. Verizon and AT&T fight back with Internet television, offering hundreds of channels and thousands of hours of on-demand programs. Like the cable companies, they want to choose the content they deliver over their broadband pipes and not simply act as common carriers. AT&T’s CEO declares that Internet content providers will have to pay extra for fast broadband delivery.

In reaction, content providers join with consumer groups to persuade Congress to preserve network neutrality. But they get a chilly reception in Washington. Instead, Congress gives telcos authority to freely offer Internet programming and decide what traffic gets priority within their network.

2010: Based on the early success of Wi-Fi in Philadelphia and San Francisco, Google launches broadband wireless nationwide in partnership with local municipalities.

Verizon and AT&T, followed by the cable operators, offer contracts to Sony, Fox, Disney, and others for fast-lane Internet delivery of their online games, movies, video, and other content. Those who choose not to pay must accept standard delivery. This slow lane is where user experimentation is allowed, the only option for user-run servers, and P2P and other applications unaffiliated with the carriers. To enforce the separation, the TCNOs now scan all data packets. Customer contracts authorize carriers to screen for viruses, spam, copyright violations, and content of interest to government agencies. These contracts also limit the bits users can upload without paying substantially higher fees.

2011: Most large content providers are enthusiastic about fast-lane delivery. They can now charge higher fees for premium media experiences. But some, like Google and Microsoft, mount court challenges to packet scanning and prioritization as violations of users’ privacy rights and of network operators’ obligations to provide common carrier services.

2012: Flush with cash from content providers, TCNOs accelerate investment in fiber infrastructure and in-network innovations to achieve high performance. Dozens of new services, such as online multiplayer sports and games, become wildly popular. With full control over individual data streams, the carriers can craft compelling multimedia experiences for their customers. TCNO interface equipment in the home also optimizes the user experience, while preventing unauthorized copying of content or the bypassing of advertising messages.

Meanwhile, Google’s broadband wireless buildout has achieved initial success with four million subscribers in twenty-eight cities. However, security and reliability concerns arise after hacker attacks disable some fifteen thousand wireless-enabled computers in Chicago and Los Angeles. The TCNOs effectively use this security failure in their broadband marketing campaigns.

2014: The U.S. Supreme Court rules in favor of the TCNOs’ right to scan data packets and prioritize Internet traffic. The decision cites the need to ensure network reliability and protect customers from hacker-induced harm.

2016: The merger of Comcast and Time Warner creates a behemoth controlling 90 percent of the U.S. cable market and 60 percent of all broadband connections.

After reporting billion dollar losses, Googlezon (formed by the recent merger of Google and Amazon.com)[31] abandons its municipal wireless partnerships. Some cities vow to keep their networks on the air, but it appears an uphill struggle against the dominance of TCNO broadband.

2017: Determining that only increased scale can compete effectively with Comcast Time Warner, the Justice Department approves the merger of AT&T and Verizon. The broadband duopoly has no serious rivals. It has brought affordable broadband to 85 percent of U.S. households, who love the network innovations that protect against spam and viruses, the e-sports leagues, and the high-definition entertainment they receive from TCNO content affiliates.

Political expression online is encouraged within the established political structure—primarily through the two dominant national parties that have negotiated fast-lane delivery for their candidates and issue messages. Other political organizations and civic groups must negotiate ad hoc arrangements, and few have the financial resources to assure fast lane delivery of their messages.

Still, some academics, artists, and other dissidents bemoan the loss of amateur content production and collaborative activity that flowed over the Internet in the early-twenty-first century. TCNO restrictions have virtually eliminated P2P communication among residential broadband users for content distribution, collaborative work, or social and political organization. Online distribution and collaboration are channeled through TCNO-controlled servers and routers. As a consequence, Internet content and applications now conform closely to established consumer tastes and traditional values. It is nearly impossible for an innovator that is unaffiliated with the TCNOs to gain a sizable Internet audience in the United States

But perhaps another eBay, Napster, Yahoo!, Amazon.com or Google is ready to emerge out of the competitive chaos in India or Brazil.

3. AutoMata

In the year 2015, the Internet addresses the last-mile challenge using a mesh network of wireless devices named AutoMatas. The traditional telcos are now relegated to routing the backbone traffic. AutoMatas are self-organizing devices that communicate with one another when in each other’s radio range. They may use access points available at hot spots for access to the telcos’ wired infrastructure. AutoMatas are now ubiquitous in society and a standard feature in all new vehicles. Their widespread use contributes to augmenting the capacity of the wireless mesh network that effectively transmits the last-mile traffic.

An early example of wireless ad-hoc networks is the use of short-range radios by truck drivers to communicate road conditions and socialize while traveling. The need to exchange information and communicate while in motion was later satisfied through the use of mobile phones, although this system does not operate on an ad-hoc mesh network. The origins of the wireless ad-hoc network organized around AutoMatas dates back to 1991, when Vic Hayes of NCR Corporation in the Netherlands invents Wi-Fi for use with cashier systems. Prior to his retirement in 2003, he shapes the design of standards such as IEEE 802.11a, b, and g, becoming the “father of Wi-Fi.” The devices that operate on this standard have a limited radio range in the order of tens of feet and offer bandwidths in the order of tens of megabits per seconds.

In the first years of the new millennium, a number of different events begin to lay the ground for the emergence of a wireless mesh network. Lightweight portable products with gigabytes of memory storage, such as the iPod, hit the market enabling the consumption of digital entertainment on the go. COVERGE 2001 is a first conference dedicated to convergence of automobiles and computers. Participants discuss standards for an in-vehicle multimedia network. Mesh networks emerge in neighborhoods using Wi-Fi devices based on IEEE 802.11a, b, and g and offering Internet services to their residents.

2009: Advent of Wi-Fi tailor-made for mobile vehicles. The technological leap to enable Wi-Fi on mobile vehicles occurs in 2009, when companies introduce products based on 802.11p, also referred to as Wireless Access for the Vehicular Environment (WAVE), offering radio ranges in the order of a thousand feet with bandwidths in the order of megabits per second. Such ranges make possible meaningful exchange of information between vehicles that are in motion and base stations.

2010: The U.S. Congress makes a significant part of the analog broadcast spectrum available for unlicensed use. Groundbreaking work in interference-avoidance techniques has enabled efficient use of this spectrum. This big-broadband wireless spectrum represents the future of wireless communications. The unlikely group of successful bidders for the spectrum is composed of EarthLink, Sony, Cingular, GM, and Toyota. While EarthLink and Sony are preparing for general distribution of wireless content, the car companies are observing this from the consumer’s end: soon vehicles will be capable of receiving massive quantities of wireless content reliably, and new vehicle designs in the works are incorporating these features.

2011: The year 2011 witnesses the development of AutoMata. These wireless devices, looking like iPods and weighing less than 10 ounces, store audio and video clips, implement navigation and GPS capabilities of an automobile, and detect one another for multi-player games that extend a physical world with virtual objects. AutoMatas include terabytes of storage, and several types of wireless cards such as cellular, 802.11g, 802.11p, and Bluetooth. These networking cards operate in a variety of radio ranges (from a few feet to miles) and bandwidths (tens of kilobits to hundreds of megabits per second). Numerous car manufacturers begin to combine the vehicle’s telematics, navigation, and entertainment systems into AutoMatas for new models. These devices plug into an in-vehicle network to guide a driver to a destination and deliver entertainment content. Its owner may carry the device and use it as a personal digital assistant and personal audio and video entertainment unit.

2012: Hackers publish the programming interface to AutoMata, adding features such as text-to-audio and a VoIP interface. Using a Bluetooth-enabled headset, a user may listen to e-mail messages. Consumer electronic vendors are quick to adopt these novel features, publishing official versions for download by all. This grassroots effort introduces many safety features for in-vehicle use. For example, when in radio range of one another, AutoMatas exchange traffic information and hazards such as icy road conditions. An elegant holder on the front dashboard of a car provides for both re-charging of the AutoMata and its view of the road in front of the automobile. A passenger may view what another vehicle’s AutoMata might be recording several minutes ahead. Insurance companies start to give incentives to drivers who use voice activated AutoMatas that warn them of road hazards.

2011: Competing AutoMata-like devices reach the market, driving down prices and improving their efficiency and capabilities. In addition an active community of programmers continue to develop applications and operating systems and standards for AutoMatas. Devices are everywhere, creating an ad-hoc mesh network that begins to produce and route growing amounts of Internet traffic. In particular, the planned pre-staging of popular content to make it accessible to the mesh network is effectively reducing the last-mile Internet traffic. Users begin to switch from home ISP subscriptions to maintaining their AutoMata connectivity, much like mobile phone usage affected standard telephone lines. Telcos point to the growing traffic created by the mesh network, and its lack of security, to insist on the need to manage traffic on the Internet backbone to justify the investment in increased capacity. There is a surge in spam and identity theft due to the increasing Internet traffic on the mesh network. Soon telcos develop enhanced certification methods for personal messages and improved cryptography techniques for security introduced by academics and other users of the network.

2014: GM and Ford introduce software for AutoMata, enabling cars from 2014 onward to drive without human assistance on certain freeway stretches. A valid driver’s license holder must sit in the driver’s seat and agree to the risks of using AutoMata. When an intelligent road stretch is encountered, AutoMata signals its driver that it may assume responsibility for driving the car. These devices communicate with one another to share information about icy road conditions, accidents such as chemical spills, and other emergency situations. Devices minimize accidents by slowing down and stopping when too close to one another.

There is a steady reduction in the use of mobile phones once the mesh network provides better coverage in most urban settings than the nonintegrated cellular networks due to the ubiquitous presence of the AutoMata. Mobile phone calls are now mostly conducted with the AutoMata using VoIP.

2015: An AutoMata is now the size of a company pin that one wears on a business suit. It records and stores hundreds of hours of phone and live conversation, along with hours of video recordings. Its battery works for days of usage. For their display, AutoMatas utilize TVs in a living room, a laptop’s display at work, a mobile phone’s display on the road, and a car’s navigation display or fold-down screen. New and inexpensive displays start to appear in automobiles and coffee houses. Menus at high-end restaurants turn into a display and a keyboard for AutoMatas.

A year after the introduction of automatic driving by AutoMatas, observing a significant reduction in the number of accidents, many insurance companies give incentives to those who employ AutoMata to drive. In collaboration with tens of insurance companies, the U.S. Department of Transportation deploys Internet wireless hubs on most highway stretches in the country. Its primary purpose is to increase road safety and enhance emergency response. This new infrastructure helps create an alternate Internet backbone to the mired telco network and its traffic-management practices. There is now an alternative backbone for delivery of delayed modes of communication, alleviating the load on the telco’s wired backbone infrastructure.

2017: In cities such as Los Angeles a growing number of people are driven by their AutoMata from place to place. A vehicle then has become an extension of both the office and home. Special content is released for in-vehicle entertainment systems, capitalizing on the enclosed nature of the vehicle to immerse passengers in an experience.

Bollywood of India finally supersedes Hollywood in revenues by generating content tailor-made for in-vehicle use. Widespread amateur production and improved social filters have enabled a broad dissemination of diverse content and ideas, effectively “fattening the long tail” for entertainment, commerce, and political movements. Every idea is capable of finding its audience, however, due to the large number of possibilities available, advertising on the Internet and on hardware is still able to shape public opinion. Much like important content producers such as Disney, powerful interest groups such as the NRA and Texans for a Democratic Majority promote their message through product placement on antivirus and antispam software and new equipment. In addition, content can be prioritized by pre-placing content through the access points to the mesh network for a fee.

Widespread use of AutoMata has lead to a decline of home ISPs; local Internet traffic is routed through the mesh network, and traditional telcos are relegated to routing part of the backbone traffic. Bandwidth and latency on the mesh network is affected by the number of AutoMata present.