This report is available in Adobe Acrobat format as well. Front Cover (2.1MB) Report (596kb) Back Cover (2.4MB)
Issued by the
Broadband Access Project
of the
Center for Democracy & Technology
December 2000
The Center for Democracy and Technology ("CDT") is dedicated to ensuring that democratic values and constitutional liberties are a central feature of the new digital age. With its unique mix of expertise in law, technology and public policy, CDT works for practical, real-world solutions that enhance free expression, privacy, open access and democracy in the rapidly evolving global communications technologies. CDT endeavors to build consensus among all parties interested in the future of the Internet, finding common ground among activists, nonprofit groups, Internet businesses and government policymakers.
Following the passage of the Communications Decency Act in 1996, CDT helped to organize the Citizens Internet Empowerment Coalition. Through the CIEC, leading members of the Internet industry challenged CDA’s constitutionality in American Library Association/ACLU v. Reno. When the case was argued in Philadelphia, the coalition brought the Internet into the courtroom through high-speed data lines. And when the case came before the Supreme Court, it was argued by the coalition’s counsel.
With its Broadband Access Project, CDT seeks to ensure that the characteristics of the narrowband Internet that were so critical in Reno, and the legal principles that came out of that case, continue to thrive as the Internet moves into the broadband world. The Project has looked at all forms of broadband access that are emerging as ways to reach the Internet, including cable modems, digital subscriber lines, satellites, and terrestrial wireless services. CDT has worked closely with a broad cross-section of the Internet, computer, and communications industries, as well as with consumer groups and other interested parties. It has developed a comprehensive and balanced assessment of where the technology is today, where it can be tomorrow, and what impact the new technology will have on speech and access to content on the Internet.
This Paper was prepared by John B. Morris, Jr., Director of the Broadband Access Project. Mr. Morris is a Partner with the Washington, D.C. office of the law firm Jenner & Block. In 1996, he was one of the lead trial counsel representing the industry and association plaintiffs in American Library Association v. Reno/ACLU v. Reno, the constitutional challenge to the Communications Decency Act. During 1999-2000, he took a leave of absence from his firm to work with CDT and its Broadband Access Project.
TABLE OF CONTENTS
The extraordinary growth and innovation of the Internet, its ability to empower individuals, and its role in promoting free expression and democratic values depend critically on openness principles that have characterized the "narrowband" world of dial-up access. Some of these principles relate to the openness of the Internet's standards and software. Some are engineering principles, designed to make the Internet function flexibly and efficiently. Others are policy choices regarding the telephone system, made before the Internet existed. These architectural elements and policy choices have implemented values as well as enabled communication.
The openness of the narrowband Internet has been responsible for the most democratic and dynamic form of mass communications ever created. As expressed by Judge Dalzell in ACLU v. Reno, "the Internet is a far more speech-enhancing medium than print, the village green, or the mails." The Internet allows anyone to reach the entire world simply and inexpensively. It enables the unprecedented ability of speakers to speak and allows listeners to receive content, free from governmental or private interference.
Preserving openness and the dynamic nature of the Internet is critical to maintaining the democratic character of this medium as the Internet is transformed from the narrowband technologies of dial-up modems and slow content delivery to the "broadband" world of cable modems, DSL, wireless, and other technologies that deliver high-speed Internet access. Emerging broadband Internet technologies offer advantages over narrowband access that will enhance and expand the Internet's usefulness to users. Broadband Internet will allow subscribers to send or receive video and audio content of digital quality and to download interactive, graphic-rich webpages. The high-speed technology will enable entrepreneurs to bring new services to market that will make the Internet interactive in real time.
However, critical differences between the narrowband and broadband Internet could change the open nature of the Internet and raise the possibility that this dynamic and democratic medium might come to be dominated and in part controlled by a small number of private companies. Until now, this open quality has allowed Internet users a wide range of choices about how to access the Internet and what to do with the communications medium once online. As the Internet evolves from narrowband technologies to broadband, it is imperative to maintain openness and the empowering and democratizing characteristics of the Internet that flow from that openness. Policymakers face the momentous challenge of ensuring that the empowering aspects of the Internet of the past ten years are carried over into the Internet of the new century.
The Center for Democracy and Technology is committed to preserving the open character of the Internet. From the outset, CDT helped define the vision of the Internet as a uniquely user-controlled, decentralized, democratic medium. CDT was at the forefront of the legal challenges to the Communications Decency Act and framed the legal strategy that culminated in the Supreme Court's landmark decision in Reno v. ACLU extending to Internet speech the highest level of constitutional protection.Since then, CDT has been instrumental in educating policymakers about the empowering potential of an open, accessible Internet.
CDT believes that it is imperative to ensure that the open and democratic characteristics of the narrowband Internet central to the Reno decision are carried over into the emerging broadband Internet. To evaluate the potential impact of the evolution from narrowband to broadband, CDT undertook its Broadband Access Project, an objective assessment of the factual and public policy issues raised by emerging broadband delivery technologies.
This paper explains what the broadband openness debate is about, and describes the major broadband delivery technologies, focusing on cable modem access. It finds that opennessis feasible and critically necessary across a broad range of technologies to preserving the free-speech and democracy enhancing character of the Internet. It provides a factual primer and considered analysis of the issues, while it leaves to policymakers and public debate the questions of what constitutes openness, and how best to achieve it.
The Narrrowband Internet
The beginnings of narrowband Internet are well known, but it is worth recalling their implications for the transition to broadband. Narrowband Internet had its origins in ARPANET, an early data communications network developed in the late 1960s under the auspices of the government's Defense Advanced Research Projects Agency ("DARPA"). A number of different individual networks of linked computers developed in the 1970s, typically to serve specific segments of the academic or governmental communities. The National Science Foundation ("NSF") spearheaded an effort in the 1980s to link these networks using common protocols. NSF, while it financed the initial Internet "backbone" carrying communications between the linked government and academic networks also through its policies encouraged the development of private competitive backbone networks. The usefulness of these networks depended upon the ability of individual users to access them via the "last mile" of the telephone line via modem.
It is accepted wisdom that the Internet has prospered because it has not been regulated. On many levels this wisdom is true, except it is also true that the Internet prospered because the Federal Communications Commission ("FCC") imposed significant regulation on the facilities over which most people access the narrowband Internet ‚ the telephone system. The successful shift of the narrowband Internet from a publicly funded effort to a private sector, commercial effort providing services to individual users through the telephone network depended on critical non-regulatory and regulatory decisions by the FCC ‚ in some cases to regulate, in others not to:
Taken together, the First Computer Inquiry and Carterfone decisions were significant and affirmative regulatory steps that permitted the creation and ultimate explosion of the Internet.
Rollout of Broadband Technologies
Technologies supporting high-speed transmission of data have been available for years, but only at a substantial price. Widely affordable broadband services are currently emerging, along with a mass interest in and market for those services. The new market for broadband services is an outgrowth of technological improvements, increased competition, and increased acceptance of and interest in the Internet.
At its simplest, broadband allows higher speed transmission of data. But it also makes it possible for the user's Internet connection to be "always on" ‚ in other words, the Internet can always be available, allowing users to seek information from the Internet far more often than with dial-up access. Broadband also allows users access to much more video and other high bandwidth content than would be possible over a narrowband connection.
These capabilities give users a far wider range of potential uses for the Internet. Professionally, users will be able to telecommute and establish home offices by enabling access to corporate networks, e-mail systems, and videoconferencing. Broadband will facilitate the creation of home-based businesses though web serving, e-commerce with customers and financial functions. Broadband Internet will allow a wider range of entertainment activities, including web surfing at higher speeds with richer video content, video on demand, and interactive, multi-player video games. Consumers will be able to shop, access telemedicine, participate in distance learning, take advantage of public services, research information and videoconference with friends and family.
Technologies for Broadband
There are a variety of different and competing technologies that promise to deliver broadband Internet access to both individual and business users. The leading technologies are Digital Subscriber Line ("DSL") service and "cable modem" service, with wireless, satellite and other technologies still emerging.
Cable: In the mid-1990s, the cable industry undertook the massive and expensive conversion of its traditional cable facilities into systems that (a) could support two-way transmission of signals, and (b) had a much higher capacity to support analog video signals, digital video signals, and data signals. As of August 1999, over one million homes in North America subscribe to cable modem service, 300,000 of which are in Canada. Cable systems pass approximately 90% of the homes in the US. The leading US cable operators forecast that 90% of their systems will have been upgraded by the end of the year 2000, and forecasts suggest that by 2001, cable modem service will be an option for as many as 80 million homes in the United States.
Digital Subscriber Line ‚ DSL: Digital Subscriber Line ("DSL") uses existing copper wires to provide high speed data services. The technology was developed in the late 1980s and was first widely used for non-Internet specific applications. With asynchronous DSL ("ADSL"), a high-speed data connection can be run over the same wire that is used to carry a regular phone line. Thus, ADSL can share an existing wire with a user's existing voice telephone line, permitting both DSL and telephone services to be used at the same time. ADSL service operates on dedicated wires that are unaffected by high usage by neighbors. Theoretically, ADSL service will provide more reliable, but somewhat slower Internet access than cable modem service. There are two primary types of DSL providers: Incumbent local exchange carriers ("ILECs") such as Verizon and Bell South (historically the primary local telephone companies in their regions), and competitive local exchange carriers ("CLECs"), such as Covad. At the end of 1999, an estimated 500,000 DSL lines were deployed in the United States, with about 75% being ILEC lines and 25% being CLEC lines. Deployment has exploded in the year 2000, and significant growth is expected in later years.
Other Technologies: Broadband will also be available over terrestrial wireless and other technologies. Although significantly less established and deployed than cable and DSL broadband services, terrestrial wireless services using land-based transmitters can also provide broadband services to businesses and individuals. While a number of different technological approaches will be available, it is not clear which will be successful in the marketplace. Nevertheless, it is likely that for some categories of users, broadband wireless will compete with cable and DSL services to provide broadband access to the Internet.
Beyond cable, DSL, and terrestrial wireless services, other broadband access methods are still in development or testing. Satellite access technology will use satellites to deliver Internet access to homes and businesses. Fiber optic lines running directly into users homes and businesses ‚ often referred to as "fiber-to-the-home" ‚ will deliver broadband. Although fiber does not share the same technological uncertainties facing satellite systems, it is not clear whether it will be economically viable to run fiber optic wires through existing residential neighborhoods. Finally, a number of schemes to use power lines to transport Internet access services have been explored over the past few years, but none of the yielded marketable service offering. If powerline based services are successfully developed, utility companies could become players in the broadband access market.
An Overview of the Policy Debate
The most significant policy questions have involved the cable industry and efforts by ISPs and others to offer services over cable systems. Other questions relate to the provision of DSL services, while other issues are more general in focus. Resolution of the cable or DSL specific issues could have direct implications in other technology areas. For example, how policymakers resolve the "access to cable" issues could affect the development and deployment of wireless technology. These decisions should be made with a broad perspective of the Internet.
Cable-Specific Issues: The key broadband issues that specifically relate to Internet access over cable systems are rooted in the broad policy question of whether cable-based Internet access should be treated similar to cable television service (which has been largely unregulated) or similar to telephone service (which has been subject to significant regulation). Some cable companies have argued that additional regulations should not be imposed on the cable industry, even if it is providing Internet access. On the other hand, proponents of "open access" have argued that the regulations applicable to the telephone system has been vital to the growth of the Internet, and similar regulations should be imposed on Internet access over cable. Finally, some local telephone service providers have argued that the incongruity in the levels of regulation for the telephone and cable industries should be resolved by reducing regulations applicable to telephone service, rather than by imposing regulation on cable-based Internet service.
DSL-Specific Issues: The broadband issues relating to Internet access over digital subscriber line ("DSL") systems generally involve the ability of competitive local exchange carriers ("CLECs") to compete with incumbent local exchange carriers ("ILECs"). Under the Telecommunications Act of 1996, ILECs are required to make certain pieces, or "elements" of their network available to CLECs to enable CLECs to compete with the ILECs in the provision of local phone service and Internet access service. The policy questions center around whether the ILECs are fully complying with their statutory and regulatory obligations to act properly toward, and compete fairly with, the CLECs. Most of the DSL specific issues reflect CLEC challenges to ILEC actions, and are already the subject of legal or regulatory proceedings at the state and/or federal levels. As a general matter, the CLECs do not seek major changes in laws or regulation, but instead seek stronger and faster enforcement of existing regulations and statutes.
The Emerging Broadband Content Distribution Model: A vitally important result of the Internet's infrastructure is that any speaker on the Internet can reach any listener. On the Internet, a single speaker and the largest media company have roughly the same abilities to speak and be heard. There is significant risk on a broadband Internet, where locally-based broadband content servers deliver broadband content quickly to consumers, the major means of broadband distribution will be the proprietary domains of large companies or wealth speakers, ending the rough equality among speakers so critical to the Internet's promotion and facilitation of democracy.
A Focus on the Cable "Third Party Access" Debate
Far and away the most contested issue relating to broadband access to the Internet is the "open access" or "forced access" issue ‚ whether cable system operators must permit unaffiliated Internet Service Providers to offer Internet access services over the cable facility. As is done in Canada, where cable modem deployment and "open access" are both more advanced that in the United States, this paper uses the term "third party access" to refer to this policy issue.
Without drawing conclusions, the paper looks closely at the specific arguments and claims made in the third party access debate. Unlike the telephone system, cable has not been subject to regulation as a common carrier. The paper examines the decision not to regulate cable and its effect on the debate about imposing a third party access requirement on the calbe industry. The paper probes the claims made by advocates of mandatory third party access, including the risk of censorship and concerns about anticompetitive behavior. It also discusses the claims made by opponents of mandatory third party access, including the incentives to upgrade the cable systems to support Internet access, the potential for a regulatory morass, and the constitutional rights of cable operators.
However, the paper does find that openness is feasible. It examines the technological issues surrounding feasibility, concerns raised by the shared nature of the cable network, and whether there is validity to the much debated (but no longer enforced) "10 minute" limit imposed on streaming video over cable systems. For each of these issues, the paper offers an informed and balanced assessment of whether the claims and arguments are justified. It finds that, as demonstrated by the recent movement by leading U.S. cable companies to accept some form of voluntary open access, no one is strongly asserting that it is not possible. It further finds that, while concerns are inherent in the provision of Internet access over a cable system, many of these are present with or without third party access, and many will be resolved as engineers turn their attention to designing equipment that supports third party access in cable systems . As the debate has evolved, there is no longer any serious question about whether third party access is feasible or desirable ‚ it is both. The main focus of the debate today is on how third party access will be implemented.
Open and decentralized, the current "narrowband" Internet is the most democratic and dynamic form of mass communications ever created. The narrowband Internet operates without gatekeepers. It is not limited by scarce spectrum or dependent on other scarce resources. Anyone who wants to can be a publisher and can reach the entire world simply and inexpensively. Barriers to entry are low, and anyone who wants to can become an Internet Service Provider and a part of the global network on an equal par with others. The Internet's broad availability gives business, non-profit organizations and individuals an unprecedented ability to speak, and allows listeners to receive information, free from governmental or private interference. The narrowband Internet is a medium that, as Judge Dalzell stated in his decision in ACLU v. Reno, "provides significant access to all who wish to speak in the medium, and even creates a relative parity among speakers." [ I-1 ]
Today, the Internet is on the threshold of a transformation from the narrowband technologies of dial-up modems and slow content delivery speeds to the broadband world of cable modems, DSL, wireless, and other technologies that deliver high-speed Internet access. Emerging broadband Internet technologies offer tremendous advantages over narrowband access that will enhance the Internet's usefulness to users. But as the Internet shifts from a narrowband foundation to one based on broadband technologies, it is imperative that its openness, and the empowering and democratizing characteristics of the Internet that flow from that openness, are not lost.
The Center for Democracy and Technology is committed to preserving the open character of the Internet. From the outset, CDT helped define the vision of the Internet as a uniquely user-controlled, decentralized, democratic medium. CDT was at the forefront of the legal challenges to the Communications Decency Act and framed the legal strategy that culminated in the Supreme Court's landmark decision in Reno v. ACLU extending to Internet speech the highest level of constitutional protection. Since then, CDT has been instrumental in educating policymakers about the empowering potential of an open, accessible Internet.
CDT believes that it is imperative to ensure that the open and democratic characteristics of the narrowband Internet central to the Reno decision are carried over into the emerging broadband Internet. To evaluate the potential impact of the evolution from narrowband to broadband, CDT undertook its Broadband Access Project to objectively assess the factual and public policy issues raised by emerging broadband delivery technologies.
CDT believes that critical differences between the narrowband and broadband could change the essential open nature of the Internet and raise the possibility that this open, dynamic and democratic medium might come to be dominated and in part controlled by a small number of private companies. Until now, the openness of the Internet has allowed users a wide range of choices about how to access the Internet and what to do once online. Policymakers must understand the impact of broadband technology on the Internet's openness, and should ensure that the empowering aspects of the Internet of the past ten years are carried over into the Internet of the next century.
The goal of this paper is to outline the basics of the broadband openness debate, to describe the major broadband delivery technologies, focusing on cable modem access, and thereby to provide policymakers and interested parties with considered analysis of the issues as they make decisions about what constitutes openness, and how best to achieve it. This paper finds that, to preserve the free-speech and democracy enhancing character of the Internet, openness is critically necessary across a broad range of Internet access technologies. It also finds, focusing on cable access, that it is technologically feasible to have openness. As the paper notes, the "open access" debate has shifted from whether there can and will be third party access to how such access will be implemented and monitored, and whether it should be mandated. As a result, while there are many challenging issues that remain to be resolved, the paper concludes both that openness is technically feasible and that implementing it is crucial to preserving the democratic essence of the Internet in the broadband world. [ I-2 ]
To make decisions on the policy issues that arise from the emergence of broadband technologies, it is important to understand how the Internet developed, how the narrowband Internet works today, and how regulation affects the Internet.
The Internet originally grew out of the ARPANET, an early data communications network developed in the late 1960's under the auspices of the Defense Advanced Research Projects Agency ("DARPA"). [ 1 ] Key to its creation was the development of open, freely available standard protocols that for the first time allowed computers to communicate with each other, and then allowed networks of linked computers to communicate similarly with other networks.
A number of different networks of linked computers developed in the 1970's, typically to serve specific segments of the academic or governmental communities. Initially, many of these networks did not communicate with each other. Through the efforts of the National Science Foundation ("NSF") in the 1980's, many of these networks were linked together using common protocols. This "network of networks" came to be known as the Internet.
In addition to spurring academic and governmental networks to "internetwork" with each other, the NSF also financed the initial Internet "backbone" — a national communications network that carried communications between the linked networks. In an effort to spread the cost and take advantage of economies of scale, the NSF encouraged local and regional networks to offer network access to commercial customers. However, commercial communications over the NSFNET backbone were not permitted. This policy of denying commercial users access to the backbone, as it was intended to do, prompted the development of private, competitive backbone networks such as PSI and UUNET.
In the late 1980's, the NSF determined that the Internet would evolve most efficiently and effectively if the network was privatized. NSF worked towards this goal until 1995, when it halted funding and operation of its NSFNET backbone and began to help regional networks purchase backbone connectivity from private networks. Private business and consumer oriented networks had evolved during the 1980's, and over time these networks linked to the Internet.
In the early 1990’s, at the same time that the Internet was coalescing as a global network of primarily private networks, a researcher in Switzerland conceived of and developed the World Wide Web. At its core, the Web is a set of "hypertext" protocols that operate over the Internet.[ 2 ] They provide an easy way both to publish and to access information over the Internet. With the development and popularization of the Web, commercial consumer networks such as Prodigy, CompuServe, and America Online experienced explosive growth, and the social and commercial phenomenon that is the Internet began.
Critical foundations for the Internet are communications protocols named "Transfer Control Protocol" and "Internet Protocol," which together are known as "TCP/IP." These fundamental protocols allow millions of otherwise incompatible computers and computer networks to communicate. A key innovation offered by TCP/IP was its use of small, independent "packets" of data to transmit communications, thereby avoiding the inefficiencies of the prior "circuit switched" world of traditional communications networks.
Prior to the Internet, most electronic communications — including ordinary telephone calls — occurred in a "circuit switched" environment. [ 3 ] When a telephone call or other communication is initiated in a circuit switched environment, the network creates a dedicated electrical path (or "circuit") to carry the communication. A telephone call, for example, has exclusive use of its electrical circuit until the call terminates. This circuit approach allows for relatively simple devices (such as telephones) to communicate reliably. Because telephone calls do not involve the continuous transmission of information, however, the dedicated use of a circuit is extraordinarily inefficient. Even with computer-to-computer communications, computers seldom transmit information constantly, and thus the communications circuits are almost never fully utilized.
In contrast to this "circuit switched" approach, TCP/IP allows "packet switched" communications. With packet switching, every communication is broken up into small, finite pieces. These pieces are transmitted across a computer network that simultaneously carries packets from numerous other communications. As described by one of the inventors of the TCP/IP protocols, Vint Cerf, [ 4 ] TCP/IP packets can be thought of as individual electronic postcards, and the Internet as a superfast postcard delivery system. When a communication is sent over the Internet (an e-mail, for example), the TCP/IP protocol on the sending computer breaks the message into discrete packets, each of which individually carries "to" and "from" addresses. Thus, a short message may be broken up and written in five parts, on five different postcards. The packets are sent into the network, and the network itself decides how to route each packet to the destination. If along the way a packet encounters congestion in the network, it might be re-routed along a different path. Thus, the various packets in a communication may not travel over the same path, may not arrive at the same time, and may not arrive in the same order as they were sent. The TCP/IP protocols on the receiving computer reassemble the packets into the correct order and request that any missing packets be sent again.
This packet switching allows for the operation of a reasonably efficient and reliable network linking many computers. The network as a whole can carry far more communications simultaneously than would be possible if dedicated circuits were required each time a computer communicated with another computer. A downside, however, is that if usage exceeds the network's optimum capacity, all packets (and thus all communications) might slow down.
Leading Internet companies, including the major telecommunications companies AT&T, WorldCom, and Sprint, and more Internet-focused companies such as PSINet, now operate nationwide or worldwide backbone networks to carry TCP/IP packets. Thus, if an AT&T customer in San Francisco seeks to access the web site of an Atlanta company that also happens to be an AT&T customer, the San Francisco consumer's TCP/IP packets would be carried by AT&T to Atlanta, and the Atlanta company's packets would be sent back to San Francisco. If the Atlanta company happens to be a PSINet customer, however, then AT&T would pass the initial request to PSINet for delivery to Atlanta, and PSINet would pass the response back to AT&T for delivery to the AT&T customer. Dozens of overlapping Internet backbone networks criss-cross the country and the world. Those networks interconnect so that packets originating anywhere on the Internet can reach computers anywhere else on the Internet.
Beyond the major Internet "backbone providers" that operate the nationwide or global interconnected networks that make up the core of the Internet, thousands of "Internet Service Providers" ("ISPs") provide Internet service directly to individuals and businesses. Although backbone providers typically also operate as ISPs to provide service directly to customers, many Internet users obtain their service from ISPs that do not operate a backbone. Alternatively, many users obtain service from an "Online Service Provider" ("OSP") such as America Online, which operates as an ISP but also offers its own proprietary content and services (which are not made generally available to non-subscribers). The OSPs and ISPs themselves connect to one or more of the backbone providers (or perhaps to a larger ISP that in turn connects to a backbone provider). Thus, a TCP/IP packet might be carried first by the sender's ISP, then by that ISP's backbone provider, then by a second backbone provider, and finally by the recipient's ISP.
Retail level ISPs (and OSPs such as America Online) offer a wide variety of services and service plans to Internet users. Some offer bare-bones access to the Internet; others offer both Internet access and the option to post a "home page" on the World Wide Web. Still others offer specialized packages aimed at business users. The Internet service market is highly competitive, and in fact some companies are offering free Internet service in return for the ability to transmit advertisements to users while they are online. Some ISPs offer specialized packages aimed at particular market niches. Some ISPs, for example, focus on the family market and offer filtered Internet access that excludes certain types of content available on the Internet, such as adult-oriented content. Other ISPs offer packages aimed at people who use the Internet for real-time chatting with other users.
Although many people have high speed connections to the Internet at their place of employment, most individuals and small businesses access the Internet by using a modem and ordinary telephone line to call a dial-up ISP or OSP.
The three dominant uses for the Internet today are (a) "surfing" the World Wide Web to find entertainment, get information, or make purchases, (b) electronic mail, and (c) "chat" and other real time communications. This Paper assumes a basic familiarity with these three types of communications. On the World Wide Web, Internet users can get access to an enormous range of non-commercial and commercial sites providing information, entertainment, political speech, commercial offerings, and online services. Electronic mail, or e-mail, is most typically used for one-to-one communications not unlike a postal letter or postcard, while "chat" or "instant messaging" are technologies that permit immediate (usually typed) communications between people worldwide.
These applications do not require high-speed broadband technology, but Web surfing in particular is dramatically enhanced by fast access speeds. Broadband access to the World Wide Web offers a much more efficient and enjoyable experience, and can deliver far more audio and video content than would be efficient with a narrowband, dial-up connection. Chat over broadband could eventually support two-way "video phone" images, and e-mail over broadband would likely make the sending and receiving of large audio or video messages more reasonable and more common.
From its origins as an academic communications and resource-sharing tool, the Internet has evolved into a unique and unprecedented medium of mass communications that empowers individuals and businesses to reach millions of listeners worldwide. No other form of communications in our history has allowed speakers to reach the entire world, and to do so relatively simply and inexpensively. Moreover, no other means of communication has emerged as a mass medium so quickly, or evolved at the rate of change seen with the Internet over the past decade. The World Wide Web did not exist ten years ago, but is now a critical method of communications for individuals and businesses trying to reach listeners and consumers. While few were even aware of the Internet five years ago, today it is a primary communications tool for tens of millions of people.
One key to its growth and success is that the Internet is widely available and relatively free of restrictions and rules that might limit its growth or the growth of new communications tools that use the Internet. Individuals and businesses are free to innovate and push the capabilities of the communications medium, and individual listeners are free to experiment with, and embrace, the new innovations.
The explosive growth and evolution of the Internet has prompted a revolution in business. Small and large businesses can reach new markets. Entrepreneurs can implement new business models rapidly and test new products without enormous up-front costs. And startup companies can become major e-commerce players almost overnight.
As one federal judge noted in a very early legal decision affecting the Internet, some important characteristics set it apart from other communications medium:
"Four related characteristics of Internet communication have a transcendent importance to [the court's deferential analysis of the Internet]. . . . First, the Internet presents very low barriers to entry. Second, these barriers to entry are identical for both speakers and listeners. Third, as a result of these low barriers, astoundingly diverse content is available on the Internet. Fourth, the Internet provides significant access to all who wish to speak in the medium, and even creates a relative parity among speakers." [ 5 ]
Because of these characteristics, that judge indicated that "[i]t is no exaggeration to conclude that the Internet has achieved, and continues to achieve, the most participatory marketplace of mass speech that this country — and indeed the world — has yet seen." [ 6 ] In determining that the Internet’s content deserved a very high level of constitutional protection, the United States Supreme Court recognized the Internet’s unprecedented reach:
"From the publishers' point of view, it constitutes a vast platform from which to address and hear from a worldwide audience of millions of readers, viewers, researchers, and buyers. Any person or organization with a computer connected to the Internet can "publish" information. Publishers include government agencies, educational institutions, commercial entities, advocacy groups, and individuals." [ 7 ]
The empowering and democratizing aspects of the Internet are direct — though perhaps neither fully anticipated nor intended — consequences of the Internet's open, and decentralized architecture. The open ability for speakers and listeners to gain access to the Internet, and the decentralized and uncontrolled nature of speech over the Internet, have been essential to realizing the Internet's promise to promote free expression, spur economic growth, create new forms of community online, and reinvigorate democracy.
Emerging broadband technologies could change the essential open nature of the Internet. One of the challenges facing policymakers today is the need to anticipate the possible impact of broadband technology on the Internet, and to ensure that the empowering aspects of the Internet seen in the past ten years can be carried over into the Internet of the twenty-first century.
To assess the potential impact of broadband technology on the nature of the Internet, it is helpful to identify the characteristics of today's Internet that promote user empowerment, democracy, technological innovation, and broad deployment of Internet access. A wide range of discrete characteristics of the narrowband Internet have combined to create an extraordinarily open and dynamic communications medium, and have fostered the environment of democratic expression and technological innovation that is the hallmark of the narrowband Internet. These vital characteristics can be roughly divided into three groupings:
Freedom to listen, speak, and be heard:
Freedom to innovate and experiment:
Freedom to choose from among numerous Internet Service Providers:
Whether or not each one of these characteristics of the narrowband Internet is essential, standing alone, it is clear that taken together, these characteristics enable the Internet to be the dynamic and open medium it is.
As long as there have been national political debates and controversies about the Internet, many policymakers and debate participants have asserted that the Internet is not regulated. Advocates both in the United States and around the world have argued against regulation of the Internet, contending that the Internet has not been regulated in the past, and that governments should not start now.
The assertion that the Internet is not regulated is certainly true, but only in part. The fact that most technological and structural aspects of the Internet have not been regulated has been vital to its explosive growth. On the other hand, the vast majority of the users of the Internet access the Internet every day using a system that historically has been highly regulated — the telephone system. Just as the growth of the Internet depended on a lack of regulation in many areas, the growth of the Internet also depended on the regulation of the telephone system. Indeed, one of the most critical developments that permitted the Internet to come into existence — the right of businesses and consumers to use modems over telephone lines — was the imposition of a regulatory burden on the owners of the telephone network.
In reality, as detailed below, regulation has played a vital role in encouraging and permitting the Internet to develop, and a full understanding of that role is important for any policymaker confronting issues relating to broadband access to the Internet. But because affirmative regulation of the means of access has been vital to the growth of the Internet does not necessarily mean that regulation is the answer to any current or future policy questions. Regulation unquestionably imposes significant costs, and the potential benefits of regulation must always be weighed against its costs.
It is difficult to construct a scheme of categories that effectively separates the aspects of the Internet that have not been regulated from those aspects that have. A "content-versus-facilities" distinction fails because although Internet content has generally not been regulated, neither have important parts of the physical facilities that make up the Internet. A distinction based on physical location within the network is flawed because unregulated elements of the Internet can be connected using regulated elements of the telephone system. The most successful distinction is one between "old" or traditional multi-purpose infrastructure and modes of communications on the one hand, and "new" Internet-specific infrastructure and modes of communication on the other hand. The following sections briefly trace the history of regulation affecting the Internet, and discuss in greater detail the distinction between multi-purpose and Internet-specific infrastructure.
According to FCC Chairman William Kennard, "[f]or the past 30 years, the FCC has created a deregulatory environment in which the Internet could flourish." [ 8 ] Similarly, upon the release of an analysis of FCC's regulatory interaction with the Internet, the FCC emphasized that the "Internet Prospers with 'Hands-Off Unregulation.'" [ 9 ] While these claims are justified — the FCC certainly contributed to the Internet by allowing it to evolve without interference —they gloss over an important part of the story. Even as the FCC took steps to deregulate or avoid regulation of the advanced data services that led to the Internet, it was taking affirmative regulatory steps that allowed those data services to flourish and to protect Internet users' competitive and flexible access to the Internet.
In 1966, the FCC initiated a regulatory inquiry that would set the stage for the later evolution and dynamic expansion of the Internet. Responding to the then emerging reality that computer technology would increasingly be used to communicate and transmit data over ordinary telephone lines, the FCC proactively raised the questions of whether those data services should be regulated, and whether and how the owners of the underlying telephone facilities should be permitted to compete in the data services market. [ 10 ] The decisions reached in what is now called its First Computer Inquiry proceeding would have a dramatic impact on whether the Internet would be burdened with regulation.
The first question that the FCC addressed was whether computer data service companies should be subject to common carrier regulations in the same way that the telephone companies, on whose telephone lines the data was communicated, were regulated. After concluding that there were no significant barriers to entry into the data services market, the FCC decided that data services should not be regulated as common carriers. [ 11 ] Fifteen years later, as the Internet was commercialized in late 1980's and early 1990's, Internet Service Providers benefited from the FCC's decision to forego regulation of data services.
The second question that the FCC answered in its First Computer Inquiry proceeding was whether telephone companies (primarily meaning, at that time, the American Telephone & Telegraph company before its breakup into AT&T and seven "local exchange carriers") should be permitted to offer data services in competition with data service companies that did not own telephone facilities. The FCC was concerned that the telephone companies might discriminate against unaffiliated data service providers in the provision of the telephone service that was essential for data companies to reach their customers. The FCC first concluded that common carriers should be permitted to compete in the data services market, but that (a) they must offer data services through a separate affiliate, and (b) they may not discriminate in favor of their data services affiliates. [ 12 ]
Thus, in its First Computer Inquiry, the FCC laid the groundwork for the explosion of thousands of ISPs competing to offer better, faster, cheaper Internet service. The FCC made the deregulatory, or unregulatory, decision that data service companies (and thus ISPs) should not be regulated. The FCC then took the affirmatively regulatory step of forcing the facility-owning companies to (a) segregate their data businesses and (b) provide vital telephone services to their data service competitors without favoritism or discrimination.
At the same basic time that the FCC initiated its First Computer Inquiry, it also took another regulatory action that — much later — directly enabled the Internet to become the mass medium it is today. In 1968, the FCC issued its Carterfone decision. That decision struck down a tariff, levied by AT&T, prohibiting customers from connecting their own telephone equipment to AT&T's network. [ 13 ] Rejecting claims by AT&T that third party equipment and services would harm the network, the FCC determined that customers could connect their own equipment to the network — so long as the equipment did not in fact harm the functioning of the network. In the mid-1970's, the FCC adopted specific rules that defined for the first time the exact standards that third-party equipment had to meet if it was to be connected to the telephone network. [ 14 ] The Carterfone decision opened the door to the development and improvement of a piece of equipment vital to today's Internet — the modem used to transmit data signals over ordinary telephone lines.
Taken together, the First Computer Inquiry and Carterfone decisions of the 1960's were significant and affirmative regulatory steps that permitted the creation and ultimate explosion of the Internet. The net effect of those decisions was to require the facility owner to carry the data services of competing service providers at published, tariffed rates, without discrimination in favor of an affiliated data service provider. Although the FCC determined in those proceedings that the data services industry itself should not be regulated, the Commission took strong steps to ensure that the new industry would have reasonable and non-discriminatory access to the telecommunications facilities necessary to provide the data services. At the same time as the FCC set the emerging data industry free to compete and evolve, the FCC ensured that the incumbent facility owner could not stand in the way of that competition.
In a follow-up proceeding to the First Computer Inquiry, the FCC created the regulatory categories of "basic" and "enhanced" services. Basic services are the ordinary telephone services subject to common carrier regulation, while enhanced services are data services (including Internet service) not subject to regulation. [ 15 ] The FCC expressly defined "enhanced services" to be data transmission and computer processing services that are offered over basic common carrier facilities. In other words, the FCC plainly recognized that the data services that it sought to promote and encourage through an "unregulatory" approach nevertheless directly relied on the existence and openness of the underlying common carrier facilities. If the "basic" services were not available on a reasonable and nondiscriminatory basis, the "enhanced" services would not have taken off as they did.
The "basic" and "enhanced" service terminology, however, does not fully encompass all aspects of Internet service and communications, and thus is not wholly satisfying as a framework with which to analyze the Internet. [ 16 ] Because so much of Internet service and communications is not regulated, it may be more productive simply to identify the basic elements of a typical communication over the Internet, along with the regulations that apply to each element.
The following are the essential steps that occur when a hypothetical user accesses a major national company's site on the World Wide Web: [ 17 ]
|
REGULATION |
NO REGULATION |
|
|
1. The user types a "URL" (or web address) into a web browser on a home computer |
Other than standards for electrical emissions from the equipment, there are no significant regulations imposed on the user's choice of computer, modem, or software. |
|
|
2. The user's computer initiates a local telephone call to the user's local ISP and transmits the web request to the ISP. |
Common carrier regulation requiring reasonable and nondiscriminatory rates for the "last mile" connection to the user. Both the user and the ISP can obtain local phone service essentially on demand. |
|
|
3. The local ISP routes the web request within its internal network (if any). |
No significant regulations. |
|
|
4. Typically using a dedicated line leased from the incumbent local phone company, the local ISP transmits the web request to a regional ISP. |
Common carrier regulation requiring reasonable and nondiscriminatory rates. |
|
|
5. The regional ISP routes the request within its internal network (if any). |
No significant regulations. |
|
|
6. Using a dedicated line leased from the incumbent local phone company or a competitive provider, the regional ISP transmits the web request to a national Internet backbone provider. |
Common carrier regulation requiring reasonable and nondiscriminatory rates. |
|
|
7. The regional ISP's backbone provider routes the web across the its national backbone network. |
No significant regulations. |
|
|
8. The backbone provider routes the web request to the backbone provider of the requested company. The transfer occurs at a public or private "peering" point pursuant to contract. |
No significant regulations. |
|
|
9. Typically using a dedicated line leased from the incumbent local phone company or a competitive provider, the second backbone provider transmits the web request to the company's web server on company premises. |
Common carrier regulation requiring reasonable and nondiscriminatory rates. |
|
|
10. The company's web server receives and responds to the web request (and transmits the requested web page back to the user using a similar sequence of steps). |
No significant regulations, either on the methods of operating a web server, or on the content of the web pages served by the company. |
As the above chart indicates, elements of communication that are specific to the Internet are not regulated, but elements that travel over the facilities of the incumbent local telephone company are. The ILEC must provide service to both the user and the user's ISP without discrimination.
The conclusion that the Internet has benefited from both regulation and a lack of regulation does not by itself suggest resolution of any policy issue. It does suggest, however, that policymakers should be aware that the history of regulation and the Internet is subtler than is often suggested by policy advocates on all sides when issues surrounding broadband technology are discussed.
Technologies supporting high-speed transmission of data have been available for years, but generally only at a substantial price. Widely affordable broadband services are currently emerging, and emerging with them are mass interest in and market for those services. The new market for broadband services is an outgrowth of technological improvements, increased competition, and increased acceptance of and interest in the Internet. To assess the policy issues raised by broadband technologies requires a basic understanding of these technologies, a discussion of their evolution, and a look at their strengths and weaknesses. This section presents the basics of broadband as concisely as possible. [ 18 ]
Today, most home consumers and many business users access the Internet using a modem and ordinary telephone line. The maximum speed of such a connection is 56 kbps (or 56 kilobits per second, or 56 thousand bits per second). This is a relatively slow speed, and is often called a "narrowband" connection (although it is in fact dramatically faster than modems could connect only ten years ago). Because of physical limitations in modem technology and the copper wire over which most telephone lines run, there is little prospect that the speed of traditional modems will see further significant increase. In contrast, a common broadband speed is 1.5 Mbps (or 1.5 megabits per second, or 1.5 million bits per second), or about 30 times faster than the 56kbps speed of an ordinary dial-up modem.
Some of these terms warrant further definition. A "bit" is the smallest piece of information stored by a typical computer. The term is short for "binary digit," which is a piece of information that can have only one of two possible states (e.g., 0 or 1, yes or no, on or off). Numerically, a bit is represented by either a zero or a one. A byte (pronounced "bite") is typically made up of 8 bits, and can be encoded to represent a single alphabetic character. Thus, in a common implementation of data storage and transmission, a word that is five letters long would take up 5 bytes of storage (which is the same as 40 bits of storage).
The speed of modems is commonly measured in terms of "bits per second" ("bps") — how many bits can be transmitted over a telephone wire every second. The earliest modems (from the early 1960's) could transmit at a maximum speed of 300 bps, the rough equivalent of 37 characters of simple text per second. Thus, at 300 bps, it might require a modem 80 seconds to transmit a single page of typed text. By 1980, common modem speed had risen to 2400 bps, cutting the time to transmit the page of text to 10 seconds. [ 19 ]
By the early 1990's, modem speed had risen enough that the common measurement was no longer "bits per second," but instead was "kilobits per second" ("kbps") representing one thousand bits per second. Common modem speed by 1993 was 28.8 kbps, which is the same as 28,800 bps. At that speed, the hypothetical page of text would take less than a second to transmit. The standard today has about doubled to 56 kbps, or 56,000 bps. At that rate, the page of text takes less than half a second to transmit. [ 20 ]
With the rise of advanced word processors and pictures on the World Wide Web, however, one is seldom transmitting only a single page of simple text. For example, a sample Microsoft Word document with 12 pages of double-spaced text may be about 50,000 bytes (or about 400,000 bits) of data. A large, high-quality photograph might have more than 1 million bytes (or more than 8 million bits). Transmitting such a file at 56,000 bps would take over two minutes.
Files of this size and even more demanding full motion video strain the capacity of a 56 kbps motion (and the patience of the computer user). Broadband technologies can speed things up dramatically. A common broadband speed is 1.5 megabits per second (or 1.5 million bits per second, or 1.5 Mbps). At that speed, the large photograph could be transmitted in less than five seconds, compared to more than two minutes over a conventional modem. With compression techniques that can reduce the size of an image file without too much loss of image quality, the photograph could be transmitted over a broadband connection in a second or two.
At its simplest, "broadband" is simply higher-speed transmission of data. There is not, however, a generally accepted level of speed that everyone agrees qualifies as "broadband." The FCC has chosen a relatively low threshold speed of 200 kbps for a service to qualify as "broadband." [ 21 ] This speed is about four times faster than a 56 kbps modem, but significantly slower than some available broadband speeds that exceed 1 Mbps. The FCC selected this speed because it is fast enough "to change web pages as fast as one can flip through the pages of a book and to transmit full-motion video." [ 22 ]
The FCC also decided that to be considered a "broadband" service, the service had to support 200 kbps in both directions — both "downstream" from the Internet to the user, and "upstream" from the user back to the Internet. Although a significant upstream capability can be useful, the 200 kbps threshold set by the FCC is probably higher than necessary for many users. Internet users who spend most of their time (a) surfing the World Wide Web, (b) "chatting" with other Internet users, or (c) sending e-mail, probably would not consistently need or utilize as much as 200 kbps in upstream bandwidth.
The two broadband technologies already available to consumers on a large scale fit the FCC's definition of broadband. Cable modem service generally provides throughput speeds of at least 1.0 Mbps, sometimes exceeding 3 — 5 Mbps, and occasionally going as high as 10.0 Mbps, depending on conditions. DSL comes in a wide variety of flavors and speeds, but common DSL offerings include downstream speeds of 384 kbps, 640 kbps, 1.5 Mbps, and as much as 7.1 Mbps. However, certain DSL services now available do not provide 200 kbps bandwidth upstream, and thus are not considered full broadband by the FCC.
Faster transmission speed is one major advantage offered by broadband technologies. But equally important are broadband connections’ ability to be "always on" — in other words, the Internet can always be available (assuming a computer is left turned on). With a dial-up modem, a user would have to initiate each connection through the modem, a process that can easily take more than a minute. With broadband services, the Internet is always available.
The significance of broadband connections being "always on" cannot be underestimated. Many broadband users report that this characteristic is more important than the higher transmission speed. With constant access to the Internet, users seek information (such as directory assistance or an up-to-date weather forecast) from the Internet far more often than with dial-up access. [ 23 ]
As the broadband market develops, the concept of "broadband" service will go well beyond simply the faster, always-on Internet connection. Companies that specialize in broadband access (such as @Home, which provides broadband service over cable systems around the country) are developing and offering access to much more video and other high bandwidth content than would be possible over a narrowband connection. Thus, from the consumers' perspective, "broadband" will likely come to mean both high-speed, always-on access, but also content tailored to take advantage of the high-speed capabilities.
In his excellent book-length analysis of the emerging broadband marketplace, modem pioneer Kim Maxwell grouped potential residential broadband applications into three general categories: "professional" (activities related to users' employment), "entertainment" (from game playing to movie watching), and "consumer" (all other non-employment and non-entertainment activities). [ 24 ] With some modifications, Maxwell groups the activities as follows: [ 25 ]
Professional Activities:
Entertainment Activities:
Consumer Activities:
These applications have different bandwidth requirements, and some of them are still out of reach today. For example, all of the "professional" activities will likely be supported with less than 1.0 Mbps of bandwidth. Similarly, web surfing and home shopping will be supported with less than 1.0 Mbps of bandwidth. [ 26 ]
Movies and video, however, demand more bandwidth. Feature length movies can probably be delivered with 1.5 Mbps of bandwidth, but broadcast quality video will probably require more— perhaps as much as 6.0 Mbps. [ 27 ] Moreover, if high definition television ("HDTV") is widely accepted as a new broadcast standard, that quality of video would require almost 20.0 Mbps of bandwidth — much higher than the current broadband technologies will support. Thus, although the technology is moving toward flexible, high-quality video-on-demand, the necessary speed is probably still more than a few years away from becoming a reality.
Modern cable televisions systems had their origin in "community antennae television" ("CATV") systems that served remote communities that could not receive distant broadcast signals. The CATV systems used a large satellite dish to receive television signals, and then distributed the signals to the community using coaxial (or "coax") cables. Cable systems were capable of delivering many more channels than were commonly broadcast even in major cities, and cable soon spread to areas that could receive broadcast signals. Original cable systems had capacity for perhaps 50 channels of analog video, and only supported one-way (downstream) transmission of video signals. Between the cable "headend" (typically where satellite dishes received the signals) and subscribers homes, numerous signal amplifiers were placed to boost the strength of the signal sent on the coax cable.
Starting in the early- to mid-1990's, the cable industry undertook the massive and expensive conversion of its traditional cable facilities into systems that (a) could support two-way transmission of signals (e.g., interactive video and video-on-demand systems), and (b) had a much higher capacity to support analog video signals, digital video signals, and data signals. These upgraded systems are commonly called "hybrid fiber-optic coax" — or "HFC" — cable systems. A fiber-optic cable (which carries signals with light instead of electricity) runs from the cable headend to "nodes" in residential communities. In addition to installing fiber trunks, cable companies also upgraded the coaxial cable running from the nodes to homes. Old amplifiers were removed and new two-way amplifiers were installed. But even with these upgrades, HFC cable systems are sensitive to electrical "noise" and other interference that can interfere with service.
A cable operator would likely upgrade its cable facility to an HFC system for reasons entirely independent of providing Internet services. An upgraded HFC plant can support significantly more cable channels, including both analog and digital channels. And HFC systems can give cable companies a competitive advantage in a market increasingly crowded with recent entries such as TV access through small satellite dishes. [ 28 ]
Once a cable system is upgraded to a two-way HFC system, cable operators can make further upgrades that will enable Internet access. Typically, the cable operator dedicates the bandwidth from a single cable channel for data delivery. The cable operator must then build a data network and install in each cable headend one or more (depending on the number of subscribers) "cable modem termination systems" (or "CMTS"), which directly support the sending and receiving of data over the cable system. In the subscriber’s home, a "cable modem" is installed in or attached to a personal computer, and it communicates with the CMTS back in the headend.
Coaxial cable and HFC cable systems can theoretically support speeds of more than 30 Mbps downstream. Speeds available to each Internet user, however, are significantly less than 30 Mbps for two reasons. First, cable modems themselves are frequently limited to speeds of 10 Mbps or less. Second and more importantly, the bandwidth supported by a CMTS is a shared resource that is shared by all Internet access customers served by a given CMTS. Thus, although cable modem service has significantly higher maximum speeds as compared to most DSL and other broadband services, users seldom in fact receive the top speed.
There is no clear and consistent answer as to the speed of cable modem service. Speed will differ between cable systems, and even within a cable system. One report based on testing at a single computer indicates that the cable modem speed varied from 160kbps to 1.3 Mbps. [ 29 ] Users report that cable modem service is often sluggish during peak afternoon and evening hours. [ 30 ]
A second concern that is more pressing with cable modem service than with other broadband options is security. In effect, a cable modem network is similar to a local area network that one would find in an office, which means that users have some ability to see (and possibly change or delete) files that are on another user's computer. There have been reports of users of cable modem systems being able to view files on a neighbor's computer. [ 31 ] These problems can generally be solved by installing software or hardware screens in each home, and increasingly cable service operators are addressing the issue.
Analysts and commentators note that, as of August 1999, over one million homes in North America subscribe to cable modem service. [ 32 ] At least 300,000 of those homes, however, are in Canada, where cable operators (compared to U.S. operators) (a) are far more complete in their deployment of upgraded HFC cable plants, and (b) have experienced a much higher percentage of installations per homes with access to the service. The Canadian experience may suggest coming trends for the U.S., or it may simply reflect the impact of Canada’s significantly lower prices for cable modem service. Estimates suggest that the residential cable modem deployment in North America will grow to over four million by 2002. [ 33 ]
Cable systems pass (and thus are available to) approximately 90% of the homes in the U.S. The leading U.S. cable operators are forecasting that 90% of their systems will have been appropriately upgraded by the end of the year 2000, an estimate consistent with the experiences of Canadian operators. Although smaller cable operators may have less aggressive deployment goals, the forecasts suggest that by 2001, cable modem service will be an option for as many as 80% of U.S. homes.
Digital Subscriber Line ("DSL") technology was developed in the late 1980’s and was first widely used for non-Internet specific applications. DSL comes in a wide variety of forms, including ADSL, HDSL, IDSL, and SDSL, and is often generically referred to as "xDSL." [ 34 ] Each differing form of DSL has different characteristics and limitations. The form emerging as the primary DSL method for mass-market broadband Internet connections is ADSL, or Asymmetric DSL.
The critical feature of ADSL (as well as all other DSL technology) is that existing copper wires used for traditional telephone service can also be used to provide high speed data services. In fact, with ADSL (though not with most other forms of DSL), the high-speed data connection can be run over the same wire that is used to carry a regular phone line. Thus, ADSL can share an existing wire with a user’s existing voice telephone line, permitting both DSL and telephone services to be used at the same time.
ADSL is called asymmetric because its downstream speed (from the telephone company to the customer) is typically much faster than its upstream speed (back to the telephone company). Most other forms of DSL are symmetric — data travels at the same speed in both directions. ADSL is well suited for Internet access where the user commonly receives more information (from Web sites, etc.) than the user transmits. [ 35 ] ADSL can theoretically support downstream transmission speeds of up to 8 Mbps, but is most commonly available at speeds from 640 Kbps to 1.5 Mbps. [ 36 ]
With ADSL technology (as with most DSL services), there is a critical limitation on the distance of a customer’s site from the telephone central office. With ADSL at speeds of 1.5 Mbps or less, the general distance limitation is about 18,000 feet, or somewhat more than 3 miles. Importantly, the limitation is on the length of the wire itself, and thus a home or business could well be less than 3 miles from a central office, but still be unable to get ADSL service because the wire — the local loop — is not run directly to the home or office.
This 18,000 foot limitation alone precludes over 20% of U.S. households from receiving ADSL service. [ 37 ] The number of candidate households is further somewhat reduced by the fact that DSL service cannot be offered on a line that has certain line conditioning equipment that telephone companies used to extend the capacity of lines.
Unlike cable modem service, ADSL service operates on dedicated wires that are unaffected by high usage by neighbors. Thus, theoretically, ADSL service will provide more reliable — but usually somewhat slower — Internet access than cable modem service. As with the Internet in general, however, there are other potential bottlenecks within a typical ADSL system that could slow down service. Specifically, all ADSL users’ data packets must be transported from the telephone central office to the ISP’s offices (before being sent on to the Internet). [ 38 ] If the data network between the central offices and an ISP is overcrowded, then ADSL service could face overload problems similar to those possible with cable modem service.
There are two primary types of DSL providers: "Incumbent local exchange carriers" ("ILECs"), such as Bell Atlantic, Bell South, and Southwestern Bell, historically have been the primary local telephone company in their regions. In competition with the ILECs are the "competitive local exchange carriers" ("CLECs") such as Covad, NorthPoint, and Rhythms. These and other CLECs essentially owe their existence to the Telecommunications Act of 1996 that promoted local telephone competition by requiring the ILECs to sell discrete elements of the telephone network to competitive carriers.
Although DSL technology was developed by the early 1990’s, the ILECs did not aggressively market DSL until 1999. [ 39 ] Since then, the ILECs’ DSL products have been primarily marketed to residential Internet users. Common ILEC offerings have included 1.5 Mbps, 768 Kbps, or 640 Kbps in downstream (to the customer) data speed, with 90 Kbps or 384 Kbps in upstream (from the customer) speed. The ILECs’ DSL products are typically offered on a "best efforts" basis, meaning that no particular speed is guaranteed. [ 40 ] The ILECs typically price their DSL service in the $50 to $60 per month range. [ 41 ]
CLECs have focused more of their efforts on the business market. [ 42 ] The CLECs commonly have "best efforts" ADSL products that compete directly with the ILECs’ offerings. In addition, CLECs have other xDSL products that typically offer symmetrical, guaranteed-bandwidth service. CLECs typically price their DSL services in the $70 to $90 per month range for their consumer oriented products, and as much as $300 to $400 per month for their business oriented products.
As of the end of 1999, an estimated 500,000 DSL lines were deployed in the United States, with about 75% being ILEC lines and 25% being CLEC lines. [ 43 ] Deployment is projected to explode in the year 2000 (to over 2,000,000 DSL lines deployed), followed by significant growth in later years. [ 44 ] A number of leading ISPs have chosen to offer DSL services to their customers. [ 45 ]
Although significantly less established and deployed than cable and DSL broadband services, terrestrial wireless services using land-based transmitters can also provide broadband services to businesses and individuals. There are a number of different technological approaches to wireless broadband services, and it is far from clear which of these approaches will be successful in the marketplace. Nevertheless, it is likely that at least for some categories of users, broadband wireless will compete with cable and DSL services (as well as with traditional dedicated business-oriented high-speed T1 lines) to provide broadband access to the Internet.
Wireless broadband services are often called "fixed wireless" because the transmitting and receiving stations (of both the service provider and the customers) are in fixed, stationary positions. In other words, terrestrial or fixed wireless services are quite distinct from mobile cellular telephone service. Typically, a company offering fixed or terrestrial wireless services will operate one or more master microwave antennas installed on top of tall buildings or possibly mountains adjacent to populated areas. The users will rely on relatively small antennas on top of their office buildings or homes–some of which, because of their size and shape, are referred to as "pizza box" antennas. In the implementations most likely to compete with cable and DSL services, users’ antennas act as both receivers of downstream Internet data and transmitters of upstream data.
As with DSL technology, transmitting data by microwave or other wireless signals is not a new idea, [ 46 ] but only recently has it been available to broad groups of Internet users. Two critical factors boosted the potential of wireless broadband services. First, the cost of equipment declined significantly over the past ten years. Second, the Federal Communications Commission made more radio spectrum available for used by data service providers. For example, in 1998 the FCC changed its rules to allow certain spectrum that had been used for one-way video transmission to be used instead for two-way data transmission. [ 47 ] This change alone led directly to the commercial feasibility of MMDS wireless service discussed below. [ 48 ]
The two leading types of wireless services that can support broadband access to the Internet are "Local Multipoint Distribution Service" ("LMDS") and "Multichannel Multipoint Distribution Service" ("MMDS"). LMDS and MMDS generally both require a "line of sight" between a central antenna and a customer’s antenna. LMDS provides faster speeds than MMDS, but can only support customers within two or three miles of a central antenna. LMDS is therefore best suited for businesses located in dense urban areas. LMDS is, however, susceptible to interference from rain and snow. Service providers such as Winstar and Teligent are using LMDS to target businesses.
In contrast, MMDS technology cannot support such high speeds, but it can reach customers who are located much farther from an antenna than would be possible using an LMDS system. Technically, MMDS’ zone of coverage could extend 35 miles in every direction from a central tower (covering over 3500 square miles compared to less than 50 square miles covered by a single LMDS station). MMDS can be deployed more cheaply and can reach more efficiently into suburban and rural areas. Both WorldCom and Sprint have invested heavily in MMDS technology.
LMDS, MMDS, and other approaches (such as the Wireless Local Loop being tested by AT&T), however, all still require a line-of-sight, or near-line-of-sight, path between master and customer locations. As a result, buildings, trees, and other impediments can prevent wireless technology from serving many potential customers in an area. Because of this limitation, many leading providers (like AT&T and MCI WorldCom) are likely to use wireless technology to fill in gaps in their DSL or cable coverage. On the whole, wireless technology will probably be an important aspect of widespread broadband availability, but it is unlikely that wireless will ever become a ubiquitous option for all consumers.
Beyond cable, DSL, and terrestrial wireless services, other broadband access methods are still in development and/or testing, and none are certain to become technologically and commercially viable. Many companies–both startups and well-established market participants–are investing substantial money to develop and deploy new forms of broadband access to the Internet. It will likely be 2001 or 2002 before it becomes clearer whether these ventures will succeed.
Satellite Access: Most often grouped with cable, DSL, and wireless, satellites access technology will use satellites to deliver Internet access to homes and businesses.
Today, one can receive high speed Internet service from Hughes Network Systems, but its DirecPC product is a one-way service. DirecPC uses a satellite to deliver content to customers but requires that customers transmit content requests to an ordinary ISP over a standard dial-up phone line. This approach is often called a "telco return" because it relies on a standard telephone connection for any transmissions from Internet users. [ 49 ] Because one-way service requires the use of a phone line and an ISP, and because it lacks the "always on" characteristic of other broadband services, it is not viewed as a long-term broadband solution. DirecPC has announced that it hopes to begin offering two-way satellite Internet access by early 2001. [ 50 ] That service will, however, generally be slower than most cable, DSL, and terrestrial wireless offerings; Hughes anticipates replacing the service by 2003 with an "advanced generation" of service on its "Spaceway satellite platform." [ 51 ]
In addition to Hughes, a number of other companies hope to deliver two-way broadband access services by satellite. For example, iSKY plans to offer 1.5 Mbps (downstream) service via satellite (with 500 Kbps upstream) before the end of 2001. [ 52 ] Teledesic, which announced its plans to offer broadband satellite service in 1994, hopes to be operational by 2004. [ 53 ] These services suffer from the uncertainties of whether the technology will work in practice and whether cable, DSL, and wireless will have won key parts of the market before the satellites become fully operational. [ 54 ] In the wake of the bankruptcy of the Iridium satellite-based telephone system, investors and analysts are skeptical of whether Teledesic (and other satellite Internet services) will succeed. [ 55 ]
Fiber-to-the-home: Another broadband alternative is fiber optic lines running directly into users’ homes and businesses. Although fiber does not share the same technological uncertainties facing satellite systems, it is not clear whether it will be economically viable to run fiber optic wires through existing residential neighborhoods. ILEC BellSouth is the most significant market participant to test "fiber-to-the-home." BellSouth began in 1999 to run fiber to a limited number of homes in Dunwoody, Georgia (an affluent suburb of Atlanta) in a permanent installation of the technology. BellSouth plans to offer both Internet access and video services over the fiber. Futureways is preparing to offer fiber optic services on a large scale in five suburbs of Toronto, Canada. Some U.S. cities, such as Palo Alto, California and Concord, Massachusetts, are experimenting with municipally supported fiber installations. [ 56 ]
Fiber optic cables can support an extraordinarily high bandwidth of Internet traffic — possibly reaching 1000 Mbps, compared to 1.5 Mbps for typical cable and DSL offerings. In practice, fiber installations will offer 100 Mbps to individual users. The downside, however, is the cost of install the fiber optics — even in brand new neighborhoods, fiber costs at least 15 percent more than ordinary copper wire. Still, BellSouth and others appear willing to install fiber in some new areas. In existing neighborhoods, where copper wire has already been laid, however, the cost of running fiber into homes may well be prohibitive in the near term. [ 57 ]
To date, no company has announced any plans to install fiber-to-the-home into existing residential neighborhoods on any wide scale. Although fiber-to-the-home may be economical in the future, it is not presently a clearly viable alternative.
Broadband over power lines: Both cable modem and DSL offerings seek to capitalize on existing "pipes" or wires, into the home. The only remaining wire into most homes is the electric utility wire. A number of schemes to use power lines to transport Internet access services have been explored over the past few years, but none have yielded a marketable service offering. The leading test of powerline service offerings encountered significant technical difficulties and ultimately was abandoned. [ 58 ]
Two companies, Oneline AG in Europe and Media Fusion in Texas, have announced that they are currently testing or developing powerline based products. [ 59 ] If the services work as hoped, utility companies could well become major players in the broadband access market. It is far from clear, however, when, or if, powerline services will become feasible.
Policy debates relating to broadband technology have raged in Congress, before the FCC, and in the press. These arguments, however, have often led to confusion rather than clarity. This section reviews and summarizes the policy issues raised.
The most prominent policy questions have involved the cable industry and efforts by ISPs and others to offer services over cable systems. Other questions relate to the provision of DSL services, while still others are more general in focus. The issues discussed below will be roughly grouped into these categories. The most prominent issues — involving the access to cable issue — are discussed in greater detail in Part IV.
Resolution of the cable- or DSL-specific issues, however, could have direct implications in other technology areas. For example, how policymakers resolve the "access to cable" issues could affect the development and deployment of wireless technology. Thus, these policy decisions should be made with a broad perspective over the entire Internet.
The key broadband issues that specifically relate to Internet access over cable systems are rooted in the broad policy question of whether cable-based Internet access should be treated similar to cable television service (which has been largely unregulated) or similar to telephone service (which has been subject to significant regulation). Some cable companies have argued that additional regulations should not be imposed on the cable industry, even if it is providing Internet access. Some "open access" proponents have argued that the regulations applicable to the telephone system has been vital to the growth of the Internet, and similar regulations should be imposed on Internet access over cable. Finally, some local telephone service providers have argued that the incongruity between these two differing levels of regulation should be resolved by reducing regulations applicable to telephone service, rather than imposing regulation on cable-based Internet service.
A logical starting point to any analysis of regulatory action is a look at where cable-delivered Internet service fits into the current regulatory scheme. The key question is whether Internet-over-cable is a "cable service" (and thus subject to minimal regulation) or a "telecommunications" or "information" service (subject to different regulatory schemes). The FCC has to date avoided deciding the issue. The United States Court of Appeals for the Ninth Circuit, however, recently considered the issue and concluded that cable-delivered Internet service is a "telecommunications service." This appears to have pushed the FCC to consider the issue directly. Although a resolution of this question may not be essential to a resolution to the "third party access" issue discussed below, any definitive resolution of this question would have a significant impact on the third-party access debate. This issue is discussed in greater detail in Part IV below.
Cable television companies have made significant investments to upgrade their cable systems to support two-way high-speed Internet access, and those companies have contended that they should be able to contract with a single ISP to provide such service to all of their cable customers (although at least two leading cable companies have subsequently shifted their positions, as discussed below). Many Internet Service Providers, initially led by America Online, have contended that owners of cable systems should be required to permit unaffiliated, third party ISPs to offer high-speed Internet service to the cable customers. Many consumer and public interest advocacy groups have actively argued in favor of requiring cable companies to allow third party access. This issue is discussed in greater detail in Part IV below.
A small ISP, Internet Ventures, Inc. ("IVI"), among others, has contended that it should have a right under the "leased access" provisions of the 1984 Cable Act to obtain access to one or more channels on a cable system for the purpose of transmitting high-speed Internet content to customers. The cable system owners contend that under the 1984 Act, the "leased access" channels can only be used by "video programmers" and IVI does not fit within either the language or statutory intent of the Act.
IVI’s approach differs significantly from that of many other ISP’s seeking to provide cable-based Internet access. While AOL and other ISP's have worked to gain access to upgraded, two-way capable cable systems, IVI is pursuing a different approach to Internet access over cable systems. IVI proposes to provide high-speed downstream Internet content, with the upstream side completed using a standard phone line and modem (called a "telco return" approach). Although not as fast, efficient, or desirable as the regular two-way cable modem service offered by upgraded cable systems, IVI's approach can — in theory — be offered over any cable system, whether or not it has been upgraded.
IVI's business model was to lease a single cable channel from a cable system, and to place downstream data intended for IVI's customers onto that single channel. IVI contended that it should be able to lease the channel under the 1984 Cable Act, which requires that cable companies make available "leased access" channels to promote competition and diversity in leased access programming. All major cable companies approached by IVI refused to lease IVI a cable channel, on the asserted bases that IVI did not meet the definition of "video programmer" under the 1984 Cable Act.
IVI filed a petition for declaratory ruling with the FCC asking that it order cable systems to permit IVI to lease a channel for downstream Internet content. After receiving comments from the public, the FCC on February 18, 2000, denied IVI’s petition, concluding that the leased access channels were only available to "video programmers" and that a Internet service provider (such as IVI) did not meet the statutory definition of "video programmer." [ 60 ] IVI does not appear to be pursuing its petition further.
The broadband issues relating to Internet access over DSL systems generally involve the ability of "competing local exchange carriers" ("CLECs") to compete with "incumbent local exchange carriers ("ILECs"). Under the Telecommunications Act of 1996, ILECs are required to make certain pieces, or "elements," of their network available to CLECs to enable CLECs to compete with the ILECs in the provision of local phone service and Internet access service. The policy questions center around whether the ILECs are fully complying with their statutory and regulatory obligations to act properly towards, and compete fairly with, the CLECs.
The first four of the DSL-specific issues set out below reflect CLEC challenges to ILEC actions, and are already the subject of current legal or regulatory proceedings at the state and/or federal levels. As a general matter, the CLECs are not seeking major changes in laws or regulations, but instead are seeking stronger and faster enforcement of existing regulations and statutes.
The fifth issue is not a complaint of the CLECs against the ILECs, but instead reflects the arguments made by ILECs that they should not be regulated more heavily than are cable system operators that provide Internet access. Under proposals that ILECs have supported, certain requirements under which CLECs have been able to offer DSL service would be modified or eliminated.
The competitive and at times hostile nature that has characterized most ILEC-CLEC relations over the past few years may be changing. In recent months, many CLECs have struck deals with ILECs for line sharing, as discussed below. More fundamentally, ILEC Verizon (formerly known as Bell Atlantic) and CLEC NorthPoint Communications have announced an agreement to merge their DSL operations into one nationwide venture. [ 61 ] Similarly, ILEC SBC and CLEC Covad have announced a strategic alliance to sell DSL services nationwide. [ 62 ] Both ILECs and CLECs are increasingly focused on competing in the broadband arena with cable modem providers.
CLECs have generally contended that the ILECs are failing to comply with their statutory and regulatory obligations to provide "collocation space" — space inside the ILECs' central switching offices — in which CLECs need to install switching and other equipment necessary to provide their services. CLECs contend that ILECs (a) incorrectly claim that there is no available collocation space, (b) inappropriately require CLECs to obtain large collocation spaces just to install a single small piece of equipment, or require them to install unnecessary equipment or facilities in the collocation space, and (c) discriminate against the CLECs in terms of price of, or access to, the collocation space. The ILECs generally contend that they are fully complying with their statutory obligations, and that certain of the physical requirements are essential to the safe and reliable operation of the telephone network.
In March 1999, the FCC reviewed a number of complaints raised by the CLECs, and issued an order addressing many of the issues. [ 63 ] In that order, the FCC required (among other things) that (a) ILECs permit CLECs to collocate without leasing a large space or installing a wire cage separating the ILEC work space from that of the CLEC, (b) ILECs make additional space available to CLECs even if the primary collocation space is full, and (c) ILECs allow CLECs to tour an entire central switching office if the ILEC claims that there is no space available in the office. On March 17, 2000, the U.S. Court of Appeals for the D.C. Circuit upheld the FCC’s order on all of these points. [ 64 ] In light of the FCC order and the appeals court decision favoring the CLECs, there is not at this time a strong call for federal legislation or additional FCC action to address collocation issues.
CLECs have contended that ILECs are failing to comply with their statutory and regulatory obligations to provide to CLECs fair, efficient, and nondiscriminatory access to "local loops," the physical copper wires that are used to provide telephone service (including DSL service) into homes and businesses.
To provide DSL service to a customer, a CLEC requires that an ILEC provision a "local loop" that runs between the CLEC’s equipment collocated in the ILEC’s "central office" and the customer’s home or business. CLECs have complained that ILECs have been slow and unreliable in provisioning loops, and have asserted that the ILECs provision loops for their affiliated ISPs far more quickly than they do for unaffiliated CLECs. The time required for an installation of a DSL line from a CLEC is generally at least a week or two longer than for the provisioning of DSL service by the ILEC. Part of the time difference may be explainable by the fact that ILECs typically run DSL service over local loops that already carry normal phone lines (and are therefore relatively easy to identify), while CLECs until recently were only able to provide service over previously unused wires into customers’ homes. Nevertheless, the CLECs believe that the ILECs’ slowness in provisioning the loops is unfair and discriminatory.
More generally, CLECs and independent ISPs have complained that ILECs have engaged in anticompetitive practices that include (a) ILECs pricing their retail DSL service at about the same prices they charge small CLECs and ISPs for use of a local loop (ensuring that the CLEC and/or ISP cannot underprice the ILEC), (b) ILECs giving their affiliated ISPs preferential treatment over unaffiliated ISPs, and (c) ILECs using their installation personnel to tout the price advantages of the ILEC ISP over unaffiliated ISPs. [ 65 ]
The ILECs have denied these allegations. These and other issues are being considered in on-going proceedings before the FCC and the courts, and there is not at this time a strong call for federal legislation or additional FCC action to address the provisioning of local loops.
On petition by CLECs and others, the Federal Communications Commission ("FCC") has declared that ILECs must allow CLECs to provide DSL service to a particular customer over the same wire that the ILEC provides local phone service to the customer. Ordinary telephone wire can carry electronic signals over a broad range of frequencies. Traditional analog telephone calls are carried on a low portion of the wire's frequency spectrum, while digital DSL service is carried on a much higher portion of the electronic spectrum. Thus, voice phone service and digital DSL service can, technologically, share a single copper wire without interfering with each other. CLECs sought a requirement that ILECs allow CLECs to use the high frequency portion of local loops while the ILECs continue to provide voice service over the low frequency portions of the same loops. ILECs objected, and contended that they are obligated only to provide entire wires to CLECs (which cost more and are slower to install than the shared spectrum of an existing phone wire).
In November 1999, the FCC agreed with the CLECs and ordered the ILECs to share their existing telephone lines with CLECs seeking to provide DSL service. Since that order, a variety of ILECs have entered into contractual arrangements with CLECs to share lines. In an action requested by the Minnesota Public Utilities Commission, US West contracted with the CLEC Covad to share lines. Similarly, US West and Rhythms NetConnection also contracted to share lines, and thereby dramatically reduced Rhythms’ costs. [ 66 ] In light of these and other agreements between ILECs and CLECs, there is not at this time a call for federal legislation or additional FCC action to address line-sharing issues.
CLECs have contended that the most effective way to ensure that ILECs do not discriminate against CLECs (and in favor of their own Internet service providers) is to require "structural separation" between the ownership of the facilities and the ISP operations. The ILECs assert that such separation is unwarranted and is outside of the FCC’s powers.
In the context of two separate merger reviews, the FCC required (as a condition for merger approval) SBC and Ameritech, and Bell Atlantic and GTE, to separate their ISPs into separate business units to ensure that CLECs can compete on a level playing field with the ILECs' ISP. [ 67 ] At the state level, the Pennsylvania Public Utilities Commission ruled that Bell Atlantic must separate its wholesale DSL and retail ISP businesses. [ 68 ]
In light of these developments, there is not at this time a strong call for federal legislation or additional FCC action to require structural separation.
There are legislative proposals that could directly affect the relationship between CLECs and ILECs with regard to DSL deployment. Legislation proposed in the U.S. House of Representatives would relieve ILECs of their obligation to provide to CLECs access to the network elements necessary for the CLECs to offer DSL service. [ 69 ] The theory behind this legislation is that (a) the cable companies do not face similar requirements as ILECs do, (b) the two industries should be treated similarly, and (c) regulation on telephone companies should be reduced (as opposed to increasing regulation on cable companies). CLECs vigorously oppose calls for "regulatory parity," since such parity would relieve ILECs of their obligation to provide the network elements that CLECs currently receive under the Telecommunications Act of 1996.
Most of the policy debates related to broadband access have focused on the "last mile" of the connection between a provider and a customer's home or business. Yet an issue of equal, or even greater concern is that of the "Digital Divide." A significant — and possibly growing — disparity exists between the telecommunications services available to affluent city and suburban dwellers, and the services within physical and financial reach of inner city and rural communities. There is a great risk that the "divide" will grow even larger as the country moves toward broadband technology. Although there is general consensus that rural and inner-city communities should have access to broadband technologies, there is little consensus on how best to accomplish the goal.
The Digital Divide itself is the subject of numerous white papers and briefing books, and this paper does not attempt to address the issue in its entirety. Two points below, however, relate to the Digital Divide and arguably directly affect broadband deployment.
One way in which the Digital Divide issue arguably affects broadband deployment concerns statutory limitations on ILECs’ ability to carry data across local service boundaries. In an effort to encourage competition in local telephone markets, the Telecommunications Act of 1996 prohibited ILECs from carrying long-distance telephone traffic. These provisions have been interpreted as prohibiting the ILECs from operating an Internet backbone (or any other kind of data network) that would carry data across local calling boundaries known as LATA ("Local Access and Transport Area") boundaries. The ILECs have argued that this limitation constrains their ability to bring broadband Internet services to rural communities. The ILECs contend that with few connections to the Internet backbone available in rural communities, an Internet backbone in local markets could significantly facilitate broadband access in rural areas. This, in theory, would lead to greater broadband options for retail consumers.
Competing Internet backbone providers contend that there is already sufficient backbone capacity to rural communities, and that allowing the ILECs to carry data long-distance would delay local phone competition. Viewed solely from the perspective of Internet development, the restriction on the ILECs’ ability to operate backbones may well slow the introduction of broadband to rural areas. The tradeoffs, however, in terms of local telephone competition, could also be significant. [ 70 ]
The exemption of small incumbent telephone companies from requirements applicable to larger ILECs raises another significant issue. Under the Telecommunications Act of 1996, the requirements that ILECs make elements of their network available to CLECs do not apply to small incumbent telephone companies, almost all of which operate in rural communities. Without that statutory requirement, CLECs cannot force many local carriers in rural areas to allow the CLECs to install equipment and offer service (as CLECs have been able to do with larger ILECs). Certain CLECs contend that broadband access will reach rural communities faster if CLECs are given that power, while the rural incumbent telephone companies contend that they will provide broadband access more quickly if they do not have to open their networks to CLECs. There are no pending legislative or regulatory proposals to address this issue.
In its current form, the Internet is an web of interconnected, overlapping networks designed to route data traffic efficiently and reliably to reach any destination on the vast web. A vitally important result of the Internet's infrastructure is that any speaker on the Internet can reach any listener. On the Internet, a lone speaker and the largest media company have roughly the same abilities to speak and be heard. This rough equality is directly threatened by changes to the infrastructure driven by broadband technologies. There is significant risk on a broadband Internet, where locally-based broadband content servers deliver broadband content quickly to consumers, the major means of broadband distribution will be the proprietary domains of large companies or wealthy speakers. Broadband technology and the content distribution technology that it has spawned could ultimately destroy the rough equality among speakers worldwide so vital to the Internet's promotion and facilitation of democracy.
This issue is extensively discussed in a separate paper issued by the Center for Democracy & Technology and its Broadband Access Project, entitled "The Broadband Internet: The End of the Equal Voice?" [ 71 ]
Over the past two years, the broadband issue that has garnered the greatest attention by companies, lobbyists, the media, and policy makers has been the "third party" access question (also referred to as "open access" or "forced access"). Most cable television systems that have upgraded their facilities to support transmission of two-way data have entered into exclusive contracts with one preferred Internet Service Provider ("ISP"–such as @Home or Roadrunner) to provide Internet service. In most cases, the cable systems have contracted with an ISP in which the cable system owns some stake. The most immediate effect of these exclusive arrangements is that cable customers have no choice but to obtain Internet service from the preferred ISP.
Some leading ISPs and public interest advocates have vigorously argued before the FCC, in Congress, and to state and local governments in favor of what they term "open access." They advocate statutory or regulatory requirements that cable systems support multiple ISPs and permit unaffiliated ISPs to compete for the Internet business of cable subscribers. Cable interests vigorously oppose what they call "forced access" to cable systems, and offer rebuttal to the arguments of the "open access" camp.
Because both "open access" and "forced access" are weighted terms, this paper will use "third party access" to refer to this debate. Not only is "third party access" a more neutral term, it is also the term used by the industry and regulators in Canada, where the issue has been considered at greater length than in the United States.
This section reviews the arguments and issues that have been at the core of this debate, but it starts with an issue that has not been a primary focus: the question of where Internet access over cable systems fits into the existing regulatory structure.
Although the FCC and others have sought to avoid or at least defer consideration of this issue, a logical starting point in considering third party access is to determine where Internet-service-over-cable fits into the current regulations of cable and telecommunications services.
The FCC derives its regulatory authority from the Communications Act of 1934, 47 U.S.C. §§ 151 et seq., and different types of wire and wireless communications have generally been grouped under the titles of that Act. Telephone and data services provided over telephone facilities are regulated under Title II, which incorporates traditional common carrier rules and a broad requirement that providers of such services not discriminate among users of the services. In contrast, video delivery services provided over cable facilities are regulated under Title VI, which does not include common carrier rules or a broad nondiscrimination provision.
At the core of this policy issue is the identification of the proper classification for Internet service provided over a cable facility. If such service is deemed to be a "cable service," then the relatively limited regulations of Title VI could apply. If, on the other hand, Internet access is deemed to be a "telecommunications service" or "information service," then relevant provisions of Title II would more likely apply. If Title II is deemed to apply, the FCC would likely be forced to reconcile or at least explain the significantly different levels of regulation imposed on Internet access over cable systems compared to Internet access over telephone systems.
In August 1999, the FCC filed a "friend of the court" brief with the U.S. Court of Appeal for the Ninth Circuit. In that brief, the Commission provided a useful summary of the arguments about the proper classification of Internet access over cable systems:
Strong arguments have been advanced in support of the argument that Internet access via cable is not a cable service. As an initial matter, . . . not every service offered over cable facilities is a "cable service" under the Communications Act. Congress first defined the term "cable service" as part of the Cable Communications Policy Act of 1984, Pub. L. No. 98-549, 98 Stat. 2779 ("1984 Cable Act"). Initially, Congress defined "cable service" as "(A) the one-way transmission to subscribers of (i) video programming, or (ii) other programming service, and (B) subscriber interaction, if any, which is required for the selection of such video programming or other programming service." 47 U.S.C. § 522(6) (1984). For purposes of this definition, video programming means "programming provided by, or generally considered comparable to programming provided by, a television broadcast station," 47 U.S.C. § 522(20), while "other programming service" means "information that a cable operator makes available to all subscribers generally." 47 U.S.C. § 522(14).
As the legislative history of the 1984 Cable Act makes clear, the original definition of "cable service" did not include the types of services that are now commonly associated with Internet access. The House Report on the legislation emphasized that the Act's definition of "cable service" restricted subscriber interaction to the selection of categories or options provided by the cable operator or the programming service provider: "By contrast, interaction that would enable a particular subscriber to engage in the off-premises creation and retrieval of a category of information would not fall under the definition of cable service." H.R. Rep. No. 934, 98th Cong., 2d Sess. 42-43 (1984) ("House Report"). In other words, under the 1984 Cable Act, "services providing subscribers with the capacity to engage in transactions or to store, transform, forward, manipulate, or otherwise process information or data would not be cable services." Id. at 42. The House Report on the statute went on to identify several services that would not be cable services, including "shop-at-home and bank-at-home services, electronic mail, one-way and two-way transmission [of] non-video data and information not offered to all subscribers, data processing, video conferencing, and all voice communications." Id. at 44.[footnote omitted]
In 1996, Congress amended the statutory definition of cable service. As amended, the Communications Act defines cable service as "(A) the one-way transmission to subscribers of (i) video programming, or (ii) other programming service, and (B) subscriber interaction, if any, which is required for the selection or use of such video programming or other programming service." 47 U.S.C. § 522(6) (emphasis added). AT&T and TCI have argued that Congress, by adding the words "or use," intended to expand the definition of "cable service" to include the wide range of interactive services encompassed by Internet access. See TCI Order ¶ 82.
The legislative history of the 1996 Telecommunications Act provides some support for AT&T's and TCI's position. The Conference Report on the statute declares: "The conferees intend the amendment [of the definition of cable service] to reflect the evolution of cable to include interactive services such as game channels and information services made available to subscribers by the cable operator, as well as enhanced services." S. Conf. Rep. No. 230, 104th Cong., 2d Sess. (1996) ("Conference Report") (emphasis added). The Conference Report's references to "information services" and "enhanced services" suggest that the amended definition of "cable service" may include the types of interactive cable broadband services that were previously excluded. The statute defines "information service" as "the offering of a capability for generating, acquiring, storing, transforming, processing, retrieving, utilizing, or making available information via telecommunications." 47 U.S.C. § 153(20). And the FCC's rules define "enhanced services" to include services that "involve subscriber interaction with stored information." 47 C.F.R. § 64.702(a).
On the other hand, notwithstanding the 1996 amendment, one basic aspect of the definition of cable service remains unchanged: A service cannot be a "cable service" unless it qualifies as "video programming" or "other programming service." The 1996 Telecommunications Act did not alter the definitions of "video programming" or "other programming service." Unless Internet access fits one of these definitions, it cannot qualify as "cable service."
Proponents of the view that Internet access is a form of "cable service" generally do not argue that it is a form of "video programming" comparable to that offered by a television broadcast station (see 47 U.S.C. § 522(20)), but instead contend that Internet access fits the definition of "other programming service" -- that is, "information that a cable operator makes available to all subscribers generally." 47 U.S.C. § 522(14). But it is not clear that Internet access meets this description. Arguably, the "information" that an individual subscriber obtains via Internet access -- for example, E-mail or access to a specific web site chosen by the subscriber -- is provided only to that particular subscriber. In that respect, this information may not be made "available to all subscribers generally."
AT&T and TCI appear to argue that a cable operator makes information "available to all subscribers generally" simply by providing subscribers with the capability to gain access to the Internet. Under this broad statutory interpretation, however, "other programming service" would arguably include any transmission capability that enables subscribers to select and receive information, including basic telephone service. And Congress stated that its 1996 amendment of the definition of cable service was not intended to eliminate the longstanding regulatory distinction between telecommunications service and cable service: "This amendment is not intended to affect Federal or State regulation of telecommunications service offered through cable facilities, or to cause dial-up access to information services over telephone lines to be classified as a cable service." Conference Report at 169.
A number of parties have argued that Internet access services "are information services or telecommunications services covered by Title II" of the Communications Act. TCI Order ¶ 83. Currently, when Internet access service is provided over telecommunications facilities, the Commission treats that service as an information service. See Federal-State Joint Board on Universal Service, Report to Congress, 13 FCC Rcd 11501, 11536-40 (¶¶ 73-82) (1998) (Exhibit 5). If the same type of Internet access service is offered over cable systems as well as telephone networks, it is not readily apparent why the classification of the service should vary with the facilities used to provide the service.
Alternatively, Internet access over cable could be classified as the sort of "advanced telecommunications capability" identified by section 706 of the 1996 Telecommunications Act. Functionally, Internet access provided through cable modems is no different from the broadband capability provided over other facilities such as the wireline telephone network, wireless telecommunications systems, or satellite facilities. If the Commission were to treat cable modem service as "advanced telecommunications capability," it would have the opportunity to develop a coherent regulatory policy that took into account the full range of broadband service providers, including cable systems. Local franchising authorities would have no such opportunity because they have no regulatory authority over broadband service providers other than cable systems. Thus, local regulation of a cable system's broadband services as "cable services" might pose a significant risk of regulatory disparity with respect to all other broadband service providers. Any such disparity might undermine the objectives of section 706 by impeding the reasonable and timely deployment of advanced telecommunications capability to all Americans.
More generally, on a conceptual level, an argument can be made that Internet access is more appropriately characterized as an information or telecommunications service rather than a cable service. At the most basic level, there are two kinds of communications service networks: (1) broadcast (one-to-many) networks, in which the distributor chooses the content and sends it to all customers; and (2) switched (one-to-one) networks, in which the customer chooses the content and sends it to the person(s) of his or her choice. The first type of network best describes cable service; the second type of network most accurately depicts telecommunications and information services. Some have argued that Internet access more closely resembles the switched network. However, the Commission has not yet conclusively resolved the issue. [ 72 ]
As the FCC noted, it has not "conclusively resolved" the question of the proper classification of Internet access over cable systems. Indeed, the FCC has specifically avoided the issues, despite being directly asked to determine proper classification in both the merger review concerning the AT&T-TCI merger, and in the "Advanced Services" proceeding pursuant to 47 U.S.C. § 706. But, to date, the agency has adopted a "hands off" approach to third party access, and has declined to address the regulatory classification issue.
As mentioned above, a recent decision of the Ninth Circuit Court of Appeals, however, may force the FCC to address the issue of regulatory classification. In that case (in which the FCC filed the amicus curiae brief quoted above), the Court did directly address the question. The case grew out of the proposed merger of AT&T and TCI. The City of Portland and the County of Multnomah, Oregon, responded to the merger by refusing to grant an AT&T/TCI request for a change in control, unless AT&T agreed to allow third party access on its cable system. AT&T and TCI filed suit asking the federal district court to declare the Portland and Multnomah County actions illegal and unauthorized. The district court ruled against AT&T, and held that the local agencies did have the authority to impose a third party access obligation. AT&T has appealed the case to the U.S. Court of Appeals for the Ninth Circuit, which heard oral argument in November 1999.
The district court's decision focused on whether the actions of the Portland local government were preempted by actions or positions taken by the FCC and the court concluded that Portland's actions were not preempted. During oral argument on appeal, however, the Ninth Circuit judges pressed the parties about the regulatory classification issue discussed above. Before the lower court, both Portland and AT&T had asserted that the Internet service at issue was a "cable service" subject to the federal regulatory regime applicable to cable systems. On appeal, the judges of the Ninth Circuit court expressed skepticism about the assumption — made by both parties — that the "cable service" classification was appropriate.
On June 22, 2000, in a well-reasoned and careful decision, the Ninth Circuit held that Internet service delivered over cable systems was a telecommunications service, not a cable service. [ 73 ] In reaching this decision, the court closely reviewed the statutory definitions of the different types of service, as well as the practical reality of what service a cable modem provider is actually offering. The court’s analysis focused on @Home, the cable modem service provider at issue in the AT&T v. Portland case:
Like other ISPs, @Home consists of two elements: a "pipeline" (cable broadband instead of telephone lines), and the Internet service transmitted through that pipeline. However, unlike other ISPs, @Home controls all of the transmission facilities between its subscribers and the Internet. To the extent @Home is a conventional ISP, its activities are one of an information service. However, to the extent that @Home provides its subscribers Internet transmission over its cable broadband facility, it is providing a telecommunications service as defined in the Communications Act. [ 74 ]
Based on this conclusion, the appeals court decided that Portland does not have authority to order third party access.
Superficially, such a result might appear to be a victory for the cable companies. It greatly lessens the likelihood that localities across the country will be able to successfully impose third party access requirements on cable carriers. If the Ninth Circuit’s analysis is followed by other courts around the country, then there is little risk that a regulatory patchwork quilt will arise, consisting of different and possibly conflicting local regulation on the third party access issue.
On the other hand, the ruling that the Internet service is not a "cable system" but a "telecommunication service" greatly increases the pressure on the FCC to address the regulatory classification issue — and the third party access issue — directly. Assuming the agency agrees that Internet access over cable is most fairly viewed as a telecommunications service, then the FCC will be forced either to consider greater regulation of such service or to explain and defend specific reasons why such regulation is not appropriate.
It appears that the Ninth Circuit decision has had the expected impact on the FCC. On September 28, 2000, the agency issued a Notice of Inquiry that will address the third-party access issue. [ 75 ]
Although not directly relevant to the classification of cable Internet access under American statutes and regulations, in 1996 the Canadian Radiotelevision & Telecommunications Commission (the "CRTC," Canadian equivalent to the FCC) confronted and resolved a directly analogous issue. The CRTC determined that under the relevant Canadian statutes, Internet service and other data delivery over a cable system should be treated and regulated as a telecommunication service. [ 76 ] The CRTC thus subjected the Canadian cable industry to the type of tariff-based regulation that applies to Canadian telephone companies.
At the highest level, advocates for "open access" have argued that there is a risk that a handful of cable company-owned ISPs (primarily @Home and Roadrunner) could come to dominate the provision of broadband access to consumers. They contend that this could lead to fundamental negative changes in the Internet and reduce the free flow of information and opinions on the Internet. Their contentions include:
In opposition, the cable industry contends that the government should continue to refrain from regulating the Internet, and that "forced access" would significantly impede and delay the deployment of broadband access facilities across the country. Their contentions have over the past two years included:
Since as early as 1998, public interest advocates, ISPs, cable companies, and other Internet interests have engaged in an aggressive debate concerning the appropriateness and viability of requiring cable companies to open their networks to unaffiliated ISPs. It was suggested early in the debate that cable systems simply could not, as a technical matter, support more than one ISP. Today, however, most acknowledge that a cable system can support multiple ISPs. Although some important technological issues remain, most of the debate over the past year focused on non-technological concerns.
Two major shifts have occurred in the alignment of companies in the different debate camps. Originally, AOL was one of the leading companies specifically calling for government action to require third-party access, and AT&T was allied with other cable companies (including Time Warner) in defending their plans to offer only a single ISP over their cable systems. [ 77 ] During the latter half of 1999, however, AT&T publicly endorsed the concept of voluntarily opening its cable networks to unaffiliated ISPs. On December 6, 1999, AT&T submitted a letter with others to FCC Chairman Kennard in which AT&T committed to allow "multiple" ISPs to offer broadband services over its systems (although AT&T would not actually implement third party access until the expiration of certain exclusive contracts in 2002 or later). [ 78 ] By making this commitment, AT&T broke ranks with other cable companies that continued to resist the possibility of unaffiliated ISPs offering service over their cable systems.
Early in 2000, AOL and Time Warner announced their planned merger, and announced a shift of position for both companies. Together, the companies pledged that they would voluntarily open the Time Warner cable systems to "multiple" unaffiliated ISPs. At the same time, AOL indicated that it no longer advocated government action to force "open access."
More recently, some other cable companies have moved toward the "voluntary open access" position. According to published reports, both Comcast and Cox [ 79 ] have committed to open their systems when certain exclusive contracts expire. [ 80 ] Other cable companies may follow suit.
Thus, there are now three camps in the debate: (1) advocates of government mandated open access, (2) proponents of voluntary open access, and (3) those that oppose government action but also have not agreed to open access voluntarily. The primary focus of the debate has shifted from whether there will be third party access on most major cable systems, to whether such access will be mandated and how will access will be implemented and monitored.
To date, the FCC has generally sided with the cable companies, in allowing the marketplace to function without regulatory intervention. In January of 1999, the FCC completed its first detailed analysis of the advanced services market, including high-speed access to the Internet. The Commission concluded that the broadband market was very new, that different broadband technologies were being deployed across the country, that regulatory action could slow the deployment of broadband services, so there was no pressing need for action at this time. [ 81 ] The FCC indicated that it would continue to monitor the developments in the market. [ 82 ] It reaffirmed this position when it approved the proposed AT&T/TCI merger without imposing any third party access requirements. [ 83 ]
FCC Chairman William Kennard has repeatedly stated his opposition to intervention in the broadband market. He has asserted that instead of a monopoly in broadband, there is a "no-opoly" since broadband has not yet been deployed to any significant extent. Kennard has also asserted that the stepped up action of the ILECs in deploying DSL came about in response to competition from cable modem services.
More recently, FCC officials have affirmatively endorsed the concept of "openness" for cable systems, but have indicated that the FCC hopes that the marketplace will achieve openness without government intervention. [ 84 ] The officials have said that "[i]f market forces fail and cable becomes the dominant means of Internet access, regulation might then be necessary to promote competition." [ 85 ]
Most recently, as discussed above, the decision of the U.S. Court of Appeals for the Ninth Circuit has effectively forced the FCC to initiate a formal proceeding to consider the issues surrounding third party access. Within days of the Ninth Circuit decision, the FCC announced that it would initiate a proceeding to consider the third party access issues. [ 86 ] As mentioned above, the FCC recently issued a Notice of Inquiry ("NOI"), although the fact that the FCC is moving forward with an NOI rather than a Notice of Proposed Rulemaking suggests that the agency has not reached any conclusion as to whether affirmative regulation would be appropriate. In an accompanying press statement, FCC Chairman Kennard summarized his current thinking on the third-party access question:
"I have strongly advocated a policy of regulatory restraint in the cable modem services market. I have been patient in allowing the marketplace time to develop, while monitoring it closely to assess its openness. It is unclear, however, whether a marketplace solution will develop absent some form of intervention. Through this inquiry, we need to determine whether a regulatory response is appropriate to ensure that this exciting new technology develops the open competition and innovation that we cherish in the Internet." [ 87 ]
A variety of bills pending in Congress address broadband issues, but most are focused on removing some of the regulations under which ILECs operate. Five bills could have direct bearing on the third party access issue. Three of these five would mandate third party access, while two would strip the FCC of authority to mandate third party access:
H.R. 1685, The Internet Growth and Development Act and H.R. 1686, The Internet Freedom Act. Both of these bills declare that "broadband access transport providers" must offer unaffiliated ISPs the same terms that are offered to affiliated ISPs, and must not otherwise favor affiliated ISPs. Both were introduced by Reps. Goodlatte and Boucher on May 5, 1999, and were referred to the Judiciary Committee and the Commerce Committee, which in turn referred the bills to its Subcommittee on Telecommunication, Trade and Consumer Protection. The Judiciary Committee held a legislative hearing on June 30, 1999.
H.R. 2637, The Consumer and Community Choice in Access Act. The bill would require cable companies that provide Internet access to allow third party ISPs open access to the cable plant. It would also empower the FCC to set financial and technical terms for such third party access, should the cable companies fail to do so. The bill was introduced on July 29, 1999, by Rep. Blumenauer, and was referred to the Commerce Committee, which in turn referred the bill to its Subcommittee on Telecommunication, Trade and Consumer Protection.
H.R. 2420, The Internet Freedom and Broadband Deployment Act of 1999. The bill would deprive the FCC and state and local authorities of regulatory authority over high-speed data service or Internet access services. The bill was introduced by Rep. Tauzin on July 1, 1999, and was referred to the Commerce Committee, which in turn referred the bill to its Subcommittee on Telecommunication, Trade and Consumer Protection.
S. 1043, The Internet Regulatory Freedom Act of 1999. The bill states that U.S. policy is to assure that all Americans have access to advanced Internet services at affordable rates by eliminating regulation that impedes the competitive deployment of advanced broadband data networks. The bill specifies that nothing in this Act grants authority to the FCC with respect to: (1) the provision of Internet services, (2) technical standards or specifications for the provision of Internet services, or (3) the imposition of wholesale discount obligations on bulk offerings of advanced services. The bill was introduced by Sen. McCain and referred to the Committee on Commerce, Science and Transportation.
In addition to these proposals, other bills have been introduced that would affect broadband deployment, but which would not directly affect the third party access issue. For example, S. 877, The Broadband Internet Regulatory Relief Act, introduced by Senator Brownback, would remove the FCC’s ability to impose regulation on DSL service, in an effort to promote DSL deployment in rural areas. S. 2307, The Rural Broadband Enhancement Act, introduced by Senator Dorgan, would make loans available to companies seeking to bring broadband services to rural areas. Interestingly, S. 2307 appears to assume that all Internet service (whether over cable, DSL, or another technology) is a "telecommunications service."
State and local governments have been an active battleground in the fight over third party access, typically in the context of reviewing AT&T's proposed mergers with TCI and/or MediaOne (or a similar cable company transaction that often requires approval or consent from local cable regulatory authorities). The local jurisdiction that has garnered the most attention has been Portland, Oregon. As described above, Portland imposed a third party access requirement as a condition of approval of AT&T’s merger.
The majority of jurisdictions — including major cities such as Los Angeles and Miami — have rejected calls to impose a third party access requirement. [ 88 ] San Francisco approved the AT&T-TCI merger without specific conditions (though it did adopt a policy of supporting third party access). Some jurisdictions have included a condition on the merger that allows the local regulator to reconsider the access issue if Portland is ultimately successful in its efforts to force third party access. [ 89 ]
However, a small but growing number of jurisdictions have imposed third party access requirements on the merger, including Broward County, Florida, Fairfax City, Virginia, Cambridge, Massachusetts, and St. Louis, Missouri. [ 90 ] As in Portland, AT&T and other cable companies have brought suits to challenge some of the local actions. [ 91 ] Also as in Portland, the district courts that have addressed the issues to date have ruled against the cable interests. [ 92 ]
The decision by the U.S. Court of Appeals for the Ninth Circuit in the AT&T v. Portland case, however, is likely to have a significant impact on any efforts by local authorities to impose third party access. As discussed more fully above, the Ninth Circuit concluded that the local government did not have the authority to impose a third party access requirement on a cable company. Its analysis, and the fact that the FCC is undertaking its Notice of Inquiry, greatly strengthens the argument that local governments lack the power to impose third party access.
In addition to action by local governments, ballot initiatives or state legislative proposals to require third party access have been proposed in a number of states, including Colorado, Massachusetts, and Pennsylvania. [ 93 ] These efforts also may well be preempted by FCC action (or an affirmative decision not to act) on the third party access issue.
Canada and Brazil have directly addressed the third party access issue, and both have required cable systems to support third party access.
On a per capita basis, Canada is the world leader in cable modem penetration, both in the percentage of homes to which the service is available and the "take up" percentage of users opting to obtain service. The Canadian Radio-television and Telecommunications Commission ("CRTC") — the Canadian equivalent of the U.S. Federal Communications Commission — first ordered third party access in January 1996. Its order expressly covered, but did not extensively discuss, the Internet. [ 94 ] In July 1998, the CRTC more specifically reaffirmed cable providers’ obligation to support third party ISP access. [ 95 ] Implementation of third party access, however, has not yet been finalized, and ISPs in Canada have accused the cable industry of dragging its feet. To pressure the cable industry, the CRTC ordered in September 1999 that cable providers make high speed Internet access available for resale by third party ISPs until such time that full third party access is implemented. Thus, although third party access has been the rule in Canada for some time, it has yet to be implemented broadly.
In November 1999, the Brazilian communications regulator, Anatel, ordered that cable systems support access by third party ISPs, but details of implementation of such access have yet to be finalized. [ 96 ]
The FCC twice looked at the emerging cable industry in the 1950's, and both times decided against imposing common carrier obligations or other regulations on cable operators. [ 97 ] In 1958, the later of the two decisions, the FCC declined to impose common carrier regulation in part because cable subscribers did not select the particular messages that the cable system would carry (as subscribers do, for example, when they place a phone call). [ 98 ]
By 1965, however, concern over cable's impact on local broadcast television led the FCC to assert jurisdiction over cable systems, and impose "must carry" rules requiring cable systems to carry certain local TV stations. [ 99 ] In 1972, the FCC issued its first set of comprehensive regulatory rules for the cable industry. By the late 1970's, however, the FCC had begun to deregulate the industry. In a continuation of that deregulatory effort, Congress in 1984 enacted the Cable Communications Policy Act; as part of that Act, Congress delineated and restricted the FCC's regulatory authority, distributed limited regulatory power among local, state, and federal governments, and generally removed most rate regulation from the cable industry. [ 100 ]
By the late 1980's, however, concerns had emerged that (a) the cable industry was operating as a largely unregulated monopoly, (b) prices were too high, and (c) there was too much vertical integration consolidating video program production and distribution. [ 101 ] In response, Congress enacted the Cable Television Consumer Protection and Competition Act of 1992 (the "1992 Cable Act"). [ 102 ] The Act imposed strong regulations on the cable industry, including rate regulation, "must carry" requirements, limitations on ownership of competing technologies by cable operators, and specific provisions designed to guard against discrimination by cable operators in favor of affiliated content providers. [ 103 ]
Certain of the findings, policies, and public interest objectives that motivated Congress to act with regard to video programming are echoed in the arguments and concerns expressed today by proponents of third party ISP access to cable systems. Among the findings made by Congress in 1992 are:
Following its articulation of the above findings, Congress declared that it is the "policy of the Congress" to:
In directing the FCC to implement cable cross-ownership limitations, Congress required that the FCC pursue the following "public interest objectives":
In the Act, Congress prohibited exclusive contracts between cable operators and affiliated program vendors, unless the FCC concluded that such exclusive contract was in the public interest, with the FCC being required by the Act to consider the following public interest factors:
Four years after passage of the 1992 Cable Act — which re-imposed significant regulation on the cable industry — Congress again changed course and removed much of the new regulation imposed in the 1992 Act. In the Telecommunications Act of 1996, [ 108 ] Congress removed most rate regulation and eased prohibitions on ownership by cable operators of competing video distribution technologies.
The issues and policies articulated by Congress in the 1992 Cable Act make clear that many of the concerns and arguments made today by proponents of third party access are not new, but have been raised before in the context of the cable industry's relationship with video content providers. By the same token, some of the provisions of the 1992 Cable Act — if applied to Internet service — would go a significant distance toward addressing the concerns of the third party access proponents. Thus, in this regard, imposing a third party access requirement on the cable industry would not be a wholly unprecedented development.
On the other hand, even this brief history of cable industry regulation shows that regulation has hardly been consistent or well settled. The cable industry has not operated under any sort of strict regulatory regime for extended periods of time. Moreover, the cable industry has never been subject to the type of common carrier regulation that has been applied for years to telephone companies. The cable industry does not have significant experience with detailed cost-based regulation and tariff requirements similar to the requirements have been placed on incumbent local exchange carriers. In light of the extensive litigation that has marked the regulatory efforts to force "open access" on telephone local exchange carriers, cable companies are understandably concerned about the prospect of any regulatory scheme forcing third party access on cable systems. [ 109 ]
A variety of technological claims and concerns have been advanced over the course of the debate about third party access. Most fundamentally, many initially questioned whether third party access was even possible on a cable system. That question has been answered, but there remain significant technological concerns and issues that are relevant to the third party access question.
Simply stated, an upgraded cable system can support more than one ISP. Although there is no clear answer as to the best or most effective method to implement such support, there is no longer significant question that multiple ISPs can be supported. In Canada, third party access has been implemented in Timmons, Ontario, Sudbury, Ontario, and a limited number of other locations. The Canadian Cable Television Association, which represents all of the major cable companies in Canada, does not dispute that such access is possible, and in fact the CCTA is involved in government-mandated trials of third party access. Similarly, Knology cable systems in the United States has implemented third party access in a limited number of systems. In Clearwater, Florida, GTE has demonstrated third party access purely on a test basis. Most significantly, AT&T has announced its first technical trial of third party access on one of its cable systems. [ 110 ] As made clear by the recent movement by leading U.S. cable companies to accept some form of voluntary open access, no one is strongly asserting that it is not possible.
It is not clear, however, how to best implement third party access. In simple terms, third party access will be implemented by the installation of a "router" or similar piece of equipment at one or more points within a cable system. A router is a commonly used piece of network electronics that would scan the TCP/IP packets and route them to the appropriate ISP. Where the routers are placed, and how the router determines the appropriate destination, are questions to which there are not currently clear answers. [ 111 ]
On the question of router placement, a router could be placed immediately "behind" the "cable modem termination system" ("CMTS") that is installed in the cable headend. This approach might favor very small ISPs, which would be able to connect to routers that are (a) located close by, and (b) are very geographically focused. An arguably more efficient approach would be to place the router at a central point of aggregation where data from more than one CMTS, or possibly even more than one cable system, can be collected and then routed. The latter approach would require fewer routers, and would allow ISPs to connect to a much larger number of potential customers at a single point. For small, local ISPs, however, the latter approach could be very expensive, since the routers to which a small ISP would need to connect could be located a long distance away.
There are also a variety of ways the router could determine the appropriate destination. One method under consideration in Canada is "source routing," where the router looks at the source of each packet (i.e., which user has sent the packet) to determine which ISP should receive the packet. This approach, however, would require routers to look at more of the packet than routers normally are designed to do. Other approaches are also possible.
Aggravating the engineering issues is the fact that the equipment manufacturers have not historically been asked to design equipment that supports third party access in cable systems. Because there has not been a defined market in the United States for such devices, manufacturers have continued to produce only equipment that meets cable systems’ current needs—and that equipment generally does not support third party access. Although third party access has been ordered in Canada, Canadian cable companies have not to date constituted a significant enough market for equipment manufacturers to dedicate their development efforts. As U.S. cable companies turn their attention to supporting third party access, leading equipment makers will likely focus on the requirements of such access.
To be clear, there are very significant engineering challenges raised by the provision of any Internet service over a cable plant. As discussed in the following section, there is a serious need for careful management of the available bandwidth on a cable system, and limitations on how any ISP or any individual uses the bandwidth may be appropriate. Many of these concerns, however, are present with or without third party access.
There are significant technological concerns and potential problems that are inherent in the provision of Internet access over a cable system. As noted, an upgraded cable system supporting Internet access is a "shared" system, in which the users all share a finite amount of bandwidth. All of the users supported by a single cable modem termination system ("CMTS") share the same bandwidth. A single CMTS may be used to support hundreds (or more) of simultaneous Internet users.
For example, if the CMTS has been configured to support 1 Mbps of upstream data, then all of the users on that CMTS share that 1 Mbps. When only a single user is actively using the Internet connection, then that one user has full use of 1 Mbps of upstream bandwidth. When a second user accesses the Internet, and both users are transmitting very large files, both users will effectively be able to use 500 kbps of bandwidth. In reality, it is unlikely that the two users would both be fully loading the upstream bandwidth at the same instant, so that both users could perceive that they had nearly 1 Mbps of bandwidth available. If, however, there are 400 people sharing the bandwidth, then none of the users is likely to have a full 1 Mbps of upstream bandwidth available. Because not all 400 users will be transmitting data upstream at the same time, and because in most surfing of the World Wide Web the upstream data flow is quite limited, it is quite possible that all 400 users will be generally satisfied with the upstream data flow.
A significant potential exists that a small number of users (or even one user) could "hog" a large portion of the available bandwidth, to the detriment of all other users. In the above hypothetical, assume that a user were to "host" a World Wide Web site using the cable modem service (such that the user's computer would have to respond to web requests by transmitting data upstream on the cable system). If the hypothetical Web site were to become popular and have large graphic images, all other users sharing the same CMTS would likely experience very significant slowdowns in upstream data flow. [ 112 ]
The potential for slowdowns caused by one or more bandwidth "hogs" is equally plausible with downstream bandwidth. A CMTS may support, hypothetically, 30 Mbps of downstream data in aggregate. If 400 users share that bandwidth and a handful of them are constantly downloading huge data files, the other users will likely experience significant drops in performance. Even without any bandwidth "hogs," the 400 hypothetical users will likely experience slowdowns during peak usage periods, but bandwidth "hogs" could bring the downstream data flow nearly to a stop.
Theoretically, a cable operator could increase the amount of bandwidth available for Internet access in one of two ways. First, the operator could install more CMTSs, thereby reducing the number of customers sharing a single CMTS, and thus increasing the bandwidth per customer. Second, a cable operator could on a system wide basis dedicate additional bandwidth to Internet access, reducing the number of ordinary cable channels offered. Although both solutions may be desirable to those who believe that more bandwidth is always better, both approaches — if ordered by a government body — would be a significant regulatory incursion into the operations of the cable system.
There are two more practical approaches to the problem of bandwidth "hogs." First, a cable operator may impose rules about the types of operations that the cable system will support. Thus, for example, it is common for cable operators to prohibit the operation of a World Wide Web server over the cable modem system. [ 113 ] Alternatively, cable operators in Canada and elsewhere have taken the approach of pricing their cable service according to the amount of bandwidth utilized. Thus, the hypothetical bandwidth "hog" would be permitted to continue use the higher amount of bandwidth, but would be required to pay for the additional bandwidth (and this increased payment could be used to finance capital expenditures to increase bandwidth).
Given the shared nature of the cable system, bandwidth usage guidelines will be necessary for cable modem service from any ISP, whether or not third party access is imposed. The need to impose some constraints on bandwidth usage (whether by rule or by higher fees for more bandwidth usage) is independent of the third party access question.
If third party access were implemented on a cable system, the need to control bandwidth might narrow the business models of ISPs seeking to provide service over the cable system. Unless a bandwidth-sensitive pricing scheme were to be implemented, it is unlikely that ISPs would be able to offer to support web servers over the cable system, or other similar bandwidth-intensive applications. As implemented by most cable systems, cable modem service simply may not support the full range of Internet Protocol ("IP") based applications, and thus cable service may be a more narrow and less flexible offering than DSL service or other high-bandwidth options.
There was in the third party access debate much discussion of a ten-minute limit that cable companies imposed on "streaming video" delivered over the Internet. The @Home company's agreements with cable operators to provide Internet service over the cable systems include a provision that prohibits Internet users from accessing more than ten minutes of "streaming video," which is a primary method of delivering full motion video over the Internet. Although the @Home company has stated that this limitation was an outgrowth of the bandwidth concerns discussed above, critics of @Home and the cable operators have asserted that the cable companies imposed the limit on streaming video because "they believe that Internet-quality streaming video competes with traditional cable television service." [ 114 ]
The ten-minute limit on streaming video was a plausible and not unreasonable response to a concern about bandwidth. But it is also certainly plausible that there was some desire to limit competition in the delivery of video. Indeed, a senior executive from @Home was recently quoted as saying that "[c]able operators imposed [the ten minute limit] to ensure Excite At Home didn’t compete with cable." [ 115 ]
In any event, this issue appears to have resolved itself. The @Home company has stated that the ten minute limit was never enforced, and the specific limit appears to have been omitted from @Home's current Acceptable Use Policy. [ 116 ] The merged AOL Time Warner has indicated that it would not impose a "ten minute limit." [ 117 ] Moreover, in the long run, there would likely be market forces that would discourage enforcement of this type of restriction.
The ten-minute limit does illustrate, however, the risk that control by a single company over broadband access could theoretically have a direct and negative affect on innovation and the free exchange of information over the Internet. If such a limitation is technologically required given the shared nature of the cable resource (as discussed in the preceding section), then a limitation would likely be appropriate.
A key issue in considering whether government-imposed third party access is appropriate is whether consumers will be able to receive broadband access to the Internet from sources other than cable companies.
Today, DSL services from incumbent and competitive local telephone companies directly compete with cable modem services. Although significant differences between the two services exist —cable modem service promises higher speeds while DSL service offers more consistent speed — they are similar enough to be effective substitutes for today's Internet applications commonly used by residential users. [ 118 ]
In a handful of markets today, wireless technologies provide broadband access, but wireless providers have not to date widely marketed their services to residential areas. Companies such as Winstar, Teligent, and Nextlink target businesses, while Sprint, WorldCom, and other companies with wireless services have only a very few residential customers. Thus, today, broadband access for individuals is largely limited to cable modem and DSL service.
Estimates of how widely cable and DSL will be deployed vary greatly. It is likely that by 2003-2004, at least 75% of U.S. households will be able to get cable modem service, and perhaps as much as 66% of households will be able to receive some form of DSL service. Over the same period, wireless services should become more available (including for residential service), but it is unclear how wide that deployment will be for residential service. Similarly, broadband satellite services are predicted to begin to be offered by 2002 or 2003, but again, it is not clear whether such services will be price competitive in the residential market. [ 119 ]
A key question is how much choice, if any, individual users will have between different broadband access technologies. One analyst who follows the broadband market closely has concluded that "[i]n the next three to four years, . . . up to 20% of the country may have a choice of three to four different broadband facilities, roughly 30% of the country may have the choice of two, and half of the country may have only one or no broadband facility to choose from." [ 120 ] This type of estimate is very difficult to confirm based on currently available information, and these numbers reflect significant (but not unreasonable) skepticism about the business viability of residential broadband service delivered by wireless or satellite means. Nevertheless, it does appear likely that a large percentage of potential residential consumers in the United States will have at most a choice of two broadband pipes, and many consumers will have one or zero choices.
To put these numbers in further perspective, it is highly likely that in 2003 a majority of citizens will still get access to the Internet either (a) at their place of employment, or (b) using traditional, dial-up, narrowband Internet access methods. Although broadband access will surge in importance, narrowband access will remain a significant force for years to come.
Proponents of government-imposed third party access have raised a variety of concerns, the most significant of which are discussed below.
Advocates of third party access have argued that in the absence of an access requirement, cable operators would be able to restrict consumers' access to Internet content or otherwise censor speech. This is a valid concern, although the risk of censorship may not be eliminated by imposition of a third party access requirement. Short of imposing common carrier obligations on cable companies, it is at least theoretically possible that a cable company could limit the type of content carried on its system. In theory, the cable companies have a First Amendment right not to carry particular content if they do not choose do to so, and this right would apply whether a cable affiliated ISP imposed the limit or a cable operator imposed the limit on affiliated and third party ISPs alike. A government action to prohibit a content-based limitation imposed by a cable owner would raise difficult constitutional questions — questions that cannot easily be answered in the abstract.
Thus, there is a risk of cable operator-imposed censorship regardless of the third party access issue. The practical risk of such censorship, however, would be significantly reduced if third party access were required. It would not likely be in the business interest of the cable operator to impose a content based limit on ISPs offering service over the cable network. If more than one ISP offers service over the cable system, there is a lower likelihood that all of the ISPs would themselves choose to impose a content-based restriction.
In the absence of third party access, there is also a greater risk that a government would legislate content-based restriction to advance social agendas unrelated to third party access. If, at some time in the future, only a handful of companies controlled the vast majority of Internet access services to the home, the government might conclude that regulating content could be accomplished by regulating the small number of companies controlling broadband access to the home. The chance of governmental regulation would likely diminish if hundreds or thousands of ISPs continue to offer a wide range of access choices.
Another significant concern advanced by proponents of third party access is that, would a cable operator would be able to discriminate in how fast content is delivered to consumers. Not only would discrimination among content providers be possible, such discrimination is already a part of many business models, if not most business models, of broadband access providers. For example, one of the strong market selling points of the leading cable broadband access provider, @Home, is that @Home has developed a sophisticated system of "Super Nodes" that are specifically designed to be able to deliver selected broadband content to @Home's subscribers.
Thus, hypothetically, if @Home were to contract with a specific automobile manufacturer, that manufacturer's broadband sales videos would be delivered more quickly and reliably than those of a competing content provider. As an equally plausible hypothetical, if America Online were to develop and offer specific broadband content to its customers, @Home cable modem subscribers would likely not be able to access AOL's broadband content as quickly as the broadband content that @Home maintains on its local servers.
In this latter hypothetical, the discrimination between broadband content would not necessarily disappear in a third party access situation. If third party access were ordered and AOL were able to offer service directly to cable modem subscribers, it is very likely that AOL's network would continue to favor its own broadband content, and non-AOL broadband content would likely be delivered to AOL subscribers more slowly than AOL-hosted content. Likewise, subscribers who choose @Home instead of AOL would likely not be able to access AOL broadband content as quickly as @Home content. In other words, there may be an element of discrimination among broadband content inherent in how broadband networks are being constructed on the Internet. This inherent discrimination is part of a larger concern about the development of broadband networks, and is discussed more fully in a separate paper issued by the Center for Democracy & Technology and its the Broadband Access Project, entitled "The Broadband Internet: The End of the Equal Voice?" [ 121 ]
Thus, for policy makers confronted with the third party access debate, the issue may not be whether certain broadband content will be favored over other, but whether cable modem subscribers will be able to choose among broadband providers (and thus be able to choose which set of broadband content will be the favored content to which they will have quick access).
Another concern is that in the absence of a third party access requirement, a cable operator might be able to charge higher than competitive prices. Although there is some validity to this concern, the current and likely competition between cable and DSL service in parts of the country may minimize this risk.
Certainly, in the absence of third party access, a cable operator would not face direct price pressure over the cable network, and might be able to maintain higher than fully efficient prices. [ 122 ] The narrowband Internet saw the emergence of vigorous price competition from the thousands of ISPs that compete for dial-up customers. That direct price competition would be lacking in the absence of third party access.
On the other hand, competition from DSL technology will exert some pressure on prices. In Canada, DSL and cable providers have viewed themselves as in direct competition with each other, and it appears that there has been price pressure across technologies. (Indeed, in Canada, both DSL and cable modem service are priced significantly lower than comparable service in the United States, [ 123 ] although it is difficult to determine if the lower prices are attributable to a greater level of competition.) Similarly, in the United States, it appears that "DirecTV" type of satellite dishes are providing competition to cable. Thus, it is likely that there would be some degree of price pressure even in the absence of third party access.
Proponents of third party access assert that without a third party access requirement, a consumer who has a preferred ISP would have to "pay twice - once for the gatekeeper's ISP, and once for the ISP they want." [ 124 ] As a concrete example, a cable modem subscriber might have to pay both @Home (for the broadband connection) and AOL (for access to AOL’s proprietary content). This argument, however, may create some confusion about what the third party access proponents really are seeking. Ultimately, the argument both overstates and understates significant concerns.
As a threshold matter, many people would have very little interest in retaining an old ISP, and thus would not even consider "paying twice" to keep a prior ISP. For individuals who currently obtain simple Internet access from an ISP that does not itself provide proprietary content, the primary effect of being forced to switch ISPs would be to require that the user obtain a new e-mail address. The main concern of millions of Internet users is simply the ability to access the Internet rather than who provides that access. For these users, the "pays twice" argument does not really apply.
Being forced to change e-mail addresses is not, of course, a trivial problem. Changing an e-mail address can be as disruptive as changing a telephone number or mailing address. But, even if third party access were ordered, many consumers would still have to change their e-mail addresses simply because some ISPs will not convert their systems to support broadband speeds. [ 125 ] In other words, the e-mail address problem will occur to a greater or lesser extent with or without third party access. Moreover, the change in e-mail address is a one-time problem that, once addressed, will not cause continuing issues.
The major group of Internet users— more than 20 million users of them —adversely affected by a forced ISP change would be subscribers to America Online. AOL subscribers are the primary group who might in fact decide to "pay twice," to be able to continue to get access to AOL addressed e-mail, and to get access to proprietary content offered by AOL. Very few ISPs other than AOL provide a significant amount of content that is not generally available to the Internet, and thus few non-AOL users would be unable to get direct access to content that previously had been available to the users. Moreover, information provided by AOL indicates that the vast majority of its users spend most of their time within AOL's proprietary service, and do not primarily use AOL as a way to access the Internet. Thus, consumers who use the services of one company — AOL — may be more severely harmed than users of other ISPs by a failure to require third party access.
AOL does offer a content-only subscription (for users who want access to AOL content and e-mail, but who already have their own connection to the Internet, such as a cable modem service). Thus, those users would be able to get access to AOL, but they would have to "pay twice." That access, however, would not be the same quality access than if AOL were able to offer access services directly over the cable. Any content-only access to AOL that cable modem users received would by necessity run through the facilities and backbone connection of the cable-affiliated ISP. Thus, AOL would have very little ability to optimize its content to take advantage of the high-speed connection. Were AOL to offer access service directly over the cable, it would be able to ensure that broadband content was provided to the consumer from a close, local, and fast content server. Indeed, this is precisely one of the advantages that the @Home company offers today — a network of local content servers that can provide fast content to users.
AOL users that want to continue using AOL's services in the absence of third party access would have to "pay twice" to get to AOL, and even then would not be able to get the same quality broadband connection that AOL would likely offer directly. For non-AOL users, the primary downside of having to change ISPs is the change in e-mail addresses, which is a significant but not insurmountable burden.
Opponents of imposed third party access have raised a variety of concerns, the most significant of which are discussed below.
Some concern has been expressed that cable companies would not upgrade their cable systems in the face of a third party access requirement.
The U.S. cable industry started down the road toward upgraded HFC cable plants in the early 1990's, for reasons wholly unrelated to the Internet. A CableLabs overview of the development of cable television — prepared in 1995, before the Internet really exploded — explained that a primary driver for the cable industry’s adoption of HFC architecture was digital video. [ 126 ] According to that report, the emergence of digital video compression — which allows cable operators to deliver dramatically more video over the coax cable — "inspired the cable industry to dramatically upgrade its physical facilities." [ 127 ] A primary "goal" of an upgrade was to "[I]ncreas[e] the [cable] plant's channel capacity." [ 128 ] Beyond digital video, the cable industry continues to pursue interactive TV as a vehicle to deliver electronic mail, online shopping, and electronic channel guides and information. [ 129 ] These services also require an upgraded cable architecture.
By the early 1990's, the cable industry had confronted the reality of serious competition in the delivery of video services (from direct broadcast satellite, wireless, and cable overbuilders). [ 130 ] It looked to HFC plant architecture and other new technologies to be able to compete. Thus, there is a significant impetus for a cable operator to upgrade to a two-way HFC cable system unrelated to the possibility of providing Internet service. [ 131 ] In Canada, where the Canadian Radio-television and Telecommunications Commission indicated as early as 1996 that it would order third party access, cable systems nevertheless have nearly fully upgraded their cable plants. U.S. cable operators would likely upgrade their plant even if third party access were to be ordered.
On the other hand, there is no question that additional investment is required to make an HFC cable plant support Internet service. Each cable modem termination system (CMTS) can cost tens of thousands of dollars, and a large system could require dozens of CMTS boxes. A cable operator must then design and build a network to connect the CMTSs (and likely must finance the cable modems installed in users' homes). Thus, although most cable systems will likely upgrade their facilities to HFC architecture, the decision also to support Internet access service still requires significant investment.
Opponents of mandated third party access have asserted that outside investment in cable companies would dry up if third party access is ordered.
A governmental decision imposing third party access would have an impact on the value of cable enterprises, which in turn could reduce the amount of capital available to cable operators to invest in system upgrades. The stock market has followed, and reacted to, developments in the third party access debate, and cable stocks have declined following victories for the third party access proponents.
On the other hand, if a third party access requirement were imposed and the requirement were structured in a way that provided a fair price to cable operators for services provided to third party ISPs, it is likely that investment dollars would continue to be available to fund cable upgrades. Internet access is not the only, or even primary, reason a cable operator would need to upgrade its system, and so long as the operator is able to obtain a fair return on the Internet-specific portion of the upgrade, the upgrade should still make business- and investment-sense. The likelihood that investments would not stop in the face of a third party access requirement is supported by the fact that Microsoft recently invested over $400 million in Canada's Rogers Communications, Inc., notwithstanding the fact that Rogers' cable systems are operating under a still-developing third party access regime. [ 132 ]
Indeed, a clear resolution of the third party access debate could even increase investment in cable facilities. Opponents of third party access have correctly asserted that investment money avoids situations with regulatory uncertainty. The opponents have argued that Congressional or FCC consideration of third party access proposals would create uncertainty and chill investment. It is possible, however, that the prolonged third party access debate itself has created market uncertainty that a resolution of that debate (even one that results in a third party access requirement) could reduce.
Finally, the fact that four leading U.S. cable companies have voluntarily committed to some form of third party access, and those companies have not experienced massive drops in stock price or investment, suggest that the U.S. financial markets are not skittish about third party access per se. If third party access were imposed by Congress or the FCC, the cable companies would incur significant additional expenses, but it is not clear that financial markets would react differently than they have already reacted to voluntary third party access.
Opponents of mandatory third party access have expressed the very significant concern that any such requirement would lead to complex regulation and litigation.
Although simpler and more focused, a third party access requirement would not be different in kind than the local telephone competition provisions of the Telecommunications Act of 1996. Those provisions, which required that incumbent local exchange carriers open their networks and allow third parties to provide telephone services, have led to truly massive litigation and regulatory disputes. The prospect of replicating that degree of litigation in the cable arena has almost certainly been a major factor underlying the FCC's "hands off" approach to third party access.
A third party access requirement would raise a host of very difficult issues, including (a) how much a third party ISP would pay to a cable company, (b) how and where the ISP would interconnect into the cable system, (c) what limitations, if any, would be placed on the ISP's use of the bandwidth, and (d) whether the cable company continued to favor an affiliated ISP even after third party access was ordered. These are all directly analogous to difficult questions raised about competition in local telephone markets.
The challenge posed by a third party access requirement may even be greater in the cable arena than in the telephone arena, for two reasons. First, because the bandwidth over the cable system is a shared resource, an individual ISP or user could (either accidentally or intentionally) interfere with the services provided by a competing ISP. The shared nature of cable increases the need for genuinely cooperative and concerted efforts between a cable company and an ISP.
Second, some data and bandwidth management functions can only be performed by a single responsible entity. To date, cable companies have contracted out those functions to their affiliated ISPs (@Home, Roadrunner, ISP Channel, etc.), and have not maintained network engineering staffs capable of implementing and maintaining the cable modem termination system and supporting network equipment. In a third party access situation, it is likely that the affiliated ISP would continue to perform those functions on behalf of the cable operators. If that were to happen, then the affiliated ISP would have significant ability to interfere with the services and capabilities available to third party ISPs.
In an effort to avoid these potential disputes, America Online and other ISPs have expressed their willingness to agree to the same contractual terms as exist between cable operators and their affiliated ISPs (@Home or Roadrunner, etc.). This assertion, however, ignores the reality that under those terms, affiliated ISPs perform management functions that cannot practically be split among more than one company. Although contracts between cable operators and their affiliated ISPs are confidential and closely held, it is possible that some affiliated ISPs perform bandwidth management functions in partial exchange for the ability to market cable modem services to the subscriber (and thus, a separate dollar value of the management services has not been established). Moreover, in the case of AT&T and @Home, the contract between those two companies may require @Home to utilize AT&T's backbone network for at least some of @Home's Internet traffic. Thus, even if third party ISPs were allowed on a cable network, it would not likely be on the same precise terms as @Home. Therefore, if the third party access requirements mandated non-discrimination between ISPs, there would be a need to evaluate the financial significance of the bandwidth management portion of contracts between cable operators and affiliated ISPs.
In the face of these challenges, it is clear that the most desirable way to achieve a goal of third party access to cable systems would be for cable operators and ISPs to negotiate workable contractual arrangements. If cable operators and ISPs were both undertake a common objective of implementing third party access, and such access could be financially beneficial for all involved, then pricing and interconnection issues could certainly be resolved. Unfortunately, the experience in Canada does not bode well for a global negotiated solution to the third party access issue in the United States. In Canada, where the Canadian Radio-television and Telecommunications Commission (CRTC) has already mandated third party access, the cable and ISP industry association engaged in confidential negotiations in an effort to reach agreement on critical terms of third party access. These negotiations failed to bear fruit. It appears that a key unresolved issue was price–how much an ISP would pay a cable operator for access to the cable system. [ 133 ]
All of this does not by itself lead to the conclusion that third party access should not be ordered. Depending on the concerns discussed elsewhere, if it should be determined that a third party access requirement is desirable, then creative lawmakers and regulators could likely craft a workable process that minimized the disputes that have been seen in the local telephone competition proceedings. As noted, the third party access issue is simpler and narrower than all of the issues and disputes concerning local telephone competition. Although legal or regulatory disputes concerning third party access would not be wholly avoided, they could be focused fairly narrowly.
If a statutory or regulatory third party access scheme be crafted, it must include the following key points:
Except for the last provision, each of these elements could be hammered out in a regulatory drafting process. For the last provision, a policymaker would need to choose between asking the regulatory agency to articulate a pricing scheme, or creating a requirement that the market negotiate a price. Whatever approach to pricing were adopted, it is quite possible that legal or regulatory challenges would ensue. Those challenges, however, need not be as complex or involved as those seen in local telephone competition.
Opponents of mandatory third party access have expressed concern that an access requirement would delay broadband deployment. It is not clear that this delay would in fact happen, if the cable industry acts to take advantage of a possible significant head start in the broadband market.
In opposing third party access, the cable industry has asserted that a third party access requirement would significantly delay broadband deployment, in part because investment would be deterred (as discussed above), and in part because of the industry threat not to deploy broadband if third party access were ordered.
If a third party access requirement were seriously proposed or adopted, the cable industry would have two basic choices. It could (a) delay deployment while it fights the requirement, or it could (b) aggressively deploy the services while it fights the requirement. If the cable industry delays deployment and wins, then cable will have run the risk of allowing DSL service to catch up and perhaps overtake cable modem service. If the cable industry delays deployment and loses, then it will have allowed DSL to catch up and it will immediately face full third party competition. On the other hand, if the cable industry aggressively deploys during a fight over third party access, it becomes almost a win-win situation for the cable company. Either the cable industry defeats third party access (in which case it has neither delayed deployment nor lost ground against DSL), or the industry fails to defeat third party access (in which case the industry has an enormous head start over would be cable modem competitors).
The cable industry’s ability to achieve a very significant head start should not be understated. In Canada (where broadband deployment is more widespread), the cable industry aggressively deployed cable modem service while continuing to participate in the regulatory process that is moving to third party access. Most Canadian cable companies will be offering cable modem services to most of their customers long before third party access is actually implemented (sometime in 2000). The Canadian ISP industry has acknowledged the cable industry’s very significant head start. If the United States cable industry acts rationally, it will deploy even in the face of a possible third party access requirement.
Opponents of mandatory third party access have asserted that such a requirement would violate the constitutional rights of cable owners.
Although the question is far from easy or certain, it does not appear likely that a court would find that third party access to cable systems violates the constitutional rights of cable operators. A court would likely find that a third party access requirement does burden the speech of cable operators and/or forces speech on them. Thus, the requirement would then be viewed with standard of heightened scrutiny. The requirement, however, is content-neutral, and thus the appropriate level of scrutiny would be intermediate (as opposed to strict) scrutiny. If third party access is undertaken by Congress or the FCC, a court would likely find that promotion of competition and diversity in the area of broadband access is an important governmental interest, and third party access promotes that interest without burdening any more speech than is essential. A court would thus likely conclude that third party access survives constitutional scrutiny.
This conclusion is consistent with the courts' upholding of video "must carry" and "leased access" provisions of the Telecommunications Act of 1996. In that Act, Congress required cable operators to carry certain local television stations channels (for "must carry") or independently produced video ("leased access"). In the Turner Broadcasting System, Inc. v. FCC series of cases, the Supreme Court and the U.S. Court of Appeals for the D.C. Circuit upheld those provisions using an analysis similar to that discussed above. [ 134 ] Third party access presents no greater constitutional issue than was presented in Turner, [ 135 ] and thus third party access would likely be upheld.
Openness has been fundamental to the narrowband Internet's free-speech and democracy-enhancing character. As the Itnernet hsifts from narrowband architecture to broadband technologies, it is critical that openness is maintained. This Broadband Backgrounder, a factual primer and analysis of the issues, finds that openness is feasible. Indeed, as the public debate has evolved, it is clear that there is no longer a question of whether open access if feasible or desirable. Whhat remains to be decided, and the discussion this paper seeks to inform, is how openness is defined, and how it best can be achieved.
Endnotes
[I-1]. American Civil Liberties Union v. Reno, 929 F. Supp. at 882.[I-2]. The questions of what criteria determine openness and what mechanisms might assure that it is achieved are addressed in a separate document CDT is filing with the Federal Communications Commission in response to its September 28, 2000 Notice of Inquiry. See, In the Matter of Inquiry Concerning High-Speed Access to the Internet Over Cable and Other Facilities, GN Docket No. 00-185, Comments of the Center for Democracy and Technology, December 1, 2000.
[1]. The history and development of the Internet is summarized in "A Brief History of the Internet," written by nine of the early architects of the network, Barry Leiner, Vinton Cerf, David Clark, Robert Kahn, Leonard Kleinrock, Daniel Lynch, Jon Postel, Larry Roberts, and Stephen Wolff. The article can be found at http://info.isoc.org/internet/history/brief.html.
[2]. A timeline detailing the initial development of the World Wide Web can be found at http://www.w3.org/History.html.
[3]. Still today, the vast majority of voice telephone calls are circuit switched.
[4]. See Vint Cerf, "How the Internet Really Works - A Modest Analogy," http://www.wcom.com/about_the_company/cerfs_up/prose/hownetworks.shtml.
[5]. American Civil Liberties Union v. Reno, 929 F. Supp. 824, 877 (E.D. Pa. 1996) (Dalzell concurring) (hereafter Reno District Court Opinion"), http://www.ciec.org/victory.shtml.
[6]. Id. at 881 (Dalzell concurring).
[7]. Reno v. American Civil Liberties Union, 521 U.S. 844, 853 (1997), http://www.ciec.org/SC_appeal/decision.shtml.
[8]. "The Unregulation of the Internet: Laying a Competitive Course for the Future," Remarks by Chairman Kennard Before the Federal Communications Bar, Northern California Chapter, San Francisco, July 20, 1999, http://www.fcc.gov/Speeches/Kennard/spwek924.html.
[9]. "Internet Prospers with 'Hands-off Unregulation'; FCC Paper Rejects Need for Precipitous Action," Press Release, Federal Communications Commission, July 19, 1999, http://www.fcc.gov/Bureaus/OPP/News_Releases/1999/nrop9004.html. The analysis was prepared by Jason Oxman, then of the FCC's Office of Plans and Policy. "The FCC and the Unregulation of the Internet," Working Paper #31, Jason Oxman, Counsel for Advanced Communications, Office of Plans and Policy, July 19, 1999, http://www.fcc.gov/Bureaus/OPP/working_papers/oppwp31.pdf.
[10]. In the Matter of Regulatory and Policy Problems Presented by the Interdependence of Computer and Communication Services and Facilities, 7 FCC 2d 11 (1966) ("First Computer Inquiry").
[11]. First Computer Inquiry, Tentative Decision, 28 FCC 2d 291, at ¶ 20 (1970).
[12]. First Computer Inquiry, Final Decision and Order, 28 FCC 2d 267, at ¶¶ 11-12 (1971).
[13]. In the Matter of Use of the Carterfone Device in Message Toll Telephone Service, 13 FCC 2d 420 (1968).
[14]. 47 C.F.R. Part 68.
[15]. In the Matter of Amendment of Section 64.702 of the Commission's Rules and Regulations (Second Computer Inquiry), 77 FCC 2d 384, 419 (1980).
[16]. Moreover, the basic and enhanced terminology has essentially been supplanted by the "telecommunications" and "information services" categories used in the Telecommunications Act of 1996. See 47 U.S.C. § 153. The exact scope of those new terms has yet to be fully determined and litigated.
[17]. If the user happens to use a national backbone provider as a local ISP, then steps 3-7 of the table might not occur.
[18]. For a good and concise overview of broadband basics, see "Broadband Today," Staff Report by the Cable Services Bureau of the FCC, October 1999, http://www.fcc.gov/Bureaus/Cable/Reports/broadbandtoday.pdf. For an excellent and in depth - but still quite accessible - analysis of the broadband technologies and the broadband market, see Kim Maxwell, "Residential Broadband: An Insider's Guide to the Battle for the Last Mile," John Wiley & Sons, 1999. Maxwell's book contains a thorough review of the potential and limitations of the technologies, and has provided the foundation for much of the discussion in this section.
[19]. In the early days of modems, modem speeds were often quantified in terms of a "baud rate." A baud rate is a measure of how quickly a modem can transfer certain data "states." At low speeds (2400 bps or lower), the baud rate and bps were usually the same (and were often used interchangeably). Modem technology evolved, however, so that more bits per second could be transmitted at a given baud rate, and a typical 9600 bps modem, for example, still operated at 2400 baud. For the sake of simplicity, "baud" should no longer be used to quantify modem speeds.
[20]. As a general matter, because of limitations in the copper telephone wires, 56 kbps modems in use in homes can receive data at a rate of 56 kbps, but can only transmit data at 33 kbps.
[21]. Inquiry Concerning the Deployment of Advanced Telecommunications Capability to All Americans in a Reasonable and Timely Fashion, and Possible Steps To Accelerate Such Deployment Pursuant to Section 706 of the Telecommunications Act of 1996, CC Docket No. 98-146, Report, 14 FCC Rcd. 2398, 2406 (1999).
[22]. Id.
[23]. "'Always-on' will drive broadband," ZDNet News, Mar. 8, 1999, available at http://www.zdnet.com/zdnn/stories/news/0,4586,2222039,00.html.
[24]. Kim Maxwell, "Residential Broadband: An Insider's Guide to the Battle for the Last Mile," John Wiley & Sons, 1999.
[25]. Maxwell closely considers each potential application in terms of its potential to contributing to the cost of implementing broadband services. He concludes that in the near term few of the applications outside of the professional area will generate enough income to pay for broadband deployment. Any public policy efforts to finance broadband deployment should carefully consider Maxwell's assessments.
[26]. None of these applications require a huge upstream data path, and the applications could be likely supported with as little as 100 kbps of upstream bandwidth. Upstream bandwidth is more critical for companies or individuals who want to run a web or other type of server from their computer - something that is clearly not contemplated by most of the current broadband access vendors.
[27]. Ironically, although many consider film movies to be of a higher quality than video, movies are shown with fewer frames per second than video, and thus movies can be stored and transmitted in less space, or at a lower bandwidth, than video. Of particular difficulty is live sports coverage, which often have many sudden movement that causes problems for video compression techniques.
[28]. According to a definition offered by the cable industries leading trade association, the National Cable Television Association, HFC stands for a "network architecture developed by the cable industry which uses a blend of fiber and coaxial cable to bring consumers interactivity, greater channel capacity, increased signal strength, and improved reliability." http://ncta.cyberserv.com/qs/user_pages/glossary.cfm. All of these features are valuable to a cable operator independent of the capability to support Internet access.
[29]. "Speed Traps for Cable Modems?," Washington Post (Apr. 14, 2000), at E01.
[30]. Id.
[31]. "Cable Modem Security: What you don't know ü.", PC World News Radio, Oct. 30, 1997, available at http://www.pcworld.com/news/daily/data/1097/971030164448.html. For a general overview of the cable modem security issue, see http://www.cable-modem.net/gc/security.html.
[32]. "Broadband Today," Staff Report to FCC Chairman William Kennard, October 1999, at 26, http://www.fcc.gov/Bureaus/Cable/Reports/broadbandtoday.pdf.
[33]. Id. Cable modem subscribers on Time Warner cable systems alone rose from 186,000 as of June 30, 1999, to 447,000 as of March 31, 2000. Compare id. with "Time Warner Nears 500K Modem Sub. Mark," Broadband Daily, April 13, 2000, http://www.broadband-daily.com/subscribers/index.htm?article_id=1209.
[34]. A good overview of DSL technology can be found at Robyn Aber, "xDSL Local Loop Access Technology: Delivering Broadband over Copper Wires," http://www.3com.com/technology/tech_net/white_papers/500624.html. See also "General Introduction to Copper Access Technologies," http://www.adsl.com/general_tutorial.html.
[35]. The asymmetry of ADSL ultimately is a function of the interference that can arise when many symmetric electronic signals are sent over a large bundle of wires. Sending symmetric DSL signals would reduce the data rate on the lines. See "General Introduction to Copper Access Technologies," http://www.adsl.com/general_tutorial.html.
[36]. See id. The faster the speed of ADSL, the shorter the length of wire on which it can be supported.
[37]. See id. Some business-oriented DSL services, such as IDSL, can be offered at distances greater than 18,000 feet, but at significantly lower speeds than ADSL. In addition, some companies are testing technology that has the potential to raise the 18,000-foot limit.
[38]. Most of the incumbent local exchange carriers use an ATM network to transport data traffic from the central offices to the ISP. See http://www.xdsl.com/library/matrix_sheets/Service_071999.xls.
[39]. Some assert that the ILECs did not want to erode their business market for T1 lines, and thus the ILECs were slow to push ADSL until cable companies started making serious in-roads into broadband.
[40]. This is comparable to cable modem service, which is not typically offered on a guaranteed basis. Because of the lack of guarantees, neither cable modem service or consumer oriented DSL service is well suited to businesses that depend on Internet connectivity.
[41]. For a matrix of prices and speeds of DSL offerings by ILECs, prepared by market strategy firm Telechoice, see http://www.xdsl.com/library/matrix_sheets/Service_071999.xls.
[42]. For a comparison of the residential versus business deployment of ILECs and CLECs prepared by Telechoice, see http://www.xdsl.com/content/resources/deployment_info.asp.
[43]. See http://www.xdsl.com/content/resources/deployment_info.asp. These deployment figures do not include the non-Internet-focused use of HDSL as a T1 substitute.
[44]. See id.
[45]. See "Bell Atlantic, Prodigy in DSL Deal," Broadband Daily, May 25, 1999.
[46]. Indeed, one part of MCI WorldCom, a prime contender in today's marketplace for wireless broadband services, was originally named "Microwave Communications, Inc." before its name was changed to MCI, Inc.
[47]. See Report & Order, In the Matter of Amendment of Parts 21 and 74 to Enable Multipoint Distribution Service and Instructional Television Fixed Service Licensees to Engage in Fixed Two-Way Transmissions, MM Docket No. 97-217 (Sept. 25, 1998), http://www.fcc.gov/Bureaus/Mass_Media/Orders/1998/fcc98231.pdf.
[48]. WorldCom, a leading MMDS provider, directly credits the 1998 FCC decision with making MMDS a more viable offering. "Interpreting the Changing MMDS Landscape," WorldCom, Inc.,, at 11, http://www.wcom.com/about_the_company/mmds_landscape/mmds_briefbook.pdf.
[49]. For a helpful explanation of how a "telco return" system works, see "How does it work?," http://www.direcpc.com/consumer/work/work.html.
[50]. See Washington Internet Daily, Sept. 22, 2000, at 6; "Hughes Network Systems Announces Upcoming Two-Way DirecPC Satellite Internet Access," Apr. 27, 2000, http://www.hns.com/news/pressrel/csp_pres/p042700.htm.
[51]. "Hughes Network Systems Announces Upcoming Two-Way DirecPC Satellite Internet Access," Apr. 27, 2000, http://www.hns.com/news/pressrel/csp_pres/p042700.htm.
[52]. See "Answers to frequently asked questions," http://isky.net/qa/face.html.
[53]. See "Bruised Teledesic at Brink: Tech Firm's 'Net-in-sky' plan depends on new investors, secret recovery strategy," USA Today, Apr. 11, 2000, http://www.usatoday.com/life/cyber/invest/in567.htm.
[54]. Id.
[55]. Id.
[56]. The current state of "fiber to the home" technology and deployment is discussed at length in "Fiber Optics to the Home," Technology Review, March/April 2000, http://www.techreview.com/articles/ma00/hecht.htm.
[57]. See id. (discussing these and other issues).
[58]. "British Net project loses power," CNET News, May 29, 1998, http://cnet.com/news/0-1004-200-329713.html.
[59]. See "Internet access over power lines nears reality," CNET News, March 6, 2000, http://cnet.com/news/0-1004-202-1564871.html.
[60]. Memorandum Opinion and Order, In the Matter of Internet Ventures, Inc., File No. CSR-5407-L (Feb. 18, 2000), http://www.fcc.gov/Bureaus/Cable/Orders/2000/fcc00037.doc.
[61]. "Verizon and NorthPoint to Merge DSL Businesses to Create Leading National Broadband Company," Press Release, Aug. 8, 2000, at http://www.northpointdsl.com/about_press.asp?PressReleaseID=1152.
[62]. "Covad and SBC Form Marketing Agreement to Deliver Broadband Nationwide," Press Release, Sept. 11, 2000, http://www.covad.com/companyinfo/pressreleases/pr_2000/091100a_press.shtml.
[63]. First Report & Order, Deployment of Wireline Services Offering Advanced Telecommunications Capability, FCC No. 99-48, CC Docket No. 98-147, 14 FCC Rcd 4761 (1999), http://www.fcc.gov/Bureaus/Common_Carrier/Orders/1999/fcc99048.pdf.
[64]. GTE Service Corp. v. Federal Communications Comm., 205 F.3d 416 (D.C. Cir. 2000), http://pacer.cadc.uscourts.gov/common/opinions/200003/99-1176a.txt.
[65]. See, e.g., Reply Comments of the Texas Internet Service Providers Association, In the Matter of Application of SBC Communications ü for Provision of In-Region, InterLATA Services in Texas, CC Docket No. 00-4 (filed Feb. 20, 2000), at 10-11; The Big DSL Squeeze, ZDNet, March 27, 2000, http://www.zdnet.com/filters/printerfriendly/0,6061,2473709-2,00.html.
[66]. http://www.broadband-daily.com/subscription/index.htm?issue_rec=232&&article_rec=1198.
[67]. See "Summary of Conditions, Sept. 6, 1999, http://www.fcc.gov/Bureaus/Common_Carrier/News_Releases/1999/nrc9077a.html; "Conditions for Bell Atlantic/GTE Merger," June 17, 2000, http://www.fcc.gov/Bureaus/Common_Carrier/Orders/2000/fcc00221b.pdf.
[68]. See News Release, Pennsylvania Public Utilities Commission, Aug. 26, 1999, http://puc.paonline.com/telecomm/8-26-99%20puc%20passes%20rules%20jumpstart%20comp.htm.
[69]. See"Internet Freedom and Broadband Deployment Act," H.R. 2420, http://techlawjournal.com/cong106/broadband/hr2420ih.htm.
[70]. The Broadband Access Project has not attempted to evaluate the local telephone competition issues raised here.
[71]. Jerry Berman and John B. Morris, Jr., "The Broadband Internet: The End of the Equal Voice?," April 2000, http://www.cfp2000.org/papers/morrisberman.pdf.
[72]. Brief of the Federal Communications Commission as Amicus Curiae, AT&T Corp. v. Portland, No. 99-35609 (9th Cir.) (filed Aug. 16, 1999), http://techlawjournal.com/courts/portland/19990816fcc.htm.
[73]. AT&T Corp. v. City of Portland, 216 F.3d 871 (9th Cir. 2000), http://techlawjournal.com/courts/portland/20000622op.htm.
[74]. Id. at 878.
[75]. See"Federal Communications Commission Notice of Inquiry FCC 00-355", Adopted September 28, 2000. http://www.fcc.gov/Bureaus/Miscellaneous/Notices/2000/fcc00355.pdf.
[76]. SeeRegulation of Broadcasting Distribution Undertakings that Provide Non-Programming Services, No. 96-1, Canadian Radiotelecommunications & Telephone Commission, Jan. 30, 1999, http://www.crtc.gc.ca/ENG/telecom/decision/1996/d961_0.txt.
[77]. See, e.g., AT&T's and TCI's Joint Reply To Comments And Joint Opposition To Petitions To Deny Or To Impost Conditions, Application of AT&T and Tele-Communications, Inc., FCC Docket No. 98-178 (filed Nov. 16, 1998), at 40-43.
[78]. AT&T's letter is available at http://techlawjournal.com/broadband/19991206let.htm.
[79]. Comcast and Cox are the third and fourth largest cable companies in the U.S. behind AT&T and Time Warner.
[80]. "Comcast Quietly Pledges to Open Cable to Rival ISPs," Communications Daily (Mar. 24, 2000).
[81]. See Inquiry Concerning the Deployment of Advanced Telecommunications Capability to All Americans in a Reasonable and Timely Fashion, and Possible Steps To Accelerate Such Deployment Pursuant to Section 706 of the Telecommunications Act of 1996, CC Docket No. 98-146, 14 FCC Rcd. 2398, 2415-24 (1999).
[82]. Id. at 2449.
[83]. Applications for Consent to the Transfer of Control of Licenses and Section 214 Authorizations from Tele-Communications, Inc. to AT&T Corp., CS Docket No. 98-178, Memorandum and Order, 14 FCC Rcd. 3160, 3207 (1999).
[84]. "FCC Cable Chief Says "Open" Internet is Primary Goal; Cites Agreement of Consumers and Industry," Press Release, Nov. 9, 1999, http://www.fcc.gov/Bureaus/Cable/News_Releases/1999/nrcb9018.html.
[85]. Id.
[86]. "FCC Chairman to Launch Proceeding on 'Cable Access,'" Press Release, June 30, 2000, http://www.fcc.gov/Bureaus/Cable/News_Releases/2000/nrcb0017.html.
[87]. Statement of Chairman William E. Kennard Concerning Notice of Inquiry into High-Speed Internet Service, Sept. 28, 2000, http://www.fcc.gov/Speeches/Kennard/Statements/2000/stwek077.html.
[88]. See, e.g. "Open Cable Requirement Dies in Florida," CNET News (Oct. 19, 1999), http://news.cnet.com/news/0-1004-200-920007.html?tag=st.ne.1002.
[89]. See, e.g. ", "AT&T wins San Francisco cable access showdown," CNET News (July 26, 1999), http://news.cnet.com/news/0-1004-200-345360.html?tag=st.ne.100.
[90]. See, e.g. ", "Virginia city supports open access," CNET News (Sept. 29, 1999), http://news.cnet.com/news/0-1004-200-367136.html?tag=st.ne.1002; "Harmon signs bill making AT&T cable open up Internet service," St. Louis Post Dispatch (Nov. 13, 1999),
[91]. See, e.g. ", "AT&T sues over Florida open access ruling," CNET News ( July 26, 1999), http://news.cnet.com/news/0-1004-200-345168.html?tag=st.ne.1002.
[92]. See, e.g. ", "Judge dismisses parts of Comcast lawsuit," CNET News (Feb. 15, 2000), http://news.cnet.com/news/0-1004-200-1550612.html?tag=st.ne.1002.
[93]. See, e.g. ", "Voters may sound off on high-speed Net plans," CNET News (Aug. 6, 1999), http://news.cnet.com/news/0-1004-200-345792.html?tag=st.ne.1002.
[94]. See Telecom Decision CRTC 96-1, Jan. 30, 1996.
[95]. See Telecom Decision CRTC 98-9, July 9,. 1998.
[96]. See Anatel Gives Green Light for ISP via Cable TV, South American Business Information, Nov. 26, 1999, http://library.northernlight.com/FC19991126650000011.html?cb=0&dx=1004&sc=0#doc.
[97]. For a good overview of the history of cable regulation, see James C. Goodale, "All About Cable" § 1.03 (1998).
[98]. Frontier Broadcasting v. Collier, 16 Rad. Reg. 1005 (FCC 1958).
[99]. See Notice of Inquiry and Notice of Proposed Rulemaking, 1 F.C.C.2d 453 (1965).
[100]. Cable Communications Policy Act of 1984, Pub. L. No. 98-549, 98 Stat. 2779 (codified in scattered sections of 47 U.S.C.).
[101]. See Goodale § 1.15.
[102]. Pub. L. No. 102-385, 106 Stat. 1460 (1992) (codified in scattered sections of 47 U.S.C.).
[103]. For a good overview of the provisions of the 1992 Cable Act, see Goodale § 1.15[3].
[104]. Pub. L. No. 102-385 § 2(a).
[105]. Id. § 2(b).
[106]. Id. § 11(c) (codified at 47 U.S.C. § 613(f)(2), but later amended).
[107]. Id. § 19 (codified at 47 U.S.C. § 628(c)(4), but later amended).
[108]. Pub. L. No. 104-104, 110 Stat. 56 (1996).
[109]. Based on interviews with representatives of the Canadian cable industry, it appears that the prospect of extensive cost-based regulatory proceedings in Canada has caused significant concern in the industry.
[110]. See "AT&T Sets First Cable Technical Trial of Multiple ISPs," Washington Internet Daily, June 8, 2000, at 4.
[111]. These and other questions are currently under active discussion between the cable operators and the ISPs in Canada.
[112]. The possibility of a small number of users using massive amounts of bandwidth is not theoretical. The music exchange program "Napster" is an example. Just as some colleges have banned the used of Napster on their networks, at least one cable system has also banned Napster because of the high amount of upstream bandwidth used by the program. See "Napster Not At Home With Cable," Wired News, Apr. 7, 2000, http://www.wired.com/news/print/0,1294,35523,00.html.
[113]. The cable operator or cable modem ISP may well offer customers space on a web server that is not directly connected to the shared resource of the cable facility. Such approach would allow the customers to create a web site while reducing the customers' need to host the web site themselves.
[114]. http://www.nogatekeepers.org/learnmore/faq.shtml.
[115]. "Excite to offer high-speed Net services," USA Today, Apr. 12, 2000, p.6B, http://www.usatoday.com/life/cyber/tech/review/crh051.htm. The reference to the 10 minute limit on cable systems was made in the context of discussing Excite@Home's contract with Rhythms NetConnection to offer the @Home service over DSL lines. The 10-minute limit will not be applied to Excite@Home's DSL service. Id.
[116]. http://www.home.com/support/aup/.
[117]. Memorandum of Understanding Between Time Warner, Inc. and America Online, Inc. Regarding Open Access Business Practices, Feb. 29, 2000, para. 6, http://media.web.aol.com/media/press_view.cfm?release_num=25100400&title=Memorandum%20of%20Understanding%20Between%20Time%20Warner%2C%20Inc%2E%20and%20America%20Online%2C%20Inc%2E
[118]. It is certainly possible that one broadband technology will prove to be better suited for certain applications, but for general "surfing" of the World Wide Web, both cable modem and DSL service offer dramatically faster access speeds than are available over a dial-up connection.
[119]. As discussed above, the expected "service availability date" of the Teledesic system has slipped from 1999 until 2004.
[120]. "The Developing Residential Broadband Gap," The Precursor Group, Legg Mason Wood Walker, Inc., February 8, 2000. Scott Cleland and the Precursor Group are viewed by some as slanted against the cable industry in the third party access debate (and certainly, many of the Group's analyses have been critical of cable's refusal to open their pipes to multiple ISPs). Cleland's data, nevertheless, represents a reasonable and realistic assessment of the likely deployment of broadband access services.
[121]. Jerry Berman and John B. Morris, Jr., "The Broadband Internet: The End of the Equal Voice?," April 2000, http://www.cfp2000.org/papers/morrisberman.pdf.
[122]. As discussed above, one of the concerns that led to the 1992 Cable Act was the perception that the cable industry was charging too much for video services as a result of a lack of competition.
[123]. For example, @Home service in the United States is typically priced at $39.95-44.95, http://www.home.com/pricing.html, while the same service in Canada is priced at $39.95 Canadian dollars, http://rogers.home.com/Value.html, or about $27.00 in U.S. dollars.
[124]. http://www.nogatekeepers.org/learnmore/faq.shtml.
[125]. An ISP that, hypothetically, has supported 10,000 dial-up users at 56 kbps speeds cannot simply shift 2,000 of those users over to a broadband connection operating at 1.5 Mbps. The hypothetical ISP would almost certainly have to increase significantly the speed of its connection back to the Internet backbone in order to support the speed demands of the new broadband customers. Not all ISPs will choose to make that investment.
[126]. Walter Ciciora, "Cable Television in the United States - An Overview," at 47 (CableLabs 1995) http://cablelabs.com/about_cl/pubs/CATV.pdf.
[127]. Id.
[128]. Id. at 29. Ironically, the CableLabs overview expressed specific doubt that Internet services would be a practical application to be offered by cable systems. Id. at 58, 62-63.
[129]. See, e.g., "Time Warner deal pushes interactive TV forward," CNET News, http://news.cnet.com/news/0-1006-202-1454162.html (viewed Nov. 22, 1999).
[130]. Id. at 54.
[131]. The two-way capability would be necessary for selecting movies on demand, and for other limited interactivity available over a cable system.
[132]. "Rogers Communications and Microsoft Announce Agreements To Develop and Deploy Advanced Broadband Television Services in Canada," Press Release, http://www.microsoft.com/PressPass/press/1999/Jul99/RogersCommPr.asp.
[133]. Ironically, the fact that retail prices for cable modem service in the United States are generally higher than those in Canada may ultimately permit industry players in the U.S. to reach agreement where their counterparts in Canada could not. A higher retail price may give the parties more negotiating room in which to find a mutually beneficial price.
[134]. See Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622 (1994) (Turner I); Turner Broadcasting System, Inc. v. FCC, 520 U.S. 180 (1997) (Turner II).
[135]. In fact, in the Internet context it is generally quite clear that the user controls what content is delivered, and thus there is even less chance than in Turner that someone will assume that the Internet content reflects the opinions and views of the cable system operator.