Broadband Connectivity Competition Policy: FTC Staff Report June

b r o a d b a n d

connect v ty

compet t on

pol cy

ftc staff report

June 2007

Federal Trade Commission

DEBORAH PLATT MAJORAS Chairman

PAMELA JONES HARBOUR Commissioner

JON LEIBOWITZ Commissioner

WILLIAM E. KOVACIC Commissioner

J. THOMAS ROSCH Commissioner

Brian Huseman Chief of Staff

Charles H. Schneider Executive Director

Jeffrey Schmidt Director, Bureau of Competition

Lydia B. Parnes Director, Bureau of Consumer Protection

Michael A. Salinger Director, Bureau of Economics

Maureen K. Ohlhausen Director, Office of Policy Planning

Jeanne Bumpus Director, Office of Congressional Relations

Randolph W. Tritell Director, Office of International Affairs

Nancy Ness Judy Director, Office of Public Affairs

William Blumenthal General Counsel

Donald S. Clark Secretary of the Commission

Report Drafters and Contributors:

Gregory P. Luib, Assistant Director, Office of Policy Planning

Daniel J. Gilman, Attorney Advisor, Office of Policy Planning

Christopher M. Grengs, Attorney Advisor, Office of Policy Planning

James F. Mongoven, Deputy Assistant Director, Bureau of Competition

Armando Irizarry, Bureau of Competition

Mary Beth Richards, Deputy Director, Bureau of Consumer Protection

Elizabeth A. Hone, Assistant Director, Bureau of Consumer Protection

Robert Schoshinski, Bureau of Consumer Protection

Denis A. Breen, Assistant Director, Bureau of Economics

Patrick J. DeGraba, Bureau of Economics

Rachel Miller Dawson, Assistant General Counsel for Legal Counsel, Office of General Counsel

Lisa M. Harrison, Office of General Counsel

Robert B. Mahini, Office of General Counsel

Kim Vandecar, Congressional Specialist, Office of Congressional Relations

Katherine Rosenberg, Office of Congressional Relations

Yael Weinman, Office of International Affairs

Inquiries concerning this Report should be directed to:

Maureen K. Ohlhausen, Director, Office of Policy Planning

(202) 326-2632 or [email protected]

This Report represents the views of the FTC staff and does not necessarily represent the views of

the Commission or any individual Commissioner. The Commission, however, has voted to

authorize the staff to issue this Report.

TABLE OF CONTENTS

INTRODUCTION AND EXECUTIVE SUMMARY.................................................................................... 1

I: THE INTERNET: HISTORICAL AND TECHNICAL BACKGROUND............................................. 13

A. H

ISTORICAL DEVELOPMENT.......................................................................................................... 13

B. MAJOR INTERNET COMPONENTS................................................................................................... 23

C. NETWORK MANAGEMENT, DATA PRIORITIZATION, AND OTHER FORMS OF DATA DISCRIMINATION 28

II: LEGAL AND REGULATORY BACKGROUND AND DEVELOPMENTS....................................... 37

A. FTC J

URISDICTION UNDER THE FTC ACT..................................................................................... 38

B. FCC J

URISDICTION UNDER THE COMMUNICATIONS ACT.............................................................. 42

C. R

EGULATORY AND JUDICIAL CLARIFICATION............................................................................... 43

III: ARGUMENTS IN FAVOR OF AND AGAINST NETWORK NEUTRALITY REGULATION..... 51

A. A

RGUMENTS IN FAVOR OF NETWORK NEUTRALITY REGULATION................................................ 52

B. ARGUMENTS AGAINST NETWORK NEUTRALITY REGULATION...................................................... 60

IV: DISCRIMINATION, BLOCKAGE, AND VERTICAL INTEGRATION.......................................... 70

A. L

AST-MILE ACCESS CONCERNS CONTINGENT ON MARKET POWER.............................................. 72

B. POTENTIAL PROBLEMS INDEPENDENT OF LAST-MILE MARKET POWER........................................ 76

C. POTENTIAL BENEFITS OF VERTICAL INTEGRATION....................................................................... 80

D. B

RIEF SUMMARY AND REMAINING QUESTIONS............................................................................. 82

V: DATA PRIORITIZATION........................................................................................................................ 83

A. W

HY PRIORITIZE DATA?................................................................................................................ 84

B. P

RIORITIZATION VERSUS CAPACITY EXPANSION........................................................................... 86

C. TYPES AND USES OF DATA PRIORITIZATION.................................................................................. 88

D. C

ONCLUSION.................................................................................................................................. 96

VI: THE CURRENT AND FUTURE STATE OF BROADBAND COMPETITION............................... 98

A. H

ISTORICAL BACKGROUND: DIAL-UP SERVICE............................................................................. 98

B. V

IEWS ON THE STATE OF BROADBAND COMPETITION................................................................... 99

C. M

UNICIPAL PROVISION OF WIRELESS INTERNET ACCESS.............................................................. 106

D. FEDERAL SPECTRUM POLICIES...................................................................................................... 109

E. I

NTERNATIONAL COMPARISONS.................................................................................................... 113

VII: ANTITRUST ANALYSIS OF POTENTIAL BROADBAND PROVIDER CONDUCT.................. 120

A. G

ENERAL PRINCIPLES UNDERLYING THE ANTITRUST LAWS......................................................... 120

B. P

OTENTIAL ANTITRUST THEORIES................................................................................................. 121

iii

VIII: CONSUMER PROTECTION ISSUES................................................................................................ 129

AN OVERVIEW OF SECTION 5 OF THE FTC ACT..................................................................................................... 129

B. APPLICABILITY OF CONSUMER PROTECTION LAWS TO BROADBAND INTERNET ACCESS SERVICES.. 130

C. ADDITIONAL MEASURES TO PROTECT CONSUMERS........................................................................................... 136

IX: PROPOSALS REGARDING BROADBAND CONNECTIVITY.........................................................138

A. E

XISTING AGENCY OVERSIGHT..................................................................................................... 138

B. FCC POLICY STATEMENT AND MERGER CONDITIONS.................................................................. 141

C. L

EGISLATIVE PROPOSALS.............................................................................................................. 145

D. OTHER PROPOSALS RELATING TO BROADBAND CONNECTIVITY................................................... 147

X: SUGGESTED GUIDING PRINCIPLES.................................................................................................. 155

A. C

OMPETITION IN BROADBAND INTERNET ACCESS SERVICES........................................................ 155

B. GROUNDS FOR PROCEEDING WITH CAUTION................................................................................. 157

C. C

ONTINUED AGENCY OVERSIGHT................................................................................................. 161

APPENDIX 1 –

BROADBAND CONNECTIVITY COMPETITION POLICY WORKSHOP PARTICPANTS 163

APPENDIX 2 – GLOSSARY OF FREQUENTLY USED ACRONYMS................................................... 165

INTRODUCTION AND EXECUTIVE SUMMARY

Background

The Internet

has profoundly impacted numerous aspects of daily life for many

people in the United States and is increasingly vital to the American economy. In

response to recent debate relating to Internet access issues, Federal Trade Commission

(“FTC” or “Commission”) Chairman Deborah Platt Majoras announced the formation of

the Internet Access Task Force (“Task Force”) in August 2006 and invited interested

parties to meet with the Task Force to discuss issues relating to Internet access generally

and net neutrality

in particular.

The Task Force held a two-day public workshop on

broadband connectivity competition policy in February 2007 (“Workshop”) to bring

together consumer advocates and experts from business, government, academia, and the

technology sector to explore competition and consumer protection issues relating to

broadband Internet access.

The purpose of this Report is to summarize the Task Force’s

learning on broadband Internet connectivity in general and network neutrality in

particular, as developed from the Workshop, meetings between the Task Force and

various interested parties, and the FTC staff’s independent research.

As discussed in more detail in Chapter I of this Report, the term “Internet” is commonly used to refer to

the decentralized, interconnected network of computer networks that allows computers to communicate

with each other. Individual networks are owned and administered by a variety of organizations, such as

private companies, universities, research labs, government agencies, and municipalities.

The terms “net neutrality” and “network neutrality” have been used to identify various policy concerns

and prescriptions raised by diverse parties to the larger social discussion of broadband Internet

connectivity. Typically, such terms are identified with positions that recommend, at least, some legal or

regulatory restrictions on broadband Internet access services that include non-discrimination requirements

above and beyond any that may be implied by existing antitrust law or Federal Communications

Commission (“FCC”) regulations. Particular concerns and positions are explored in some detail throughout

the Report, but the terms “net neutrality” and “network neutrality” are used here, interchangeably, to refer

to this larger family of views. Unless otherwise clarified, our terminological choice is not meant to endorse

any particular policy position.

See Deborah Platt Majoras, Chairman, FTC, Luncheon Address, The Progress & Freedom Foundation’s

Aspen Summit, The Federal Trade Commission in the Online World: Promoting Competition and

Protecting Consumers (Aug. 21, 2006), available at

http://ftc.gov/speeches/majoras/060821pffaspenfinal.pdf

The agenda, transcript, public comments, and other information relating to the Workshop are available on

the FTC’s Web site at http://www.ftc.gov/opp/workshops/broadband/index.shtm

. In addition, Appendix 1

to this Report provides the identity and affiliation of the Workshop participants.

Throughout this Report, citations to “Public Comments” refer to comments submitted to the FTC

in response to its request for public comments on the topics addressed at the Workshop. In addition,

citations to “Tr.” refer to the Workshop transcript, which is comprised of two volumes. Volume I

corresponds to the proceedings on February 13, 2007; Volume II corresponds to the proceedings on

February 14, 2007. Speakers are identified by last name. Finally, citations to “Participant Presentations”

refer to presentations, including slide presentations and commentary, provided by Workshop participants.

Originally, the Internet developed out of efforts by researchers at American

universities and the U.S. Department of Defense Research Projects Agency (“DARPA”)

in the 1960s and 1970s to create and test interconnected computer networks that would

communicate via data packet switching rather than traditional circuits. Today, the

Internet – which enables applications such as e-mail and browsers that search the World

Wide Web (the “Web”) – connects many millions of end users (and more than one

hundred million Web sites worldwide) to content, applications, and each other. End users

include the initial government and academic centers, corporate entities across all sectors

of the economy, and individuals and associations.

Individual end users (and networks of end users) arrange for Internet access via a

“last mile” connection to an Internet service provider (“ISP”),

which provides, in turn,

routing and connections from the ISP’s own network to the Internet. Content and

applications providers offer their products and services to end users via network

operators, which enable connectivity and transport into the middle, or “core,” of the

Internet. Before the turn of the century, most computer users connected to the Internet

using “narrowband,” dial-up telephone connections and modems to transmit data over the

telephone system’s traditional copper wirelines. Much faster “broadband” connections

recently have been deployed using various technologies, including coaxial cable

wirelines, upgraded copper digital subscriber lines (“DSL”), and to a lesser extent fiber-

optic wirelines, wireless, satellite, and broadband over powerlines (“BPL”).

Traditionally, data traffic has traversed the Internet on a “first-in-first-out” and

“best-efforts” basis. This protocol for data transmission was established principally as a

result of DARPA’s original priority, which was to develop an effective technique for

communications among existing interconnected networks, and which placed network

survivability – or the potential for robust network operation in the face of disruption or

infrastructure destruction – as the top goal in designing the overall architecture of this

network of networks. Since the Internet’s earliest days, however, computer scientists

have recognized that network resources are scarce and that traffic congestion can lead to

reduced performance. Although different data transmission protocols and the viability of

usage-based pricing mechanisms were explored throughout the 1980s and 1990s, the

debate over broadband connectivity policy did not reach critical mass until recently.

Technical, business, legal, and regulatory developments all appear to have contributed to

the acceleration of the discussion.

Regulatory jurisdiction over broadband services generally is subject to the shared

jurisdiction of the FCC, the FTC, and the Department of Justice (“DOJ”).

FCC

jurisdiction comes chiefly from the Communications Act of 1934, as amended

(“Communications Act”).

FTC jurisdiction over broadband arises chiefly under its

Appendix 2 to this Report provides a glossary of acronyms that are frequently used herein.

In this Report, we also refer to broadband ISPs as “broadband providers” and “access providers.”

See infra Chapters II and IX.A for discussion of various jurisdictional issues.

47 U.S.C. §§ 151 et seq.

statutory mandate to prevent “unfair methods of competition” and “unfair or deceptive

acts or practices in or affecting commerce” under the FTC Act.

The FTC’s authority to

enforce the federal antitrust laws generally is shared with DOJ’s Antitrust Division. The

FCC, FTC, and DOJ have exercised their existing authority in various ways. All three

agencies have scrutinized proposed mergers in Internet-related markets and have

negotiated significant conditions on certain mergers allowed to go forward.

In addition,

the FTC has enforced the consumer protection laws, bringing a variety of cases against

Internet service providers that have engaged in allegedly deceptive marketing and billing

practices.

Certain judicial and regulatory decisions in recent years have clarified the scope

of broadband regulation in two fundamental regards. First, since about 2000, the FCC

has undertaken a substantial and systematic deregulation of broadband services and

facilities, concluding that cable, wireline, powerline, and wireless broadband Internet

access services are “information services” that are not subject to common carrier

requirements.

The first of these decisions was sustained by the Supreme Court in

National Cable & Telecommunications Association v. Brand X Internet Services.

Second, these decisions have served to reinforce and expand FTC jurisdiction

over broadband Internet access services. That jurisdiction had once been regarded as

limited to the extent that the FTC’s general enforcement authority under the FTC Act did

not extend to entities that were “common carriers” under the Communications Act. The

regulatory and judicial decisions at issue, however, confirmed that the larger categories of

broadband Internet access services, as information services, are not exempt from FTC

enforcement of the FTC Act.

In recent years, changes in both user demand and technology have prompted some

broadband providers openly to consider prioritizing certain data traffic to improve

network management and provide premium services. The demand for bandwidth has

increased dramatically, as a growing number of users seek access to increasingly data-

rich Internet content, such as streaming video, which often requires considerable

bandwidth or has particular quality-of-service requirements. That demand has prompted

15 U.S.C. §§ 41 et seq.

See, e.g., Am. Online, Inc. & Time Warner, Inc., FTC Dkt. No. C-3989 (Dec. 17, 2000) (complaint),

available at http://www.ftc.gov/os/2000/12/aolcomplaint.pdf. See infra Chapters II and IX for discussion

of FCC, FTC, and DOJ scrutiny of mergers in the area of broadband Internet access.

See, e.g., Am. Online, Inc. & CompuServe Interactive Servs., Inc., FTC Dkt. No. C-4105 (Jan. 28, 2004)

(decision and order), available at http://www.ftc.gov/os/caselist/0023000/040203aolcsdo.pdf; Juno Online

Servs., Inc., FTC Dkt. No. C-4016 (June 29, 2001) (decision and order), available at

http://www.ftc.gov/os/2001/06/junodo.pdf

Particular rulemaking and other administrative decisions along these lines are discussed in more detail in

Chapters II and IX, infra.

545 U.S. 967 (2005).

concern about present and future congestion and about the need for further infrastructure

investment and development. At the same time, technological developments have made

feasible differentiation in delivery of data of various types, or from various sources,

based on payment to or affiliation with a network operator.

In response, various interested parties, including some content and applications

providers and commentators, have expressed concern about network operators’ use of

these technologies in an environment that is not subject to common carrier regulations.

Some of these providers and commentators, therefore, have proposed that the

transmission of data on the Internet be subject to some type of “net neutrality” regulation

that forbids or places restraints on some types of data or price discrimination by network

operators. Opponents of net neutrality regulation assert that it is not just unnecessary, but

potentially harmful, and that allowing network operators to innovate freely across

technical and business contexts, and to differentiate their networks, will lead to enhanced

service offerings for both end users and content and applications providers.

Before turning to the policy discussion that follows, it is worth clarifying that this

Report reflects the views of the staff of an agency that enforces the federal antitrust and

consumer protection laws. The statutory mission of the FTC is to protect both

competition and consumers by safeguarding and encouraging the proper operation of the

free market. In carrying out that mission, the FTC primarily is focused on maximizing

consumer welfare, as that term is defined in an economic sense in modern antitrust and

consumer protection jurisprudence. We recognize that preserving the diversity of views

expressed on the Internet is one of the animating principles of many of the most ardent

proponents of network neutrality. In this Report, however, we do not attempt to balance

consumer welfare (as we use it, in the economic sense) and free expression.

Instead,

the Report focuses on the consumer welfare implications of enacting some form of net

neutrality regulation.

Further, although the goal of increasing competition in broadband Internet access

is fundamental to the FTC staff’s interest and may be widely shared, how best to achieve

that goal is a point of sharp disagreement. What the FTC can offer in this debate is an

explanation of which behavior the antitrust and consumer protection laws already

proscribe and a framework for analyzing which conduct may foster or impede

competition in particular circumstances.

The Report is organized as follows. Chapter I provides technical information on

the functioning of the Internet, and Chapter II provides background information on the

See, e.g., Mercatus Center, Public Comment 27, at 10 (“If the desired outcome is that anyone willing to

pay the monthly price for Internet access can communicate with others at some minimum speed, then a

policy that promotes ‘neutral’ treatment of everyone on the network may be appropriate. But if the desired

outcome is to have as many people as possible connected to the Internet so they can speak if they so

choose, then a different policy, aimed at reducing the consumer’s total cost of Internet access as well as

usage, may be most effective, even if it does not mandate ‘neutrality.’”); Feld, Tr. II at 75 (“It is a question

about balancing. . . . I can say that something does introduce a certain amount of economic inefficiency

and it is still extraordinarily valuable for the contribution that it gives to us as a society, as a democracy . . .

. I would argue that is something we should be willing to consider.”).

legal and regulatory developments that have fueled the debate over net neutrality

regulation. The purpose of these Chapters is to inform the subsequent policy discussion.

Chapter III identifies and briefly describes the various arguments for and against net

neutrality regulation that have been put forth to date. Chapter IV analyzes potential

conduct by ISPs and other network operators, including vertical integration into content

and applications and discrimination against non-affiliated providers of content and

applications. Chapter V analyzes the potential use of data prioritization technologies by

network operators. Chapter VI considers the current and future state of competition in

the area of broadband Internet access. Chapter VII explores the application of the

antitrust laws to certain potential conduct and business arrangements involving ISPs and

other network operators. Chapter VIII addresses consumer protection issues relating to

broadband Internet access. Chapter IX identifies regulatory, legislative, and other

proposals for broadband Internet access that have been put forth to date. Finally, Chapter

X identifies guiding principles for policy makers to consider prior to enacting any new

laws or regulations in this area.

The Contours of the Debate

Proponents of network neutrality regulation include, among others, some content

and applications providers, non-facilities-based ISPs, and various commentators. They

generally argue that “non-neutral” practices will cause significant and wide-ranging

harms and that the existing jurisdiction of the FCC, FTC, and DOJ, coupled with

Congressional oversight, are insufficient to prevent or remedy those harms. Proponents

suggest that, with deregulation of broadband services, providers of certain broadband

Internet services have the legal ability, as well as economic incentives, to act as

gatekeepers of content and applications on their networks.

Principally, these advocates express concern about the following issues: (1)

blockage, degradation, and prioritization of content and applications; (2) vertical

integration by ISPs and other network operators into content and applications; (3) effects

on innovation at the “edges” of the network (that is, by content and applications

providers); (4) lack of competition in “last-mile” broadband Internet access markets; (5)

remaining legal and regulatory uncertainty in the area of Internet access; and (6) the

diminution of political and other expression on the Internet. Not all proponents of net

neutrality regulation oppose all forms of prioritization, however. For example, some

believe that prioritization should be permitted if access to the priority service is open to

all content and applications providers on equal terms; that is, without regard to the

identity of the content or application provider.

Opponents of network neutrality regulation include, among others, some

facilities-based wireline and wireless network operators and other commentators. They

maintain that net neutrality regulation will impede investment in the facilities necessary

to upgrade Internet access and may hamper technical innovation. They also argue that

the sorts of blocking conduct described by net neutrality proponents are mainly

hypothetical thus far and are unlikely to be widespread and thus are insufficient to justify

a new, ex ante regulatory regime.

Principally, opponents of net neutrality regulation argue that: (1) neutrality

regulations would set in stone the status quo, precluding further technical and business-

model innovation; (2) effective network management practices require some data

prioritization and may require certain content, applications, or attached devices to be

blocked altogether; (3) new content and applications are likely to require prioritization

and other forms of network intelligence; (4) allowing network operators to innovate

freely and differentiate their networks permits competition that is likely to promote

enhanced service offerings; (5) prohibiting price differentiation would reduce incentives

for network investment generally and may prevent pricing and service models more

advantageous to marginal consumers; (6) vertical integration by network operators into

content and applications and certain bundling practices may benefit consumers; and (7)

there is insufficient evidence of either the likelihood or severity of potential harms to

justify an entirely new regulatory regime, especially given that competition is robust and

intensifying and the market generally is characterized by rapid technological change.

Competing Concerns about Integration and Differentiation

Proponents of net neutrality regulation have raised various concerns about the

effects of data or price differentiation in broadband markets.

Certain of these concerns

are tied to vertical integration (broadly construed), as broadband Internet access providers

have begun to offer online content and applications in addition to their primary access

services. Other concerns are independent of such integration.

In particular, proponents are concerned that vertical integration by Internet access

providers into content and applications markets could prompt them to block, degrade, or

charge higher prices to competing content or applications. New information

technologies, such as deep packet inspection, may allow network operators to identify the

source and content of much of the data traffic they handle. Hence, a broadband provider

with significant market power in a given access market, which has an interest in content

or applications generally, could have an incentive to block or degrade competing content

or applications.

Independent of market power considerations, some net neutrality proponents have

raised concerns about the so-called “terminating access monopoly problem,” which could

result from broadband Internet access providers charging content or applications

providers terminating fees for delivery to end users over the last mile. Some proponents

also have expressed concern that if broadband providers are allowed to sign exclusive

deals with content or applications providers, end users may be unable to access much of

the content they desire, thus “balkanizing” the Internet.

On the other hand, because vertical integration may offer efficiencies that are

procompetitive and pro-consumer, not all vertical integration is problematic. More

particularly, opponents of net neutrality regulation maintain that some degree of vertical

See infra Chapters IV and V for more detailed discussion of data differentiation and price differentiation,

respectively.

integration by Internet access providers into content and applications may facilitate

investment in infrastructure, investment in content or applications, optimization of fit

between content and delivery systems, and pricing benefits for consumers. They assert

that such vertical integration also may facilitate entry and thereby increase competition in

broadband Internet access markets. Further, the incentives of broadband providers may

cut both ways: for example, despite potentially having an incentive to favor affiliated

content and applications, access providers have argued that they have an interest in

providing access to a wide range of content and applications, which are essential

complements to the services they sell.

As is the case with data discrimination, it is impossible to determine in the

abstract whether allowing content and applications providers (or even end users) to pay

broadband providers for prioritized data transmission will be beneficial or harmful to

consumer welfare.

Such prioritization may provide benefits, such as increased

investment and innovation in networks and improved quality of certain content and

applications that require higher-quality data transmission, as net neutrality opponents

claim. Network neutrality proponents have raised concerns, however, regarding potential

adverse effects of data prioritization, including, among others: (1) a diminution in

innovation by content and applications providers – particularly those unable to pay for

prioritization; (2) the intentional or passive degradation of non-prioritized data delivery;

and (3) increased transaction costs resulting from negotiations between broadband

providers and content and applications providers over prioritization.

The balance between competing incentives on the part of broadband providers to

engage in, and the potential benefits and harms from, discrimination and differentiation in

the broadband area raise complex empirical questions and may call for substantial

additional study of the market generally, of local markets, or of particular transactions.

Again, further evidence of particular conduct would be useful for assessing both the

likelihood and severity of any potential harm from such conduct.

Present and Future Broadband Competition

Proponents and opponents of net neutrality regulation have fundamentally

different views on the present (and likely future) state of competition in the broadband

industry. Proponents argue either that a national market for broadband Internet access is,

in effect, a cable-telephone duopoly or that there are significant failures of competition in

many local markets. Opponents characterize the market as highly competitive.

Broadband Internet access generally is a relatively new industry characterized by high

levels of demand growth from consumers, high market shares held by incumbent cable

and telephone providers, and many new entrants trying to capture some share of the

market.

See infra Chapter V.

Broadband competition issues are discussed throughout this Report, particularly in Chapters VI and VII.

FTC staff did not conduct independent empirical research regarding competition

in local broadband Internet access markets for the purposes of this Report. We note that

opponents of net neutrality regulation have pointed to evidence on a national scale that

(1) access speeds are increasing, (2) prices (particularly speed-adjusted or quality-

adjusted prices) are falling, and (3) new entrants, including wireless and other

competitors, are poised to challenge the incumbent cable and telephone companies. We

note, too, that statistical research conducted by the FCC has tended to confirm these

general trends.

For example, broadband deployment and penetration have increased

dramatically since 2000. The FCC estimated that by 2006, broadband DSL service was

available to 79 percent of the households that were served by a telephone company, and

cable modem service was available to 93 percent of the households to which cable

companies could provide cable television service.

Jurisdiction and the Application of Antitrust Law

The competitive issues raised in the debate over network neutrality regulation are

not new to antitrust law, which is well-equipped to analyze potential conduct and

business arrangements involving broadband Internet access. The antitrust laws are

grounded in the principle that competition serves to protect consumer welfare. In

conducting an antitrust analysis, then, the ultimate issue would be whether broadband

providers engage in unilateral or joint conduct that is likely to harm competition and

consumers in a relevant market.

Many proponents of net neutrality regulation are concerned that broadband

Internet access suppliers have market power in the last-mile access market and that they

will leverage that power into adjacent content and applications markets in a way that will

harm competition in those markets and, ultimately, consumers. Such leveraging may

take the form of exclusive dealing arrangements, refusals to deal, vertical integration, or

certain unilateral conduct. All of these types of conduct can be anticompetitive and

harmful to consumers under certain conditions. They also, however, can be

procompetitive, capable of improving efficiency and consumer welfare, which involves,

among other things, the prices that consumers pay, the quality of goods and services

offered, and the choices that are available in the marketplace. Accordingly, such conduct

would be analyzed under the antitrust laws to determine the net effect of such conduct on

consumer welfare.

See, e.g., FCC, HIGH-SPEED SERVICES FOR INTERNET ACCESS: STATUS AS OF JUNE 30, 2006 (2007)

[hereinafter FCC,

HIGH-SPEED SERVICES], available at

http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-270128A1.doc

. Although some have questioned

whether the methodology used in compiling this data allows the FCC to provide a reliable analysis of

competition in particular markets, the FCC data does provide an overall picture of the significant growth in

broadband penetration over the past few years.

See, e.g., id. at 2-4, 5 tbl.1, 6 tbl.2, 7 tbl.3, 19 tbl.14.

There nonetheless remains significant disagreement with respect to the adequacy

of existing agency oversight. Some proponents of net neutrality regulation have argued

that existing laws, regulations, and agency oversight are inadequate to safeguard

competition in broadband Internet access markets. Those opposed to net neutrality

regulation, however, have argued that current competition law is adequate, that careful

rule-of-reason application of the law is critical to the preservation of competition, and

that additional regulations likely would be over-intrusive and, on balance, a burden to

vibrant competition in broadband markets.

Consumer Protection Issues

Effective consumer protection in the broadband marketplace is essential to robust

competition in that market – regardless of the outcome of the current broadband

connectivity debate. The FTC has been active in enforcing relevant consumer protection

law, bringing a variety of cases against ISPs that have engaged in allegedly deceptive

marketing and billing practices. The Workshop highlighted various consumer protection

concerns. Several Workshop participants argued that such concerns were best addressed

under FTC jurisdiction, given the FTC’s statutory mandate, its interest and experience in

consumer protection issues generally, and its interest and experience in consumer

protection aspects of various Internet services in particular.

Internet access implicates two broad areas of consumer protection: (1) clear and

conspicuous disclosure of material terms of Internet access services; and (2) security and

privacy issues created by broadband Internet access services. Current federal consumer

protection law can address both sets of concerns, although consumer protection issues in

the broadband marketplace may present unique technical and jurisdictional challenges,

both to consumers and law enforcement agencies. Commentators within and without the

Workshop have suggested that federal law enforcement fruitfully could be augmented by

industry self-regulation and expanded federal guidance on pertinent issues.

Suggested Guiding Principles

The FTC’s Internet Access Task Force has conducted a broad examination of the

technical, legal, and economic issues underpinning the debate surrounding broadband

connectivity competition policy. Based on this examination, as well as our experience

with the operation of myriad markets throughout the economy, we identify guiding

principles that policy makers should consider in evaluating options in the area of

broadband Internet access.

We have provided an explanation of the conduct that the

antitrust and consumer protection laws already proscribe and a framework for analyzing

which conduct may foster or impede competition in particular circumstances. In

evaluating whether new proscriptions are necessary, we advise proceeding with caution

before enacting broad, ex ante restrictions in an unsettled, dynamic environment.

See infra Chapter X.

There is evidence that the broadband Internet access industry is moving in the

direction of more, not less, competition, including fast growth, declining prices for

higher-quality service, and the current market-leading technology (i.e., cable modem)

losing share to the more recently deregulated major alternative (i.e., DSL). We

nonetheless recognize that not every local broadband market in the United States may

enjoy vigorous competition.

This Report does not reflect a case-by-case analysis of the

state of competition in each of the localities that may represent relevant antitrust markets.

There also appears to be substantial agreement on the part of both proponents and

opponents of net neutrality regulation that greater competition in the area of broadband

Internet access would benefit consumers. Thus, to the extent that policy makers are not

content to wait for the market to increase competition, they should consider pursuing

various ways of increasing competition in the provision of broadband Internet access.

Based on what we have learned through our examination of broadband

connectivity issues and our experience with antitrust and consumer protection issues

more generally, we recommend that policy makers proceed with caution in evaluating

proposals to enact regulation in the area of broadband Internet access. The primary

reason for caution is simply that we do not know what the net effects of potential conduct

by broadband providers will be on all consumers, including, among other things, the

prices that consumers may pay for Internet access, the quality of Internet access and other

services that will be offered, and the choices of content and applications that may be

available to consumers in the marketplace.

With respect to data discrimination, broadband providers have conflicting

incentives relating to blockage of and discrimination against data from non-affiliated

providers of content and applications.

In the abstract, it is impossible to know which of

these incentives would prove stronger for each broadband provider. Further, even

assuming such discrimination were to take place, it is unknown whether the net effect on

consumer welfare would be adverse. Likewise, it is not possible to know in the abstract

whether allowing content and applications providers to pay broadband providers for

prioritized data transmission will be beneficial or harmful to consumers.

Several open questions that likely will be answered by either the operation of the

current marketplace or technological developments provide additional reasons for

caution. These questions include, among others: (1) How much demand will there be

from content and applications providers for data prioritization?; (2) Will effective data

prioritization, throughout the many networks comprising the Internet, be feasible?; (3)

Would allowing broadband providers to practice data prioritization necessarily result in

the degradation of non-prioritized data delivery?; (4) When will the capacity limitations

of the networks comprising the Internet result in unmanageable or unacceptable levels of

See infra Chapter VI.B.

See infra Chapter IV.

See infra Chapter V.

congestion?; and (5) If that point is reached, what will be the most efficient response

thereto: data prioritization, capacity increases, a combination of these, or some as yet

unknown technological innovation? The eventual answers to these questions may give

policy makers key information about the net effects on consumer welfare arising from the

conduct and business arrangements that network neutrality regulation would prohibit or

limit.

Policy makers also should carefully consider the potentially adverse and

unintended effects of regulation in the area of broadband Internet access before enacting

any such regulation. Industry-wide regulatory schemes – particularly those imposing

general, one-size-fits-all restraints on business conduct – may well have adverse effects

on consumer welfare, despite the good intentions of their proponents. Even if regulation

does not have adverse effects on consumer welfare in the short term, it may nonetheless

be welfare-reducing in the long term, particularly in terms of product and service

innovation. Further, such regulatory schemes inevitably will have unintended

consequences, some of which may not be known until far into the future. Once a

regulatory regime is in place, moreover, it may be difficult or impossible to undo its

effects.

Two aspects of the broadband Internet access industry heighten the concerns

raised by regulation generally. First, the broadband industry is relatively young and

dynamic, and, as noted above, there are indications that it is moving in the direction of

more competition. Second, to date we are unaware of any significant market failure or

demonstrated consumer harm from conduct by broadband providers. Policy makers

should be wary of enacting regulation solely to prevent prospective harm to consumer

welfare, particularly given the indeterminate effects that potential conduct by broadband

providers may have on such welfare.

The federal antitrust agencies, the FTC and the DOJ, and the FCC share

jurisdiction over broadband Internet access, with each playing an important role in

protecting competition and consumers in this area. Further, as a byproduct of the

ongoing debate over network neutrality regulation, the agencies have a heightened

awareness of the potential consumer harms from certain conduct by, and business

arrangements involving, broadband providers. Perhaps equally important, many

consumers are now aware of such issues. Consumers – particularly online consumers –

have a powerful collective voice. In the area of broadband Internet access, they have

revealed a strong preference for the current open access to Internet content and

applications.

The FTC has been involved in the Internet access area for over a decade and will

continue to be involved in the evolving area of broadband access. The FTC Act is

sufficiently flexible to allow the FTC to enforce the antitrust and consumer protection

laws in most industries, including those involving new and ever-changing technologies.

The fundamental principles of antitrust and consumer protection law and economics that

we have applied for years are as relevant to the broadband industry as they are to other

industries in our economy.

The FTC will continue to devote substantial resources to maintaining competition

and protecting consumers in the area of broadband Internet access, using a variety of

tools. The FTC will continue to enforce the antitrust and consumer protection laws in

evaluating conduct and business arrangements involving broadband access. Further, the

FTC’s Broadband Connectivity Competition Policy Workshop and this Report exemplify

some of the diverse resources the agency may bring to bear on Internet access issues, in

addition to specific law enforcement actions. The Workshop and Report reflect the

agency’s interest in and commitment to developing competition and consumer protection

policy. Finally, the agency will continue to expend considerable efforts at consumer

education, industry guidance, and competition advocacy in the important area of Internet

access.

I. THE INTERNET: HISTORICAL AND TECHNICAL BACKGROUND

The Internet is a decentralized network of computer networks that enables

millions of private and public computers around the world to communicate with each

other. This interconnection of multiple computer networks, which otherwise would

function only as a series of independent and isolated islands, gives rise to the term

“Internet” as we know it today.

This Chapter is organized as follows. Section A

summarizes the historical development of the Internet and describes how data is routed

over it; Section B discusses the relationship between “last-mile” Internet service

providers, Internet “backbone” networks, and content and applications providers; and

Section C explores the technical aspects of network management, data prioritization, and

other forms of data “discrimination.”

A. Historical Development

The Internet developed out of research efforts funded by the U.S. Department of

Defense Advanced Research Projects Agency in the 1960s and 1970s to create and test

interconnected computer networks.

The fundamental aim of computer scientists

working on this “ARPANET” was to develop an overall Internet architecture that could

connect and make use of existing computer networks that might, themselves, be different

The Federal Networking Council, a group of U.S. federal agency representatives involved in the early

development of federal networking, for example, adopted this definition of the term “Internet” in 1995:

“Internet” refers to the global information system that–

(i) is logically linked together by a globally unique address space based on the Internet

Protocol (IP) or its subsequent extensions/follow-ons;

(ii) is able to support communications using the Transmission Control Protocol/Internet

Protocol (TCP/IP) suite or its subsequent extensions/follow-ons, and/or other IP-

compatible protocols; and

(iii) provides, uses or makes accessible, either publicly or privately, high level services

layered on the communications and related infrastructure described herein.

U.S. Federal Networking Council, Resolution dated October 24, 1995, in Robert E. Kahn & Vinton G.

Cerf, What Is the Internet (and What Makes It Work) n.xv (1999), available at

http://www.cnri.reston.va.us/what_is_internet.html

The convention of writing “internet” in lower case letters typically refers to interconnected

networks generally, while writing “Internet” with an uppercase “I” is generally used to refer to the original

or current version of the Internet. D

OUGLAS E. COMER, THE INTERNET BOOK 60 (4th ed. 2007).

Sometimes, though, individual networks are also referred to as being alternative “Internets.” E.g.,

NTERNET2, ABOUT US (2007), available at http://www.internet2.edu/about.

See generally David D. Clark, The Design Philosophy of the DARPA Internet Protocols, COMPUTER

COMM. REV., Aug. 1988, at 106, available at http://nms.csail.mit.edu/6829-papers/darpa-internet.pdf;

BARRY M. LEINER ET AL., A BRIEF HISTORY OF THE INTERNET,

http://www.isoc.org/internet/history/brief.shtml

(last visited June 18, 2007); COMER, supra note 24, at 62.

both architecturally and technologically.

The secondary aims of the ARPANET project

were, in order of priority: (1) Internet communication must continue despite the loss of

networks or gateways between them; (2) the Internet architecture must support multiple

types of communications services; (3) the architecture must accommodate a variety of

networks; (4) it must permit distributed, decentralized management of its resources; (5)

the architecture must be cost-effective; (6) the architecture must permit attachment by

computer devices with a low level of effort; and (7) the resources used in the Internet

architecture must be accountable.

That is to say, ARPANET’s first priority was

network survivability in a potentially hostile environment, and its last priority was

providing a system for allocating charges for passing data packets from network to

network.

By the late 1960s, computer scientists were experimenting with non-linear

“packet-switched” techniques to enable computers to communicate with each other.

Using this method, computers disassemble information into variable-size pieces of data

called “packets” and forward them through a connecting medium to a recipient computer

that then reassembles them into their original form. Each packet is a stand-alone entity,

like an individual piece of postal mail, and contains source, destination, and reassembly

information. Unlike traditional circuit-switched telephone networks, packet-switched

networks do not require a dedicated line of communication to be allocated exclusively for

the duration of each communication. Instead, individual data packets comprising a larger

piece of information, such as an e-mail message, may be dispersed and sent across

Clark, supra note 25, at 106 (“The top level goal for the DARPA Internet Architecture was to develop an

effective technique for multiplexed utilization of existing interconnected networks.”).

Id. at 107.

Id. Besides survivability, “[t]here were also other concerns, such as implementation efficiency,

internetwork performance, but these were secondary considerations at first.” L

EINER ET AL., supra note 25.

David D. Clark, who served as chief Protocol Architect for TCP/IP from 1981-89, has noted that the

ARPANET’s original goals differ from what an architecture designed for commercial purposes might have

looked like:

This set of goals might seem to be nothing more than a checklist of all the

desirable network features. It is important to understand that these goals are in order of

importance, and an entirely different network architecture would result if the order were

changed. For example, since this network was designed to operate in a military context,

which implied the possibility of a hostile environment, survivability was put as a first

goal, and accountability as a last goal. During wartime, one is less concerned with

detailed accounting of resources used than with mustering whatever resources are

available and rapidly deploying them in an operational manner. While the architects of

the Internet were mindful of [resource] accountability, the problem received very little

attention during the early stages of the design, and is only now being considered. An

architecture primarily for commercial deployment would clearly place these goals at the

opposite end of the list.

Clark, supra note 25, at 107.

See generally LEINER ET AL., supra note 25.

multiple paths before reaching their destination and then being reassembled.

This

process is analogous to the way that the individual, numbered pages of a book might be

separated from each other, addressed to the same location, forwarded through different

post offices, and yet all still reach the same specified destination, where they could be

reassembled into their original form.

By the mid-1970s, computer scientists had developed several software

communications standards, or protocols, for connecting computers within the same

network. At about the same time, ARPANET scientists developed a protocol for

connecting different networks to each other, called the Transmission Control

Protocol/Internet Protocol (“TCP/IP”) software suite.

The TCP component of the suite

controls the disassembly and reassembly of data packets sent from a computer server,

where the data resides.

The IP component specifies the formatting and addressing

scheme for transmitting data between sender and recipient computers.

This approach requires that individual networks be connected together by gateway

interface devices, called switches or routers.

Thus, interconnected networks are, in

See generally JONATHAN E. NUECHTERLEIN & PHILIP J. WEISER, DIGITAL CROSSROADS: AMERICAN

TELECOMMUNICATIONS POLICY IN THE INTERNET AGE 39-45 (paperback ed., 2007) (comparing circuit-

switched and packet-switched networks).

See id. at 42.

Vinton G. Cerf & Robert E. Kahn, A Protocol for Packet Network Intercommunication, 22 IEEE

TRANSACTIONS ON COMM. 637 (1974), available at

http://www.cs.princeton.edu/courses/archive/fall06/cos561/papers/cerf74.pdf.

In the original paper describing the TCP/IP protocol, Cerf and Kahn explain:

Processes that want to communicate present messages to the TCP for transmission, and

TCP’s deliver incoming messages to the appropriate destination processes. We allow the

TCP to break up messages into segments because the destination may restrict the amount

of data that may arrive, because the local network may limit the maximum transmission

size, or because the TCP may need to share its resources among many processes

concurrently. . . .

From this sequence of arriving packets (generally from different HOSTS

[computers]), the TCP must be able to reconstruct and deliver messages to the proper

destination processes.

Id. at 640.

“Since the GATEWAY [(router)] must understand the address of the source and destination HOSTS, this

information must be available in a standard format in every packet which arrives at the GATEWAY. This

information is contained in an internetwork header prefixed to the packet by the source HOST.” Id. at 638.

“If the TCP is to determine for which process an arriving packet is intended, every packet must contain a

process header (distinct from the internetwork header) that completely identifies the destination process.”

Id. at 640.

See id. at 638.

effect, a series of routers connected by transmission links. Packets of data are passed

from one router to another, via the transmission links. Typically, each router has several

incoming transmission links through which packets arrive and several outgoing links

through which the router can send packets. When a packet arrives at an incoming link,

the router will use a software algorithm to determine the outgoing link through which the

packet should be routed. If that outgoing link is free, the packet is sent out immediately.

If the relevant outgoing link is busy transmitting other packets, however, the newly

arrived packet must wait. Usually, the packet will be temporarily held, or “buffered,” in

the router’s memory, waiting its turn until the relevant outgoing link is free. Thus,

buffering is a method of dealing with temporary surges in Internet traffic, which can be

variable or “bursty.” If too many packets are buffered during a period of congestion,

however, the router may have no choice but to reroute or drop altogether some of those

packets.

Because no transmission mechanism can be completely reliable, computer

scientists also developed methods of retransmitting data to deal with dropped or

otherwise incorrectly transmitted packets.

Two of the resulting features of this TCP/IP protocol are that it transmits data

between networks on a “first-in-first-out” and “best-efforts” basis.

Therefore, although

the resulting interconnected networks are generally able to transmit data successfully

See generally Edward W. Felten, Nuts and Bolts of Network Neutrality 1-2 (AEI-Brookings Joint Center,

Working Paper No. RP-06-23, 2006), available at http://www.aei-

brookings.org/publications/abstract.php?pid=1106. See also Jon M. Peha, The Benefits and Risks of

Mandating Network Neutrality and the Quest for a Balanced Policy, 34th Research Conference on

Communication, Information, & Internet Policy 5-6 (2006), available at

http://web.si.umich.edu/tprc/papers/2006/574/Peha_balanced_net_neutrality_policy.pdf

(describing the use

of algorithms to manage traffic flows across a network).

As Cerf and Kahn explained:

No transmission can be 100 percent reliable. We propose a timeout and positive

acknowledgement mechanism which will allow TCP’s to recover from packet losses

from one HOST to another. . . . [T]he inclusion of a HOST retransmission capability

makes it possible to recover from occasional network problems and allows a wide range

of HOST protocol strategies to be incorporated. We envision it will occasionally be

invoked to allow HOST accommodation to infrequent overdemands for limited buffer

resources, and otherwise not used much.

Cerf & Kahn, supra note 32, at 643.

See generally DAVID CLARK ET AL., NEW ARCH: FUTURE GENERATION INTERNET ARCHITECTURE: FINAL

TECHNICAL REPORT (2003), available at http://www.isi.edu/newarch/iDOCS/final.finalreport.pdf

(sponsored by DARPA Information Technology Office). “The original Internet provided a very simple and

minimally specified packet transfer service, sometimes called ‘best effort’. Crudely, what ‘best effort’

means is that the network makes no specific commitments about transfer characteristics, such as speed,

delays, jitter, or loss.” Id. at 7.

between senders and receivers using TCP/IP, congestion or other technical issues can

affect transmission and, as a result, no particular quality-of-service level is guaranteed.

Also, during the Internet’s early years, network architectures generally were based

on what has been called the “end-to-end argument.”

This argument states that computer

application functions typically cannot, and should not, be built into the routers and links

that make up a network’s middle or “core.” Instead, according to this argument, these

functions generally should be placed at the “edges” of the network at a sending or

receiving computer.

This argument also recognizes, however, that there might be

certain functions that can be placed only in the core of a network. Sometimes, this

argument is described as placing “intelligence” at or near the edges of the network, while

leaving the core’s routers and links mainly “dumb” to minimize the potential for

transmission and interoperability problems that might arise from placing additional

complexity into the middle of the network.

Throughout the 1970s and 1980s, the interconnection of computer networks using

TCP/IP continued to grow, spurred by uses such as e-mail.

In the mid-1980s, the

National Science Foundation (“NSF”) recognized that computer networks were having an

important impact on scientific research by facilitating communications between

researchers working in different locations. NSF and DARPA had been jointly funding a

network to connect computer science researchers (“CSNET”) since the late 1970s. In

1985, NSF announced a plan to connect one hundred universities to the Internet, in

addition to five already-existing supercomputer centers located around the country.

In the original paper describing the TCP/IP protocol, Cerf and Kahn recognized that because individual

networks have differing characteristics, “[t]he transmit time for this data is usually dependent upon internal

network parameters such as communications media data rates, buffering and signaling strategies, routing,

propagation delays, etc.” Cerf & Kahn, supra note 32, at 637. “The success or failure of a transmission

and its performance in each network is governed by different time delays in accepting, delivering, and

transporting the data.” Id. “TCP may need to share its resources among many processes concurrently.” Id.

at 640. Likewise, resources needed to buffer high volumes of incoming packets may also be “limited.” Id.

at 643. Thus, “[c]ongestion at the TCP level is flexibly handled owing to the robust retransmission and

duplicate detection strategy.” Id. at 645.

See, e.g., J.H. Saltzer et al., End-to-End Arguments in System Design, 2 ACM TRANSACTIONS ON

COMPUTER SYS. 277 (1984).

Id. at 277 (“The argument appeals to application requirements, and provides a rationale for moving

function upward in a layered system, closer to the application that uses that function.”).

See, e.g., Adam Thierer, Are “Dumb Pipe” Mandates Smart Public Policy? Vertical Integration, Net

Neutrality, and the Network Layers Model, in N

ET NEUTRALITY OR NET NEUTERING: SHOULD BROADBAND

INTERNET SERVICES BE REGULATED? 73, 79 (Thomas M. Lenard & Randolph J. May, eds., 2006).

LEINER ET AL., supra note 25 (“Thus, by 1985, Internet was already well established as a technology

supporting a broad community of researchers and developers, and was beginning to be used by other

communities for daily computer communications. Electronic mail was being used broadly across several

communities . . . .”).

COMER, supra note 24, at 72-76.

Recognizing the increasing importance of this interconnected network to U.S.

competitiveness in the sciences, however, NSF embarked on a new program with the goal

of extending Internet access to every science and engineering researcher in the country.

In 1988, NSF, in conjunction with a consortium of private-sector organizations,

completed a new long-distance, wide-area network, dubbed the “NSFNET” backbone.

Although private entities were now involved in extending the Internet, its design

still reflected ARPANET’s original goals. Although the original ARPANET was

decommissioned in 1990, its influence continued because TCP/IP had supplanted or

marginalized most other wide-area computer network protocols in existence at that

time,

and because its design, which provided for generality and flexibility, proved to be

durable in a number of contexts.

At the same time, its successful growth made clear

that these design priorities no longer matched the needs of users in certain situations,

particularly regarding accounting and resource management.

By 1992, the volume of traffic on NSFNET was approaching capacity, and NSF

realized it did not have the resources to keep pace with the increasing usage.

Consequently, the members of the consortium formed a private, non-profit organization

called Advanced Networks and Services

(“ANS”) to build a new backbone with

transmission lines having thirty times more capacity.

For the first time, a private

organization – not the government – principally owned the transmission lines and

computers of a backbone.

At the same time that privately owned networks started appearing, general

commercial activity on the NSFNET was still prohibited by an Acceptable Use Policy.

Thus, the expanding number of privately owned networks were effectively precluded

from exchanging commercial data traffic with each other using the NSFNET backbone.

Several commercial backbone operators circumvented this limitation in 1991, when they

established the Commercial Internet Exchange (“CIX”) to interconnect their own

backbones and exchange traffic directly. Recognizing that the Internet was outpacing its

ability to manage it, NSF decided in 1993 to leave the management of the backbone to

the competing commercial backbone operators. By 1995, this expanding network of

LEINER ET AL., supra note 25.

“In the context of its priorities, the Internet architecture has been very successful. The protocols are

widely used in the commercial and military environment, and have spawned a number of similar

architectures.” Clark, supra note 25, at 113.

Id.

COMER, supra note 24, at 75-76.

“On the NSFNET Backbone – the national-scale segment of the NSFNET – NSF enforced an

‘Acceptable Use Policy’ (AUP) which prohibited Backbone usage for purposes ‘not in support of Research

and Education.’” L

EINER ET AL., supra note 25.

commercial backbones had permanently replaced NSFNET, effectively privatizing the

Internet.

The growth of the Internet has been fueled in large part by the popularity of the

World Wide Web, created in 1989.

The number of Web sites on the Internet has grown

from one in 1989, to 18,000 in 1995, to fifty million in 2004, and to more than one

hundred million in 2006.

This incredible growth has been due to several factors,

including the realization by businesses that they could use the Internet for commercial

purposes, the decreasing cost and increasing power of personal computers, the

diminishing complexity of creating Web sites, and the expanding use of the Web for

personal and social purposes.

From its creation to its early commercialization, most computer users connected

to the Internet using a “narrowband” dial-up telephone connection and a special modem

to transmit data over the telephone system’s traditional copper wirelines, typically at a

rate of up to 56 kilobits per second (“Kbps”).

Much faster “broadband” connections

have subsequently been deployed using a variety of technologies.

These faster

technologies include coaxial cable wirelines, upgraded copper digital subscriber lines,

fiber-optic wirelines, and wireless, satellite, and broadband-over-powerline

technologies.

Michael Kende, The Digital Handshake: Connecting Internet Backbones 5 (FCC Office of Plans and

Policy, Working Paper No. 32, 2000), available at

http://www.fcc.gov/Bureaus/OPP/working_papers/oppwp32.pdf

See generally WORLD WIDE WEB CONSORTIUM, ABOUT THE WORLD WIDE WEB CONSORTIUM (W3C),

http://www.w3.org/Consortium (last visited June 22, 2007). Other popular uses of the Internet include: the

transfer of data files from one computer to another through a File Transfer Protocol (“FTP”); electronic

mail using Simple Mail Transfer Protocol (“SMTP”); and the use of TELetype NETwork (“TELNET”) to

use one computer to access a different computer at another location. See generally N

UECHTERLEIN &

WEISER, supra note 30, at 130. The Internet is often described as being comprised of multiple “layers,”

including: a physical layer consisting of the hardware infrastructure used to link computers to each other; a

logical layer of protocols, such as TCP/IP, that control the routing of data packets; an applications layer

consisting of the various programs and functions run by end users, such as a Web browser that enables

Web-based e-mail; and a content layer, such as a Web page or streaming video transmission. See id. at

118-21.

Marsha Walton, Web Reaches New Milestone: 100 Million Sites, CNN, Nov. 1, 2006,

http://www.cnn.com/2006/TECH/internet/11/01/100millionwebsites/index.html

(last visited June 15,

2007).

See NUECHTERLEIN & WEISER, supra note 30, at 134-35.

See id. at 134-47. Broadband has been defined by the FCC as services that provide transmission speeds

of 200 Kbps or higher in at least one direction. E.g., FCC,

HIGH-SPEED SERVICES, supra note 18, at 5 tbl.1.

Some critics, however, believe this definition is outdated. See, e.g., G. Sohn, Tr. I at 97 (“[I]t defines

broadband at a ridiculously slow speed, 200 kilobits per second.”).

See infra Chapter VI for a discussion of various broadband technologies.

The thousands of individual networks that make up the global Internet are owned

and administered by a variety of organizations, such as private companies, universities,

research labs, government agencies, and municipalities. Data packets may potentially

travel from their originating computer server across dozens of networks and through

dozens of routers before they reach a “last-mile” Internet service provider

and arrive at

a destination computer. This process of disassembly, transmission, and reassembly of

data packets may take as little as a fraction of a second for a simple piece of information

like a text e-mail traveling along a high-speed network, or it may take several hours for a

larger piece of information like a high-resolution video traveling a long distance along a

low-speed network.

This network of networks connects millions of individuals and organizations in a

way that allows almost instantaneous communications using computers, computerized

mobile devices, and other network attachments. End users interact with each other

through an ever-expanding universe of content and applications, such as: e-mail, instant

messaging, chat rooms, commercial Web sites for purchasing goods and services, social

networking sites, Web logs (“blogs”), music and video downloads, political forums,

voice over IP (“VoIP”) telephony services, streaming video applications, and multi-

player network video games. Internet users include individuals of virtually all ages and

walks of life, established businesses, fledgling entrepreneurs, non-profit groups,

academic and government institutions, and political organizations.

The TCP/IP protocol suite has been updated periodically since its introduction.

In recent years, however, some computer experts and other interested parties have

questioned the TCP/IP suite’s thirty-year-old first-in-first-out and best-efforts

characteristics.

Likewise, in light of the increasing deployment of applications that may

See infra Chapter I.B.1 for a discussion of last-mile ISPs.

See, e.g., NUECHTERLEIN & WEISER, supra note 30, at 136.

Kahn & Cerf, supra note 24 (“Refinement and extension of these protocols and many others associated

with them continues to this day by way of the Internet Engineering Task Force.”). See also I

NTERNET

ENGINEERING TASK FORCE, OVERVIEW OF THE IETF, http://www.ietf.org/overview.html (last visited May

16, 2007) (“The Internet Engineering Task Force (IETF) is a large open international community of

network designers, operators, vendors, and researchers concerned with the evolution of the Internet

architecture and the smooth operation of the Internet.”). IETF activities take place under the umbrella of

the Internet Society. See generally I

NTERNET SOCIETY, ABOUT THE INTERNET SOCIETY,

http://www.isoc.org/isoc

(last visited May 16, 2007) (The Internet Society “is the organization home for the

groups responsible for Internet infrastructure standards, including the Internet Engineering Task Force

(IETF) and the Internet Architecture Board (IAB).”).

E.g., David Farber & Michael Katz, Op-Ed., Hold Off On Net Neutrality, WASH. POST, Jan. 19, 2007, at

A19 (“The current Internet supports many popular and valuable services. But experts agree that an updated

Internet could offer a wide range of new and improved services, including better security against viruses,

worms, denial-of-service attacks and zombie computers; services that require high levels of reliability, such

as medical monitoring; and those that cannot tolerate network delays, such as voice and streaming video.

To provide these services, both the architecture of the Internet and the business models through which

services are delivered will probably have to change.”); Christopher S. Yoo, Network Neutrality and the

Economics of Congestion, 94 G

EO. L.J. 1847, 1863 & n.74 (2006) (noting the opinion of computer scientist

David Farber that the current Internet architecture is “getting old”).

operate better in a non-end-to-end environment, some have reexamined the end-to-end

design argument.

Some also have explored what a next generation Internet architecture

might look like, with the goal of managing the emerging tension between the Internet’s

open characteristics and more technologically demanding new applications.

addition, some observers have suggested that the Internet’s continued exponential growth

and the proliferation of resource-intensive content and applications like video file sharing

and the prospect of Internet Protocol television (“IPTV”) may outstrip the Internet’s

current capacity and cause it to become significantly congested or crash altogether.

The problem of network congestion, in particular, was recognized in the original

paper describing the TCP/IP suite and, although it received less attention than

ARPANET’s other original design priorities, computer scientists continued to be mindful

of the issue. Some, therefore, continued to explore different transmission protocols and

the viability of market-based pricing mechanisms through the 1980s and 1990s.

Further, as data-routing technologies have advanced in recent years, some network

operators have begun openly to consider using prioritization and other active

management practices to improve network management and provide certain premium

See, e.g., Marjory S. Blumenthal & David D. Clark, Rethinking the Design of the Internet: The End-to-

End Arguments vs. the Brave New World, 1 ACM TRANSACTIONS INTERNET TECH. 70 (2001) (concluding

that the open, general nature of the Internet historically associated with the end-to-end argument should be

preserved); R

OBERT E. KAHN, CORPORATION FOR NATIONAL RESEARCH INITIATIVES, INTERNET

EVOLUTION, GOVERNANCE AND THE DIGITAL OBJECT ARCHITECTURE: WORKSHOP ON SCORM

SEQUENCING AND NAVIGATION 8 (Feb. 23, 2005), available at

http://www.handle.net/presentations_plugfest9/PlugFest9_Plenary_kahn.ppt

(discussing whether the

Federal Network Council’s 1995 Internet definition, see supra note 24, should be updated to also include

services “integrated with” communications and related infrastructures); Press Release, Stanford Center for

Internet and Society, The Policy Implications of End-to-End (Dec. 1, 2000), available at

http://cyberlaw.stanford.edu/e2e

(workshop chaired by Professor Lawrence Lessig) (“In an increasing

range of contexts . . . e2e [(end-to-end)] is being questioned. Technologies that undermine e2e are

increasingly being deployed; other essential services, such as quality of service, are being developed in

ways that are inconsistent with e2e design.”).

E.g., CLARK ET AL., supra note 38, at 4 (“The goal of this project was to consider the following question:

if we could now design the Internet from scratch, knowing what we know today, how would we make the

basic design decisions?”).

E.g., DELOITTE TOUCHE TOHMATSU, TELECOMMUNICATIONS PREDICTIONS: TMT TRENDS 2007 (2007),

available at

http://www.deloitte.com/dtt/cda/doc/content/us_tmt_%202007_Telecom_Predictions_011606.pdf

According to this report, “[o]ne of the key possibilities for 2007 is that the Internet could be approaching its

capacity. The twin trends causing this are an explosion in demand, largely fueled by the growth in video

traffic and the lack of investment in new, functioning capacity.” Id. at 4.

E.g., Jeffrey K. MacKie-Mason & Hal R. Varian, Pricing the Internet, in PUBLIC ACCESS TO THE

INTERNET 269 (Brian Kahin & James Keller eds., 1995). According to MacKie-Mason and Varian:

“Congestion is likely to be a serious problem in the future Internet, and past proposals to control it are

unsatisfactory. We think an economic approach to allocating scarce Internet resources is warranted.” Id. at

284. “Our objective is not to raise profits above a normal rate of return by pricing backbone usage. Rather,

our goal is to find a pricing mechanism that will lead to the most efficient use of existing resources, and

will guide investment decisions appropriately.” Id.

services for a fee.

As a result, computer scientists, network operators, content and

applications providers, and other interested parties have increasingly debated the

significance of the Internet’s historical and current architecture and its implications for

the Internet’s future development.

See, e.g., At SBC, It’s All About “Scale and Scope,” BUS. WK., Nov. 7, 2005,

http://www.businessweek.com/@@n34h*IUQu7KtOwgA/magazine/content/05_45/b3958092.htm

(interview with SBC Telecommunications’ CEO Edward Whitacre). According to Whitacre:

[T]here's going to have to be some mechanism for these people who use these pipes to

pay for the portion they're using. Why should they be allowed to use my pipes?

The Internet can't be free in that sense, because we and the cable companies

have made an investment and for a Google or Yahoo! or Vonage or anybody to expect to

use these pipes [for] free is nuts!

Id. See also Marguerite Reardon, Qwest CEO Supports Tiered Internet, ZDN

ET NEWS, Mar. 15, 2006,

http://articles.techrepublic.com.com/2100-1035_11-6050109.html. Qwest CEO Richard Notebaert has

stated his company would like to offer prioritized data transmission in the same way that express parcel

service may be purchased from Federal Express or UPS. In his view, “[i]t’s possible that (these companies)

would like to have differentiated service. . . . And if you have enough money, we can make a lot of things

happen.” Id. “Would this give some content providers an advantage over others? . . . Well, yeah. We’re

all trying to provide a bit of differentiation for a competitive edge. That’s what business is about.” Id.

For example, some of the Internet’s early designers have offered the following account:

One should not conclude that the Internet has now finished changing. The

Internet, although a network in name and geography, is a creature of the computer, not

the traditional network of the telephone or television industry. It will, indeed it must,

continue to change and evolve at the speed of the computer industry if it is to remain

relevant. It is now changing to provide such new services as real time transport, in order

to support, for example, audio and video streams. The availability of pervasive

networking (i.e., the Internet) along with powerful affordable computing and

communications in portable form (i.e., laptop computers, two-way pagers, PDAs, cellular

phones), is making possible a new paradigm of nomadic computing and communications.

This evolution will bring us new applications – Internet telephone and, slightly

further out, Internet television. It is evolving to permit more sophisticated forms of

pricing and cost recovery, a perhaps painful requirement in this commercial world. It is

changing to accommodate yet another generation of underlying network technologies

with different characteristics and requirements, from broadband residential access to

satellites. New modes of access and new forms of service will spawn new applications,

which in turn will drive further evolution of the net itself.

The most pressing question for the future of the Internet is not how the

technology will change, but how the process of change and evolution itself will be

managed. As this paper describes, the architecture of the Internet has always been driven

by a core group of designers, but the form of that group has changed as the number of

interested parties has grown. With the success of the Internet has come a proliferation of

stakeholders – stakeholders now with an economic as well as an intellectual investment

in the network. We now see, in the debates over control of the domain name space and

the form of the next generation IP addresses, a struggle to find the next social structure

that will guide the Internet in the future. The form of that structure will be harder to find,

given the large number of concerned stake-holders. At the same time, the industry

B. Major Internet Components

1. “Last-Mile” Internet Service Providers

“Last-mile”

Internet service providers offer the network connections that link

end users to the wider Internet.

By connecting its end-user customers to the many

networks comprising the Internet backbone, an ISP provides its customers access to the

end-user computers of any other ISP in the world connected to that backbone. Computer

users in the United States have had nearly ubiquitous last-mile access to dial-up Internet

connections of 56 to 280 Kbps since the late 1990s through telephone modems.

recent years, faster broadband connections have supplanted dial-up service for a rapidly

growing number of computer users who demand faster access to the increasingly

sophisticated and data-rich content and applications available on the Internet.

Principally, end users receive last-mile broadband Internet service through coaxial cable

wireline or upgraded copper digital subscriber wireline connections; other platforms,

such as fiber-optic wirelines, wireless, satellite, and broadband over powerlines, are also

increasingly available to connect end users to the Internet.

Basic residential service packages are typically available on a flat-rate basis to

home computer users.

ISPs may require that end users with more demanding needs,

like a medium or large business, purchase a business-class or other type of premium

struggles to find the economic rationale for the large investment needed for future

growth, for example to upgrade residential access to more suitable technology. If the

Internet stumbles, it will not be because we lack for technology, visions, or motivation. It

will be because we cannot set a direction and march collectively into the future.

EINER ET AL., supra note 25.

Networks that connect end users to the broader Internet are generally referred to as “last-mile” ISPs.

Networks that transmit data from a content or applications provider’s computer server(s) to the broader

Internet are sometimes referred to as “first-mile” ISPs.

Today, major last-mile wireline broadband ISPs include: AT&T, Comcast, Covad, Cox

Communications, and Verizon. Major wireless broadband ISPs include: AT&T, Sprint Nextel, T-Mobile,

and Verizon Wireless.

See NUECHTERLEIN & WEISER, supra note 30, at 134-35.

See id. at 134-47.

According to the most recent data available from the FCC, most broadband consumers access the Internet

today by cable modem or DSL. Of the 64.6 million high-speed lines in the United States as of June 30,

2006, 44.1% were cable modem, 36.4% DSL or other high-speed telephone line, 17.0% mobile wireless,

1.1% fiber-to-the-premise, 0.8% satellite, 0.5% fixed wireless, and 0.01% broadband over powerlines (and

other lines). FCC,

HIGH-SPEED SERVICES, supra note 18, at 5 tbl.1.

See generally Lehr, Tr. I at 37 (discussing “the market’s current attraction to . . . flat-rate pricing”);

Brenner, Tr. II at 96. See also, e.g., VERIZON, VERIZON HIGH SPEED INTERNET,

http://www22.verizon.com/content/consumerdsl/plans/all+plans/all+plans.htm

(last visited May 17, 2007).

service package.

In addition, end users can purchase for a premium fee access to a

specialized virtual private network (“VPN”) offering a defined quality-of-service level

over a reserved portion of an ISP’s network.

Last-mile broadband wireline architecture can take various forms. A last-mile

ISP can extend a fiber-optic wireline from a backbone connection to either a

neighborhood node, to the curb of a premise, or all the way to the end user’s premise. If

the fiber runs only to the node or curb, the ISP can then use a cable or DSL connection

for the remaining distance to the end user’s premise.

DSL wirelines provide a

dedicated amount of bandwidth to each end user, but can transmit data up to only about

three miles without the use of a repeater. Accordingly, transmission speeds can vary

depending on an end user’s distance from a repeater.

Cable wirelines offer shared

bandwidth among many customers. Thus, the transmission speed for an individual cable

modem customer can vary with the number of customers who are using the network

simultaneously.

Last-mile wireless networks using wireless fidelity (“Wi-Fi”) or worldwide

interoperability for microwave access (“Wi MAX”) technologies can be set up by

deploying multiple antennas on street lights, traffic signals, and buildings, so that

multiple wireless hotspots overlap each other to form a continuous “mesh” network of

wireless signals. An initial connection to a backbone network also must be made in order

to provide access to the wider Internet.

Several major telecommunications companies

also offer mobile wireless Internet services over their wireless phone networks.

Three

satellite providers offer broadband Internet service via satellite.

An end user must have

a computer or other device that is configured for wireless Internet use to access these

E.g., COMCAST, COMCAST WORKPLACE, http://www.comcast.com/wa-business/internet.html (last visited

May 14, 2007). Last-mile access for large enterprise customers, particularly those with multiple locations,

typically involves the use of dedicated, high-capacity facilities often referred to as special access or

dedicated access services. See In re Special Access Rates for Price Cap Local Exch. Carriers, 20 FCC Rcd

1994, 1995-96 (2005) (order and notice of proposed rulemaking) [hereinafter Special Access NPRM].

See, e.g., CHARLES B. GOLDFARB, ACCESS TO BROADBAND NETWORKS: CONGRESSIONAL RESEARCH

SERVICE REPORT TO CONGRESS 10-11 (2006), available at

http://www.ipmall.info/hosted_resources/crs/RL33496_060629.pdf

Id. at 9-11.

See generally FCC, FCC CONSUMER FACTS: BROADBAND ACCESS FOR CONSUMERS,

http://www.fcc.gov/cgb/consumerfacts/dsl2.html

(last visited June 22, 2007).

See generally id.

Wireless broadband providers that do not have their own facilities connecting their transmitters (e.g., cell

towers) to their switches typically purchase special access services from an incumbent local exchange

carrier or other provider of such services. See Special Access NPRM, 20 FCC Rcd at 1995-96.

GOLDFARB, supra note 73, at 10.

Id. at 10-11.

networks. In addition, there are now over forty deployments of broadband–over-

powerline technologies in the U.S., most of which are in trial stages.

Today’s last-mile networks generally are partitioned asymmetrically to provide

more bandwidth for data traveling from an ISP’s facilities to the end user’s computer

(“downstream”) than in the other direction (“upstream”). Typically, this is done because

end users request much more data from other server computers than they, themselves,

send out.

As a result, asymmetric architecture may constrain content and applications

that require the end user simultaneously to send and receive content at the same speeds

and volumes, such as two-way video transmissions.

Also, ISPs have the technical

capability to reserve portions of last-mile bandwidth for specific applications.

2. Internet Backbone Operators

Since 1995, when the expanding number of commercial backbone networks

permanently replaced NSFNET, commercial backbones have generally interconnected

with each other through voluntary, market-negotiated agreements.

To this day, there

are no general, industry-specific regulations that govern backbone interconnection in the

U.S.

Instead, commercial backbone operators independently make decisions about

interconnection by weighing the benefits and costs on a case-by-case basis.

Typically,

Id. at 11-12.

Id. at 4, 9.

Id. at 9.

For example, Verizon reserves one fiber of its downstream fiber-to-the-home service specifically for the

company’s video service, while a separate fiber carries all other incoming traffic. Id. at 10. AT&T

reserves 19 of 25 megabits of downstream end-user bandwidth specifically for the company’s video

service. Id. at 11. AT&T customers can purchase between 1.5 and 6 Mbps of the remaining downstream

bandwidth for Internet access and voice services. Id.

Observers have noted that:

Particularly in the Internet’s early days, many backbone providers exchanged traffic at

government-sponsored Network Access Points (NAPs)–the Internet’s equivalent to

public airports, where the routes of many different carriers converge. (When the

government privatized the Internet, it transferred control of these points to commercial

providers.) Internet backbone providers now increasingly rely on privately arranged

points of interconnection, largely because of congestion at the NAPs.

UECHTERLEIN & WEISER, supra note 30, at 132.

See generally id. at 133 (“These peering and transit agreements are completely unregulated. Neither the

FCC nor any other governmental authority regulates the prices that a larger backbone network may charge

a smaller one for transit services or mandates that backbone providers interconnect at all.”).

As one commentator notes:

Currently, there are no domestic or international industry-specific regulations that govern

how Internet backbone providers interconnect to exchange traffic, unlike other network

backbones connect to each other under one of two types of arrangements. In a “peering”

arrangement, backbones of similar size engage in a barter arrangement in which

backbone A carries traffic for backbone B in exchange for backbone B carrying a similar

amount of traffic for backbone A. In this arrangement, exchanged traffic generally is

destined only for the other backbone’s end users.

In a “transit” arrangement, a smaller

backbone pays a larger backbone to carry its customers’ traffic to all end users on the

Internet.

To date, market forces have encouraged interconnection among backbones

and between backbones and last-mile ISPs.

Today, these backbones make up the core or “middle” of the Internet. Generally,

individual backbone networks are made up of a multiplicity of redundant, high-speed,

high-capacity, long-haul, fiber-optic transmission lines that join at hubs or points of

interconnection across the globe.

Transmission over the backbone is generally reliable

even when one component fails because there are multiple different routes of

transmission from one computer to another.

A backbone’s customers include ISPs

providing last-mile connectivity to end users, providers of content and applications that

wish to connect their computer servers directly to a backbone, and specialized companies

that lease space on shared or dedicated computer servers to smaller content and

applications providers.

3. Providers of Content and Applications

Millions of organizations and individuals connected to the Internet’s edges

provide an ever-expanding universe of content and applications to end users.

Commercial entities and other organizations provide a large portion of such content and

applications, but individuals are increasingly contributing content and applications to the

Internet for personal, social, and creative purposes.

services, such as long distance voice services, for which interconnection is regulated.

Rather, Internet backbone providers adopt and pursue their own interconnection policies,

governed only by ordinary laws of contract and property, overseen by antitrust rules.

Kende, supra note 50, at 2.

See generally NUECHTERLEIN & WEISER, supra note 30, at 132-33.

Cf. Ryan, Tr. I at 237.

NUECHTERLEIN & WEISER, supra note 30, at 131-38. See also Li Yuan & Gregory Zuckerman, Level 3

Regains Luster Amid Web-Video Boom, W

ALL ST. J., Dec. 21, 2006, at C1 (providing a map of Level 3’s

fiber-optic backbone). Today, major U.S. backbone operators include: Verizon, AT&T, Global Crossing,

Level 3, Qwest, SAVVIS, and Sprint-Nextel.

COMER, supra note 24, at 137-42.

Popular examples include: Blogger.com (Web logs); flickr.com (photo sharing); YouTube.com (audio

and video files); and MySpace.com (social networking pages, Web logs, photo sharing, audio and video

files). See also Lev Grossman, Time’s Person of the Year: You, T

IME, Dec. 25, 2006, at 38, available at

http://www.time.com/time/magazine/article/0,9171,1569514,00.html

Content and applications providers use various methods to distribute their

offerings over the Internet. Smaller organizations and individuals typically lease space

on a shared or dedicated computer server from a specialized company that provides a

connection to the wider Internet, typically through a negotiated agreement with a

backbone operator.

Large companies may build their own server farms with direct

access to an Internet backbone.

Some companies also provide Web sites where users

can post self-generated content, such as photos, blogs, social networking pages, and audio

and video files, while the companies themselves manage the site’s underlying technical

aspects.

Increasingly, content and applications providers are also copying their content

and applications to multiple computer servers distributed around the world, a technique

called local caching.

This practice allows data to be transmitted to end users more

quickly, over a shorter physical distance, and using fewer routers. This strategy, in turn,

generally decreases the potential for transmission problems such as the delay or dropping

of data packets.

Today, many applications can be delivered from a provider’s computer server via

the Internet to a customer’s computer and installed automatically. This ability to transmit

applications cheaply and directly to end users allows applications providers to update

their programs frequently and to deliver new versions to customers quickly. Likewise,

the Internet allows content providers to transmit cheaply an expanding array of content,

such as music and video downloads.

Originally, most Web content consisted of static text and graphics files that could

be viewed graphically using a basic Web browser and a narrowband connection. Some

of the newest content and applications, however, are time-sensitive, bandwidth-intensive,

or both. VoIP, for example, is sensitive to both “latency” – the amount of time it takes a

packet of data to travel from source to destination – and “jitter” – on-again, off-again

See, e.g., TheHostingChart, http://www.thehostingchart.com (last visited June 22, 2007).

See, e.g., Pepper, Tr. I at 93. Pepper notes that “a lot of these large providers made enormous

investments in big server farms to bring content closer to consumers with their caching servers. Bringing

content closer to consumers reduces the need to go across multiple hops [between networks].” Id. See also

Yoo, supra note 59, at 1881-83; John Markoff & Saul Hansell, Hiding in Plain Sight, Google Seeks More

Power, N.Y.

TIMES, June 14, 2006, at A1, available at

http://www.nytimes.com/2006/06/14/technology/14search.html?ei=5090&en=d96a72b3c5f91c47&ex=130

7937600.

See supra note 91.

Content and applications providers may construct multiple server farms in various locations. See supra

note 93. Alternatively, they can contract with a third party to manage this function. See, e.g., Misener, Tr.

II at 191 (“Essentially, you have a company that has set up edge serving facilities. That is to say server

farms outside major metropolitan areas.”). See also Yoo, supra note 59, at 1881-83; William C. Symonds,

Traffic Cops of the Net, B

US. WK., Sept. 25, 2006, at 88, available at

http://www.businessweek.com/magazine/content/06_39/b4002094.htm

(profiling third-party content

distribution company Akamai Technologies).

See Pepper, Tr. I at 93; Yoo, supra note 59, at 1882.

delay associated with bursts of data traffic.

High-resolution video files and streaming

video applications are examples of bandwidth-intensive content and applications that

some observers suggest are already challenging the Internet’s capacity.

C. Network Management, Data Prioritization, and Other Forms of Data

“Discrimination”

The differential treatment of certain data packets by network operators, such as

prioritizing some packets over others, is often referred to as data “discrimination.”

This

Section addresses Internet congestion (one of the primary reasons cited for engaging in

such data discrimination), the various types and uses of data discrimination, and the

feasibility of end users detecting and avoiding certain types of data discrimination.

1. Internet Congestion

As explained above, the problem of network congestion has been recognized

since the Internet’s earliest days. Network resources such as computer processing power,

transmission media, and router buffer memory are finite, like other resources.

Congestion, therefore, can occur at any point on the Internet. Of course, end users can

purchase more powerful computers and network operators can expand the capacity of

their networks, but the computers, physical transmission media, and routers that comprise

the Internet can still transport and process only a certain amount of data at any given

time. Although it happens rarely, if too many computers send bursts of packets at the

same time, a network may become temporarily overloaded.

The TCP/IP protocol generally has enabled the Internet to function at a workable

level, even as Internet use has undergone tremendous growth during the last decade.

100

Nonetheless, Internet transmissions are still subject to variable performance and periods

of congestion. Some observers suggest that the use of bandwidth-intensive applications

like certain peer-to-peer file-sharing protocols by even a small minority of users is

already consuming so many network resources as to be worrisome. This situation is of

particular concern to some experts, who believe that the use of such applications by even

a small portion of Internet users may effectively degrade service for the remaining

See, e.g., Blumenthal & Clark, supra note 60, at 72-73; GOLDFARB, supra note 73, at 2-3 & n.4.

See, e.g., GOLDFARB, supra note 73, at 3-4.

“Unfortunately, engineers, economists, and lawyers have different definitions for discrimination.” Peha,

supra note 36, at 3. Some technology experts distinguish between so-called “minimal” or “needs-based”

discrimination, where packets are discarded or otherwise treated differently only when absolutely necessary

(as in the case of congestion), and “non-minimal” or “active” discrimination, where packets are treated

differently for some other, discretionary reason. See, e.g., Felten, supra note 36, at 4. The introduction to

Chapter IV below includes a discussion of how we use the term “discrimination” in analyzing the potential

effects on consumer welfare of various conduct by ISPs and other network operators.

100

COMER, supra note 24, at 165-69.

majority of end users.

101

Some observers suggest that such applications are already

testing the Internet’s existing capacity and may even potentially crash the Internet, or

parts of it.

102

2. Alleviating Internet Congestion

Several techniques have been used to alleviate short-term Internet congestion.

Non-linear packet switching enables data to be dispersed and, in turn, allows networks to

reroute individual data packets around points of congestion and avert delays. The TCP

component of the TCP/IP suite also monitors delays and slows the packet-transmission

rates accordingly.

103

Some applications, however, such as certain peer-to-peer file-

sharing protocols, operate in a different manner. When congestion occurs, these

applications do not slow their rates of data transmission. Rather, they aggressively take

advantage of TCP’s built-in reduction mechanism and, instead, send data as fast as they

can.

104

Therefore, some networks have actively restricted or blocked altogether these

kinds of applications, on the grounds that the networks need to preserve an equitable

level of service for the majority of their end users.

Networks may also use “hot potato” routing policies that hand off to other

networks at the earliest possible point data that is not destined for termination on their

own networks, thus reducing the use of network resources.

105

Local caching of data by

content and applications providers further helps to alleviate congestion by reducing the

101

According to Peha, “[t]raffic from a very small number of users can dominate the network and starve

everybody else out. Peer-to-peer, in particular, is a problem today, and other applications might come

along. ” Peha, Tr. I at 22. See also S

ANDVINE, INC., NETWORK NEUTRALITY: A BROADBAND WILD WEST?

4 (2005), available at http://www.sandvine.com/general/getfile.asp?FILEID=37

(reporting that it is

common for less than 20% of users/applications/content to consume 80% of a network’s resources);

NDREW PARKER, CACHELOGIC, P2P IN 2005 (2005), available at

http://www.cachelogic.com/home/pages/studies/2005_01.php

(reporting that in 2004 peer-to-peer traffic

constituted 60% of overall Internet data traffic and 80% of upstream data traffic); Press Release, Sandvine,

Inc., EDonkey – Still King of P2P in France and Germany (Sept. 13, 2005), available at

http://www.sandvine.com/news/pr_detail.asp?ID=88

(reporting that P2P file-sharing traffic in the UK and

North America represents up to 48% of all downstream bandwidth and 76% of all upstream traffic).

102

See, e.g., Brenner, Tr. II at 99 (recounting that “[w]e all know the famous story of downloading the

Victoria’s Secret streaming video when so much demand was placed on it, nobody could get a download”).

Beyond this oft-cited example, however, staff has not been presented with any specific evidence of an

instance where a significant portion of the Internet has substantially crashed, apart from general examples

of temporary network congestion. See also D

ELOITTE TOUCHE TOHMATSU, supra note 62, at 4.

103

TCP sends and receives acknowledgements each time a packet is sent to and received from a computer.

Also, TCP automatically starts a timer whenever a computer sends a packet. The timed period depends on

the distance to the recipient computer and delays on the Internet. If the timer runs out before the sending

computer receives an acknowledgement, TCP retransmits the packet and lengthens the timed period to

accommodate the network delay, effectively slowing the transmission rate. Once enough computers in the

network slow down, the congestion clears. See C

OMER, supra note 24, at 140-41.

104

Peha, supra note 36, at 7.

105

NUECHTERLEIN & WEISER, supra note 30, at 132.

distance over which data must travel and the number of routers that might potentially

delay or drop packets. In addition, as discussed below, some networks have proposed

prioritizing data and providing other new types of quality-of-service assurances to

alleviate the effects of congestion.

3. Packet-inspection and Flow-control Technologies

To treat some data packets differently than others, as opposed to simply using a

first-in-first-out and best-efforts approach, a network operator must be able to identify

certain relevant characteristics of those packets.

106

One source of identifying information

is the packet’s header, which contains the IP address of its source and destination. The

packet header also contains several types of information that suggest the type of

application required to open the data file, such as the source and destination port

numbers, the transport protocol, the differentiated service code point or traffic class, and

the packet’s length.

107

Additionally, the header contains the Media Access Control

(“MAC”) address of the packet’s source and destination, which provides information

about the manufacturer of the device attached to the network.

108

In recent years, router manufacturers have refined packet-inspection technologies

to provide network operators with a wide range of information about the data traffic on

their networks, including information not provided in packet headers.

109

These

technologies were developed in part to help local area networks direct traffic more

efficiently and to thwart security risks.

110

Deep packet inspection may also be

implemented on the Internet to examine the content of packet streams – even search for

keywords in text – and to take action based on content- or application-specific policies.

111

Such actions could involve tracking, filtering, or blocking certain types of packet streams.

Further, deep packet inspection can map the information it accumulates to databases

containing, for instance, demographic or billing information.

112

106

Peha, supra note 36, at 3 (discussing the criteria that networks can consider when deciding how to

prioritize packets).

107

Id. at 4. Some computer scientists believe that port numbers have become an unreliable tool for

determining a packet’s associated application. According to Peha, “[o]nce upon a time, you could learn

who the application was, through something called a port number, but that hasn't been reliable or

meaningful for a number of years.” Peha, Tr. I at 18.

108

Peha, supra note 36, at 4.

109

See, e.g., Pepper, Tr. I at 83-87.

110

E.g., Tim Greene, The Evolution of Application Layer Firewalls, NETWORK WORLD, Feb. 2, 2004,

available at http://www.networkworld.com/news/2004/0202specialfocus.html

(“Now the latest Internet

defense technology – deep packet inspection firewalls – is being touted as the best line of defense against

worms that can sneak past earlier technology to wreck havoc in corporate networks.”).

111

Peha, supra note 36, at 4-5.

112

Id.

Another relatively new technology that may be implemented to reveal information

about packet streams is flow classification.

This technology monitors the size of packets

in a data stream, the time elapsed between consecutive packets, and the time elapsed

since the stream began, with the goal of making reasonable determinations about the

nature of the packets in the stream. Thus, flow classification may reveal information

about a packet stream even if the individual packets themselves are encrypted against

packet inspection.

113

With the development of these two technologies, it is now cost-

effective for a network operator to gain extensive knowledge about the nature of the data

traveling across its network.

114

4. Data Prioritization and Other Forms of Data Discrimination

Recently, some network operators have suggested that they would like to use

these new technologies to prioritize certain data traffic or to provide other types of

quality-of-service assurances to content and applications providers and/or end users in

exchange for a premium fee.

115

In contrast to the practice of transmitting data on a first-

in-first-out and best-efforts basis, network operators could use a router algorithm to favor

the transmission of certain packets based on characteristics such as their source,

destination, application type, or related network attachment. One or more of these

strategies could be employed to manage network traffic generally. Or, they might be

used by a network operator to actively degrade certain non-favored traffic.

Packets going to or from certain favored addresses could be given priority

transmission. Likewise, network operators could give priority to packets for latency-

sensitive applications such as VoIP or network video games. In the alternative, routers

could be programmed to reroute, delay, or drop certain packets.

116

For example, a

network operator could block packets considered to be a security threat.

117

It could drop

or otherwise delay packets associated with unaffiliated or otherwise disfavored users,

content, or applications.

118

A network could apply such treatment only in certain

113

Id. at 4. For example, if a network operator detects a steady stream of packets flowing at 30 Kbps across

its network for a period of time, it might conclude those packets are part of a VoIP telephony transmission.

Id.

114

Id. at 5.

115

See supra note 64. Quality of service “typically involves the amount of time it takes a packet to traverse

the network, the rate at which packets can be sent, and the fraction of packets lost along the way.” Peha,

supra note 36, at 5.

116

E.g., Peha, supra note 36, at 4-6.

117

E.g., Craig McTaggart, Was the Internet Ever Neutral?, 34th Research Conference on Communication,

Information, & Internet Policy 9 (2006), available at

http://web.si.umich.edu/tprc/papers/2006/593/mctaggart-tprc06rev.pdf

(discussing blocking as a tool to

control network abuse).

118

E.g., Peha, supra note 36, at 12–13 (describing scenarios in which network operators might block rival

services, specific content, or software).

circumstances, such as during periods of congestion, after a quota of packets has been

met, or, until certain usage fees are paid.

119

Some observers, however, question whether

implementing wide-scale prioritization or similar schemes across multiple networks

having differing technical characteristics is, in fact, even technically possible.

120

Network operators also could provide separate physical or logical channels for

different classes of traffic.

121

Another method for favoring certain Internet traffic is to

reserve capacity on last-mile bandwidth for certain packet streams to provide a minimum

level of quality.

122

Similarly, a network operator could limit the amount of bandwidth

available to an end user, thereby degrading or effectively blocking altogether the use of

119

See, e.g., id. at 5-6.

120

See, e.g., Alcatel-Lucent, Public Comment 1. According to Alcatel-Lucent, an opponent of network

neutrality regulation:

[I]ndustry standards would have to be adopted that put in place common policies for the

labeling and prioritization of data packets. . . . The vast majority of Internet traffic must

traverse the networks of numerous broadband service providers. This means that in order

to favor the traffic of Service A over Service B during its entire trip through the Internet,

each service provider and backbone network would have to prioritize and label packets in

exactly the same way – a scenario that does not exist today. The idea that a service

provider could maintain priority routing for its “preferred data packets” between a user in

Washington, DC and Los Angeles, CA is not possible absent a comprehensive agreement

between all network service providers to treat and identify data packets based on a

common standard not currently in existence. Absent such developments, the data would

almost certainly change hands at least once, likely stripping it of any prioritization it

might have enjoyed inside the network of a sole provider.

Id. at 5. Likewise, a representative of Google, a network neutrality proponent, states that:

[L]ast mile providers who want to give some sort of priority service, you know, only

have control over their own network. It’s not obvious to us how you can offer this kind

of end-to-end service. It’s not obvious to us how you identify the traffic in order to

segregate it, that you’re going to give priority to. And how do you do this segregation

without degrading other traffic?

Davidson, Tr. I at 230-31.

121

For example, a network operator could physically send favored data traffic over a lightly used

connection, while sending other data traffic over a more heavily used connection. Or, the network could

use logical separation to send traffic on the same physical connection, but use different service flows, as in

the case of a virtual local network (“VLN”). Peha, supra note 36, at 6.

122

For example, AT&T’s Project Lightspeed and Verizon’s FiOS services reserve portions of last-mile

bandwidth for their proprietary video services. G

OLDFARB, supra note 73, at 10-11, 17-18. These network

operators also could sell reserved capacity to content or applications providers in return for a quality-of-

service guarantee. Verizon, for example, has such plans for its FiOS service. Id. at 10.

bandwidth-intensive content or applications.

123

A network operator also could treat data

packets differently by providing preferential access to services, such as local caching.

124

Data also can be treated differently through the use of pricing structures, such as

service tiers, to provide a certain quality-of-service level in exchange for payment.

125

a fee-for-priority system, content and applications providers and/or end users paying

higher fees would receive quicker, more reliable data transmissions. Sometimes, such an

arrangement is referred to as a “fast lane.” Other data might simply be provided on a

best-efforts basis. Similarly, a network operator might assess fees to end users based on

their behavior patterns, a practice sometimes referred to as “content billing” or “content

charging.”

126

5. Detecting Data Discrimination

127

Although differential data treatment may be easy to detect in some instances, like

outright blocking, in many instances it may be more difficult for an end user to

distinguish between performance problems resulting from deliberate discrimination and

problems resulting from other, more general causes.

128

For example, an end user whose

Internet traffic is treated differently than other traffic might experience poor performance

in one or more aspects, such as delays in transmitting data, delays in using applications,

or sporadic jitter. Such effects, however, can also result from general network

123

See Network Neutrality: Competition, Innovation, and Nondiscriminatory Access: Hearing Before the S.

Comm. on Commerce, Sci., & Transp., 109th Cong. 13 (2006) (testimony of Earl W. Comstock, President

and CEO, COMPTEL), available at http://www.digmedia.org/docs/comstock-020706.pdf

124

Id. at 14.

125

Peha, supra note 36, at 6.

126

Id.

127

The difficulties associated with end-user detection of data discrimination discussed in this Section would

appear to be equally applicable to enforcement of any network neutrality regulation that prohibited data

discrimination by ISPs and other network operators.

128

See, e.g., Pepper, Tr. I at 93. According to Pepper:

[T]here are techniques that consumers actually have readily available to them to test their

own bandwidth and performance latency between . . . the home, or the office, and the

first POP [(point of presence)], right?

And so, those techniques are actually relatively available. The problem is that,

depending on the service you’re trying to download, the application that you’re using, it

may – you may be going through two or three hops [between networks], or as many as a

dozen hops across the Internet. When you go across multiple hops across multiple

networks, it’s more difficult for a consumer to know.

Id. See also Brenner, Tr. II at 98 (“[T]here are many points between the key strokes of the customer and

the download in which the speed can be affected.”).

congestion.

129

Distinguishing the two may be particularly difficult for end users not

possessing a technical background. Researchers, however, are working to develop

diagnostic tools to detect the differential treatment of data.

130

6. Potential End-user Responses to Data Discrimination

a. Bypassing Discriminatory Networks

Some computer experts have suggested that the prospect of networks treating

some data differently than others might give rise to a kind of arms race between network

operators seeking to employ technical measures to manage their networks and end users

seeking to employ countermeasures to avoid them.

131

They suggest, for example, that

end users can bypass networks to a limited degree through cooperative access sharing.

132

On a small scale, a group of neighbors with access to multiple, distinct broadband

Internet service providers might each set up an open-access Wi-Fi router, giving

everyone in the group access to each other’s service provider. If one provider engages in

data discrimination, members of the cooperative could bypass it by accessing the Internet

through another provider in the pool. Such a strategy, however, depends on a last-mile

network operator allowing the use of open-access Wi-Fi access points in the first place.

133

To the extent that last-mile networks allow the resale of their services through open-

access wireless networks, competition from resellers might have a similar effect.

134

Alternatively, a municipality might set up its own wireline or wireless network if its

residents are not satisfied with the service provided by private providers. It is

conceivable, however, that a municipal network could also engage in certain practices

that some of its residents consider to be discriminatory.

135

129

See, e.g., Felten, supra note 36, at 4.

130

Robert McMillan, Black Hat: Researcher Creates Net Neutrality Test, COMPUTERWORLD, Aug. 2, 2006,

available at

http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9002154

131

See generally William H. Lehr et al., Scenarios for the Network Neutrality Arms Race, 34th Research

Conference on Communication, Information, & Internet Policy (2006), available at

http://web.si.umich.edu/tprc/papers/2006/561/TPRC2006_Lehr%20Sirbu%20Peha%20Gillett%20Net%20

Neutrality%20Arms%20Race.pdf. See also Lehr, Tr. I at 52.

132

Lehr et al., supra note 131, at 10-13. See also Lehr, Tr. I at 41-43.

133

Lehr et al., supra note 131, at 10-13.

134

Id. at 13-14 (describing the Wi-Fi resale business model of FON); Lehr, Tr. I at 42-43. See also FON,

What’s FON, http://www.fon.com/en/info/whatsFon (last visited May 14, 2007).

135

Lehr et al., supra note 131, at 15; Lehr, Tr. I at 43.

b. Technical Measures to Counter Data Discrimination

Countering data discrimination, like detecting it in the first place, may be

difficult, especially for end users without technical backgrounds. Several technical

measures to counter data discrimination do exist, however, at least to a limited degree.

Several potential methods for circumventing applications-based degradation or blocking

involve the computer port numbers that typically indicate which software application a

computer should use to open a packet. Computer users and applications developers can

prevent networks from identifying the application associated with a packet by employing

port numbers not commonly associated with a particular application or by assigning and

reassigning port numbers dynamically.

136

Alternatively, applications developers can use

TCP port 80, the number used by most hypertext transfer protocol (“HTTP”) traffic and,

thus, potentially make an application’s traffic indistinguishable from most other Web

browser-based traffic.

137

To evade differential treatment based on a sender or receiver’s IP address, an end

user could access information from the Internet through a proxy that reroutes data

through another server, camouflaging its source and destination.

138

Likewise, packets

might be encrypted so that a network cannot use packet inspection to identify their

contents or related application.

139

Such encrypted packets could also be transmitted

through a VPN to a gateway computer outside the ISP’s network, where the packets

could be decrypted and forwarded to their recipient.

140

In such a scenario, the last-mile

ISP would see only streams of encrypted packets traveling from the end user through the

VPN, thus preventing the ISP from identifying the computers with which the sender is

communicating.

141

Some ISPs have responded to these measures by banning the use of

VPNs and encryption protocols or charging a fee for their use.

142

Alternatively, a

network might simply relegate or drop altogether encrypted packets when it cannot

identify their contents.

An alternate encryption system called “onion routing” conceals packets’ content,

source, and destination without the use of a VPN. A packet is enveloped in several layers

of encryption and then sent through a special network of links and unique routers called

136

Lehr et al., supra note 131, at 19-20. See also Lehr, Tr. I at 45-46.

137

Lehr et al., supra note 131, at 20-21.

138

Id.

139

For example, some P2P software has been rewritten using the Internet IP Security protocol (“IPSec”) to

encrypt everything in the packets except the IP header. Id.

140

Felten, supra note 36, at 8-9.

141

Id.

142

Lehr et al., supra note 131, at 22.

“routing anonymizers” or “onion routers.”

143

A layer of encryption is removed at each

router until the packet is stripped of encryption and delivered to its destination. Onion

routing prevents network operators from knowing who is communicating with whom,

and the content of the communication is encrypted up to the point where the traffic leaves

the onion-routing network.

144

Even with encryption, however, a network might be able to infer the type of

packet through flow classification and continue to target certain packets for

discrimination.

145

An end user might try to evade flow classification by altering the size

and timing of packets, adding blank packets to the flow, or mixing packets from multiple

flows.

146

A network might respond, however, by degrading or blocking all of the user’s

traffic or by manipulating that traffic in a way that affects one type of application much

more than it does other types of traffic.

147

Alternatively, end users might be able to offset the effects of certain kinds of

discrimination to some extent by using buffering techniques to preload data streams into

a computer’s memory and then accessing them after a period of time, thereby alleviating

problems with latency or jitter. Such techniques, however, may not be useful for real-

time applications like VoIP and streaming video.

148

In some circumstances, caching

content closer to end users might also effectively circumvent discriminatory practices that

are implemented further into the core of the Internet.

149

* * *

The text above provides historical and technical background regarding the

Internet to help inform the policy discussion in this Report. In the next Chapter, we

address the jurisdiction of the relevant federal agencies in the area of broadband Internet

access, as well as the legal and regulatory developments that have prompted the current

debate over network neutrality.

143

Id.

144

Id. See also generally U.S. Navy, Onion Routing: Executive Summary, http://www.onion-

router.net/Summary.html (last visited June 15, 2007).

145

Felten, supra note 36, at 8-9; Lehr et al., supra note 131, at 23; see Peha, supra note 36, at 4.

146

Lehr et al., supra note 131, at 23.

147

Felten, supra note 36, at 9.

148

Lehr, Tr. I at 48-49.

149

Id. at 49.

II. LEGAL AND REGULATORY BACKGROUND AND DEVELOPMENTS

If recent years have seen considerable change in the development and deployment

of platforms for broadband Internet access, they also have seen considerable flux in the

field of broadband regulation. A comprehensive review of federal and state law issues

pertinent to the provision of broadband Internet access would go well beyond the scope

of this Report.

150

This Chapter, however, provides a basic legal and regulatory

framework for the policy discussion to follow in the remainder of the Report. To that

end, it sketches the central elements of FTC (in Section A) and FCC (in Section B)

jurisdiction over broadband services, including the statutory bases of that jurisdiction.

This Chapter also reviews (in Section C) certain decisions of the courts and the agencies,

including recent enforcement activity, rulemaking, and policy statements that have served

to clarify both jurisdictional and substantive questions about broadband Internet access.

In brief, federal regulatory jurisdiction over broadband services generally is

subject to the shared jurisdiction of the FCC, the FTC, and the DOJ. FCC jurisdiction

comes chiefly from the Communications Act,

151

which established the FCC and provides

for the regulation of “interstate and foreign commerce in communication by wire and

radio.”

152

FTC jurisdiction over broadband services comes chiefly from its statutory

mandate to prevent “unfair methods of competition” and “unfair or deceptive acts or

practices in or affecting commerce” under the FTC’s enabling legislation, the FTC Act.

153

The FTC’s authority to enforce the federal antitrust laws generally is shared with DOJ’s

Antitrust Division.

154

150

For a more detailed treatment of the pertinent legal background, see, e.g., PETER W. HUBER ET AL.,

FEDERAL TELECOMMUNICATIONS LAW (2d ed. 1999) (especially Chapters 3, 10-12, Supp. (2005), and

Supp. (2006)). See also NUECHTERLEIN & WEISER, supra note 30 (discussing Internet commerce, policy,

and law).

151

47 U.S.C. §§ 151 et seq. Significant amendments to the Communications Act of 1934, 48 Stat. 1064

(1934), were imposed by the Telecommunications Act of 1996. See Pub. L. No. 104-104, 110 Stat. 56

(1996). Although broad in scope, the Telecommunications Act of 1996 did not replace the

Communications Act, but amended it.

152

47 U.S.C. § 151.

153

15 U.S.C. §§ 41 et seq. Although the FTC Act is central to the FTC’s jurisdiction over broadband

Internet access, and competition and consumer protection issues generally, it is not the only statutory basis

of FTC authority pertinent to the larger Internet debate. With regard to competition concerns, the FTC is

also charged under, for example, the Clayton Act (15 U.S.C. §§ 12-27); the Hart-Scott-Rodino Antitrust

Improvements Act of 1976 (15 U.S.C. § 18a) (amending the Clayton Act); and the International Antitrust

Enforcement Assistance Act of 1994 (15 U.S.C. §§ 46, 57b-1, 1311, 1312, 6201, 6201 note, 6202-6212).

154

The FTC and DOJ share antitrust authority with regard to most areas of the economy. The two antitrust

agencies have long-standing arrangements, first established in 1948, that allow them to avoid inconsistent

or duplicative efforts. See infra notes 218-19 for a discussion of various DOJ merger reviews in the area of

Internet broadband access.

A. FTC Jurisdiction under the FTC Act

The FTC Act gives the FTC broad authority with regard to both competition and

consumer protection matters in most sectors of the economy.

155

Under the FTC Act,

“[u]nfair methods of competition in or affecting commerce, and unfair or deceptive acts

or practices in or affecting commerce,” are prohibited,

156

and the FTC has a general

statutory mandate “to prevent persons, partnerships, or corporations,” from engaging in

such prohibited methods, acts, and practices.

157

At the same time, the FTC Act cabins this general grant of statutory authority

with regard to certain activities. In particular, the FTC’s enforcement authority under the

FTC Act does not reach “common carriers subject to the Communications Act of 1934,”

as amended.

158

An entity is a common carrier, however, only with respect to services that

it provides on a common carrier basis.

159

As discussed below in Chapter II.C, because

most broadband Internet access services are not provided on a common carrier basis, they

are part of the larger economy subject to the FTC’s general competition and consumer

protection authority with regard to methods, acts, or practices in or affecting commerce.

Exercising its statutory authority over competition matters, the FTC has, where

appropriate, investigated and brought enforcement actions in matters involving access to

content via broadband and other Internet access services. For example, the FTC

challenged the proposed merger between America Online (“AOL”) and Time Warner, on

the basis that the merger threatened to harm competition and injure consumers in several

markets, including those for broadband Internet access and residential Internet transport

services (i.e., “last mile” access).

160

The consent order resolving the agency challenge

required the merged entity to open its cable system to competitor Internet service

155

The FTC’s authority is defined broadly to deal with “methods . . . acts or practices in or affecting

commerce.” 15 U.S.C. § 45(a)(2). But for certain limited market sectors that are expressly excluded from

the FTC’s enforcement authority, and for the areas in which FTC jurisdiction over various market sectors is

shared, the FTC’s authority ranges broadly over “commerce,” without restriction to particular segments of

the economy. See id. (FTC authority generally; express exclusion for, e.g., common carriers); supra note

154 and accompanying text (shared FTC/DOJ antitrust authority).

156

15 U.S.C. § 45(a)(1). In 1994, Congress defined an “unfair” act or practice over which the FTC has

authority as one that “causes or is likely to cause substantial injury to consumers which is not reasonably

avoidable by consumers themselves and not outweighed by countervailing benefits to consumers or to

competition.” Id. § 45(n).

157

Id. § 45(a)(2).

158

Id.

159

47 U.S.C. § 153(44) (provider of telecommunications services deemed a common carrier under the

Communications Act “only to the extent that it is engaged in providing telecommunications services”).

160

Am. Online, Inc. & Time Warner, Inc., FTC Dkt. No. C-3989 (Dec. 17, 2000) (complaint), available at

http://www.ftc.gov/os/2000/12/aolcomplaint.pdf

providers on a non-discriminatory basis, for all content.

161

The order also prevented the

company from interfering with the content of non-affiliated ISPs or with the ability of

non-affiliated providers of interactive TV services to access the AOL/Time Warner

system.

162

Moreover, the order required the company, in areas where it provided cable

broadband service, to offer AOL’s DSL service in the same manner and at the same retail

pricing as in areas where it did not provide cable broadband service.

163

The FTC has addressed Internet access and related issues in a number of other

merger investigations as well.

164

For example, the FTC investigated the acquisition by

Comcast and Time Warner of the cable assets of Adelphia Communications and, in a

related matter, the exchange of various cable systems between Comcast and Time

Warner. In the course of that investigation, the FTC examined, among other things, the

likely effects of the transactions on access to and pricing of content. The investigation

eventually was closed because a majority of the Commission concluded that the

acquisitions were unlikely to foreclose competition or result in increased prices.

165

In addition to such competition issues are various consumer protection issues that

have been raised in the larger Internet access context. Over the past decade, the FTC has

brought a variety of cases against Internet service providers that have engaged in

allegedly deceptive marketing and billing practices.

166

For example, in 1997, the FTC

separately sued America Online, CompuServe, and Prodigy, alleging that each company

had offered “free” trial periods that resulted in unexpected charges to consumers.

167

One

Prodigy advertisement, for example, touted a “Free Trial” and “FREE 1

MONTH’S

MEMBERSHIP” conspicuously, while a fine print statement at the bottom of the back

panel of the advertisement stipulated: “Usage beyond the trial offer will result in extra

161

Id. (Apr. 17, 2001) (consent order), available at http://www.ftc.gov/os/2001/04/aoltwdo.pdf.

162

Id.

163

Id.

164

See, e.g., Cablevision Sys. Corp., 125 F.T.C. 813 (1998) (consent order); Summit Commun. Group, 120

F.T.C. 846 (1995) (consent order).

165

See Statement of Chairman Majoras, Commissioner Kovacic, and Commissioner Rosch Concerning the

Closing of the Investigation into Transactions Involving Comcast, Time Warner Cable, and Adelphia

Communications (Jan. 31, 2006) (FTC File No. 051-0151); see also Statement of Commissioners Jon

Leibowitz and Pamela Jones Harbour (Concurring in Part, Dissenting in Part), Time

Warner/Comcast/Adelphia (Jan. 31, 2006) (FTC File No. 051-0151). Both statements are available at

http://www.ftc.gov/opa/2006/01/fyi0609.htm

166

See, e.g., Am. Online, Inc. & CompuServe Interactive Servs., Inc., FTC Dkt. No. C-4105 (Jan. 28, 2004)

(consent order), available at http://www.ftc.gov/os/caselist/0023000/0023000aol.shtm

; Juno Online Servs.,

Inc., FTC Dkt. No. C-4016 (June 25, 2001) (consent order), available at

http://www.ftc.gov/os/caselist/c4016.shtm

167

See Am. Online, Inc., FTC Dkt. No. C-3787 (Mar. 16, 1998) (consent order), available at

http://www.ftc.gov/os/1997/05/ameronli.pdf; CompuServe, Inc., 125 F.T.C. 451 (1998) (consent order);

Prodigy, Inc., 125 F.T.C. 430 (1998) (consent order).

fees, even during the first month.”

168

Other alleged misrepresentations included AOL’s

failure to inform consumers that fifteen seconds of connect time was added to each online

session (in addition to the practice of rounding chargeable portions of a minute up to the

next whole minute),

169

as well as its misrepresentation that it would not debit customers’

bank accounts before receiving authorization.

170

The settlement orders in these matters

prohibited the companies from, among other things, misrepresenting the terms or

conditions of any trial offer of online service. Although all three matters involved dial-

up, or narrowband, Internet access, the orders are not limited by their terms to

narrowband services.

More recently, in the matter of FTC v. Cyberspace.com,

171

the federal district

court for the Western District of Washington granted summary judgment in favor of the

FTC, finding, among other things, that the defendants had violated the FTC Act by

mailing false or misleading purported rebate or refund checks to millions of consumers

and businesses without disclosing, clearly and conspicuously, that cashing the checks

would prompt monthly charges for Internet access services on the consumers’ and

businesses’ telephone bills. Following a trial on the issue of consumer injury, the court

ordered the defendants to pay more than $17 million to remedy the injury caused by their

fraudulent conduct. The Court of Appeals for the Ninth Circuit affirmed the trial court’s

liability finding last year.

172

In addition, the FTC has brought numerous cases involving the hijacking of

consumers’ modems.

173

For example, in FTC v. Verity International Ltd.,

174

the

Commission alleged that the defendants orchestrated a scheme whereby consumers

seeking online entertainment were disconnected from their regular ISPs and reconnected

to a Madagascar phone number. The consumers were then charged between $3.99 and

168

Prodigy, 125 F.T.C. at 430 exhibit A (complaint). Similar complaints were lodged against America

Online and CompuServe.

169

For example, “an online session of 2 minutes and 46 seconds, with the 15 second supplement, totals 3

minutes and 1 second and is billed as 4 minutes.” Am. Online, FTC Dkt. No. C-3787 at 4 exhibit E

(complaint).

170

See id. at 5-6 exhibit F.

171

No. C00-1806L, 2002 U.S. Dist. LEXIS 25565 (W.D. Wash. July 10, 2002), aff’d, 453 F.3d 1196 (9th

Cir. 2006).

172

Cyberspace.com, 453 F.3d at 1196.

173

A list of FTC enforcement actions involving the Internet and online services generally, and modem

hijacking allegations in particular, can be found at http://www.ftc.gov/bcp/internet/cases-internet.pdf

These actions include the following: FTC v. Sheinkin, No. 2-00-3636-18 (D.S.C. 2001); FTC v. RJB

Telcom, Inc., No. CV 00-2017 PHX SRB (D. Ariz. 2000); FTC v. Ty Anderson, No. C 00-1843P (W.D.

Wash. 2000); FTC v. Audiotex Connection, Inc., No. CV-97-0726 (DRH) (E.D.N.Y. 1997).

174

335 F. Supp. 2d 479 (S.D.N.Y. 2004), aff’d in part, rev’d in part, 443 F.3d 48 (2d Cir. 2006), cert.

denied, 127 S. Ct. 1868 (2007).

$7.78 per minute for the duration of each connection. In that case, AT&T and Sprint –

which were not parties to the FTC enforcement action – had carried the calls connecting

the consumers’ computers to the defendants’ servers. Consumers were billed at AT&T’s

and Sprint’s filed rates for calls to Madagascar. The defendants therefore argued that the

entertainment service in question was provided on a common carrier basis and thus

outside the FTC’s jurisdiction. One defendant also claimed to be a common carrier itself

and hence beyond FTC jurisdiction. Although both the District Court and the Court of

Appeals rejected those arguments, the FTC had to expend substantial time and resources

litigating the question of jurisdiction.

175

As the Verity case demonstrates, enforcement difficulties posed by the common

carrier exemption are not merely speculative. The FTC regards the common carrier

exemption in the FTC Act as outmoded and, as it creates a jurisdictional gap, an obstacle

to sound competition and consumer protection policy. As the FTC has explained before

Congress, technological advances have blurred traditional boundaries between

telecommunications, entertainment, and high technology.

176

For example, providers

routinely include telecommunications services, such as telephone service, and non-

telecommunications services, such as Internet access, in bundled offerings. As the

telecommunications and Internet industries continue to converge, the common carrier

exemption is likely to frustrate the FTC’s efforts to combat unfair or deceptive acts and

practices and unfair methods of competition in these interconnected markets.

Finally, based on the above discussion of the FTC’s jurisdiction over broadband

services, three general points may be in order. First, as the investigations and

enforcement actions described above suggest, the FTC has both authority and experience

in the enforcement of competition and consumer protection law provisions pertinent to

broadband Internet access. Second, the FTC Act provisions regarding “[u]nfair methods

of competition in or affecting commerce, and unfair or deceptive acts or practices in or

affecting commerce,” are general and flexible in nature, as demonstrated by judicial and

administrative decisions across diverse markets.

177

Third, the FTC’s investigative and

enforcement actions have been party- and market-specific; that is, neither the general

body of antitrust and consumer protection law nor the FTC’s enforcement and policy

record determines any particular broadband connectivity policy or commits the

Commission to favoring any particular model of broadband deployment.

175

In response to a request from the district court, the FCC filed an amicus brief in support of the FTC’s

jurisdiction in this matter. See Verity, 443 F.3d at 56, 61.

176

See FTC Jurisdiction over Broadband Internet Access Services: Hearing Before the S. Comm. on the

Judiciary, 109th Cong. 9-11 (2006) (statement of William E. Kovacic, Comm’r, FTC), available at

http://www.ftc.gov/opa/2006/06/broadband.shtm

177

“Congress has deliberately left these phrases undefined so that the parameters of the FTC’s powers and

the scope of its administrative and judicial functions could be responsive to a wide variety of business

practices.” ABA

SECTION OF ANTITRUST LAW, ANTITRUST LAW DEVELOPMENTS 643 & n.4 (6th ed. 2007)

(citing FTC v. Sperry & Hutchinson Co., 405 U.S. 233, 239-44 (1972); FTC v. R.F. Keppel & Bro., 291

U.S. 304, 310-12 (1934)).

B. FCC Jurisdiction under the Communications Act

As noted above, FCC jurisdiction over broadband services arises under the

Communications Act.

178

Central to the broadband discussion is a distinction under that

Act between “telecommunications services” and “information services.”

179

The former,

but not the latter, are subject to substantial mandatory common carrier regulations under

Title II of the Communications Act.

180

While not subject to the Title II common carrier

regulations, information services are treated by the FCC as subject to its general,

ancillary jurisdiction under Title I of the Communications Act.

181

Under Title II, providers of telecommunications services are bound to, among

other things, enable functional physical connections with competing carriers,

182

at “just

and reasonable” rates,

183

which the FCC may prescribe,

184

and are prohibited from

178

47 U.S.C. §§ 151 et seq.

179

Under the Communications Act, an “information service . . . means 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). In contrast, “‘telecommunications

service’ means the offering of telecommunications for a fee directly to the public . . . regardless of the

facilities used,” id. § 153(46), and “‘telecommunications’ means the transmission, between or among points

specified by the user, of information of the user’s choosing, without change in the form or content of the

information as sent and received.” Id. § 153(43). In brief, to act simply as a transmitter or transducer of

information is to provide a telecommunications service, whereas to act as a transformer of information is to

provide an information service.

180

The Communications Act is divided into seven Titles. See generally 47 U.S.C. §§ 151 et seq. Under

Title I are “General Provisions,” including, for example, the purposes of the Act, definitions, the

establishment of the FCC, and the structure and operations of the FCC. Under Title II are the “Common

Carriers” provisions, including, among others, common carrier regulations and “Universal Service”

requirements. Under Title III are “Provisions Relating to Radio.” Under Title IV are “Procedural and

Administrative Provisions.” Under Title V are “Penal Provisions.” Under Title VI are provisions relating

to “Cable Communications.” Finally, miscellaneous additional provisions are included under Title VII.

181

See, e.g., In re Appropriate Framework for Broadband Access to the Internet Over Wireline Facilities,

20 FCC Rcd 14853, 14914 (2005) (report and order and notice of proposed rulemaking) (“We recognize

that . . . the predicates for ancillary jurisdiction are likely satisfied for any consumer protection, network

reliability, or national security obligation that we may subsequently decide to impose on wireline

broadband Internet access service providers.”). Although the scope of the FCC’s ancillary jurisdiction over

broadband services has not been defined by the courts, it should be noted that the Supreme Court, in dicta,

has recognized the application of the FCC’s ancillary jurisdiction over information service providers. See

Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 976 (2005).

182

47 U.S.C. § 201(a).

183

Id. § 201(b).

184

Id. § 205.

making “any unjust or unreasonable discrimination in charges, practices, classifications,

regulations, facilities, or services . . . .”

185

There are, however, several important qualifications on these Title II common

carrier requirements. First, the Communications Act expressly provides for regulatory

flexibility to facilitate competition. In particular, with regard to telecommunications

carriers or services, the FCC

shall forebear from applying any regulation or any provision of this Act . .

. if the Commission determines that–(1) enforcement . . . is not necessary

to ensure that the charges, practices, classifications, or regulations . . . are

just and reasonable and are not unjustly or unreasonably discriminatory;

(2) enforcement . . . is not necessary for the protection of consumers; and

(3) forbearance from applying such provision or regulation is consistent

with the public interest.

186

In addition, in determining such “public interest,” the FCC must “consider whether

forbearance from enforcing the provision or regulation promotes competitive market

conditions.”

187

Finally, the Communications Act expressly states that “[i]t shall be the

policy of the United States to encourage the provision of new technologies and services

to the public.”

188

As a consequence, any person “(other than the Commission) who

opposes a new technology or service proposed to be permitted under this Act shall have

the burden to demonstrate that such proposal is inconsistent with the public interest.”

189

C. Regulatory and Judicial Clarification

As noted above, a series of regulatory and judicial decisions have helped to clarify

both the distinction between information and telecommunications services and the status

of broadband services as information services. That clarification is, to an extent, in

tension with early regulatory and judicial attempts to grapple with the novel technologies

that enabled the provision of Internet access. For example, in 1980, the FCC

promulgated rules designed to address, among other things, the growing commerce in

data-processing services available via telephone wires (the “Computer II Rules”).

190

With reference to those rules, the FCC subsequently applied certain common carrier

obligations, such as non-discrimination, to local telephone companies providing early

185

Id. § 202.

186

Id. § 160(a).

187

Id. § 160(b).

188

Id. § 157(a).

189

Id. § 160(b).

190

See In re Amendment of Section 64.702 of the Comm’n’s Rules & Regulations (Second Computer

Inquiry), 77 F.C.C.2d 384, 417-23 (1980) [hereinafter Computer II Rules].

DSL services.

191

Further, as recently as 2000, the Court of Appeals for the Ninth Circuit

held that “the transmission of Internet service to subscribers over cable broadband

facilities is a telecommunications service under the Communications Act.”

192

Still, the FCC’s current view that broadband services are information services has

its roots in earlier decisions by the FCC and the courts. The same Computer II Rules that

grounded the early DSL determination distinguished between “basic” and “enhanced”

services and did not subject the latter to Title II common carrier regulation.

193

In the

following decade, the FCC recognized that ISPs provide not just “a physical connection

[to the Internet], but also . . . the ability to translate raw Internet data into information

[consumers] may both view on their personal computers and transmit to other computers

connected to the Internet.”

194

Moreover, the 1998 Universal Service Report regarded

“non-facilities-based” ISPs – those that do not own their own transmission facilities –

solely as information service providers.

195

Indeed, even the Ninth Circuit opinion that

held that ISPs offering cable broadband were offering telecommunications services

recognized that, under the Communications Act and FCC implementing regulations, a

significant portion of those services were information services.

196

In 2000, the FCC issued a Notice of Inquiry to resolve, among other things, the

application of the Communications Act’s information/telecommunications distinction to

cable broadband ISPs.

197

In its subsequent declaratory ruling in 2002, the FCC

concluded that broadband cable Internet access services were information services, not

191

In a 1998 order, the FCC found, among other things, that incumbent local exchange carriers are subject

to various interconnection obligations under Title II of the Communications Act. See In re Deployment of

Wireline Servs. Offering Advanced Telecomms. Capability, 13 FCC Rcd 24011 (1998) (memorandum

opinion and order and notice of proposed rulemaking). The FCC noted that, although DSL and other

advanced services could “also be deployed using other technologies over satellite, cable, and wireless

systems, [it would] limit the discussion here to wireline services, because none of the petitioners raise

issues about these other technologies.” Id. at 24016 n.11. See also GTE Operating Cos. Tariff No. 1, 13

FCC Rcd 22466 (1998).

192

AT&T Corp. v. City of Portland, 216 F.3d 871, 880 (9th Cir. 2000).

193

See Computer II Rules, 77 F.C.C.2d at 428-32.

194

In re Fed.-State Joint Bd. on Universal Serv., 13 FCC Rcd 11501, 11531 (1998).

195

See id. at 11530.

196

See AT&T, 216 F.3d at 877-78.

197

In re Inquiry Concerning High-Speed Access to the Internet Over Cable & Other Facilities, 15 FCC Rcd

19287 (2000) (notice of inquiry). As noted above, this notice of inquiry had been expressly limited in its

application to broadband services provided by local telephone companies over wireline. Prior to 2000, the

FCC had not ruled on the application of common carrier obligations to broadband services provided via

cable. It sought, in this notice of inquiry, “to instill a measure of regulatory stability in the market,” and to

resolve a split in the Circuit courts regarding the regulatory status of “cable modem” broadband services.

See id. at 19288 & n.3 (comparing AT&T, 216 F.3d 871 with Gulf Power Co. v. FCC, 208 F.3d 1263 (11th

Cir. 2000)).

telecommunications services, and hence not subject to common carrier regulation under

Title II.

198

In reaching that conclusion, the FCC emphasized the information coding,

storage, and transformation processes that were central to such services, as it had in

concluding that non-facilities-based services were information services in its Universal

Service Report.

199

Moreover, the FCC concluded that there was no principled or

statutory basis for treating facilities-based and non-facilities-based services differently, as

both offered “a single, integrated service that enables the subscriber to utilize Internet

access service . . . .”

200

In response, several parties sought judicial review of the FCC’s determination in a

dispute eventually heard by the Supreme Court, in National Cable &

Telecommunications Association v. Brand X Internet Services (“Brand X”).

201

In Brand

X, the Court upheld the FCC’s determination that cable broadband is an information

service as a reasonable construction of the Communications Act, reversing a Ninth

Circuit decision that had relied on City of Portland as precedent.

202

In the wake of the Brand X decision, the FCC has continued to expand, platform

by platform, upon the broadband policy defended in that case. In 2005, the FCC released

the Appropriate Framework for Broadband Access to the Internet over Wireline

Facilities (“Wireline Order”), in which it reclassified wireline broadband Internet access

service by facilities-based carriers as an information service.

203

That reclassification

pertains to both “wireline broadband Internet access service . . . [and] its transmission

component,”

204

and is independent of the underlying technology employed.

205

The

198

In re Inquiry Concerning High-Speed Access to the Internet Over Cable & Other Facilities, 17 FCC Rcd

4798, 4821-22 (2002) (declaratory ruling and notice of proposed rulemaking).

199

Id. at 4820-23.

200

Id. at 4823.

201

545 U.S. 967 (2005).

202

Id. at 973-74. It should be noted that Brand X is fundamentally a Chevron decision. That is, the Court

did not examine the question of the status of cable broadband services as an abstract or de novo issue of

statutory construction. Rather, the Court held that the FCC’s ruling was – because based on reasonable

policy grounds – a permissible resolution of ambiguous statutory language in the Telecommunications Act

of 1996, given the FCC’s authority under the Communications Act, the Administrative Procedures Act, and

standards of agency deference the Court had articulated in Chevron v. NRDC. See id. at 973 (citing

Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984) and 5 U.S.C. §§ 551 et seq.).

203

In re Appropriate Framework for Broadband Access to the Internet over Wireline Facilities, 20 FCC

Rcd 14853 (2005) (report and order and notice of proposed rulemaking).

204

Id. at 14856.

205

Id. at 14860 n.15 (“We stress that our actions in this Order are limited to wireline broadband Internet

access service and its underlying broadband transmission component, whether that component is provided

over all copper loops, hybrid copper-fiber loops, a fiber-to-the-curb or fiber-to-the-premises (FTTP)

network, or any other type of wireline facilities, and whether that component is provided using circuit-

switched, packet-based, or any other technology.”).

Wireline Order does, however, permit facilities-based wireline carriers to elect to provide

broadband transmission service on a common carrier basis.

206

In 2006, the FCC released an order in which it classified broadband-over-

powerline Internet access services as information services.

207

Also in 2006, the FCC

granted – by operation of law – Verizon’s petition for forbearance from Title II and

Computer Inquiry Rules

208

with respect to its broadband services.

209

Verizon had asked

for forbearance “from traditional common-carriage requirements for all broadband

services,” seeking relief chiefly with regard to certain commercial broadband services not

expressly addressed in the Wireline Order or other rulemaking.

210

Most recently, the FCC clarified more generally the status of wireless services as

information services, issuing in 2007 a declaratory ruling finding: (1) “that wireless

broadband Internet access service is an information service”; (2) that while the

underlying transmission component of such service is “telecommunications,” offering

telecommunications transmission “as a part of a functionally integrated Internet access

service is not ‘telecommunications service’ under section 3 of the Act”; and (3) “that

206

Id.

207

In re United Power Line Council’s Petition for Declaratory Ruling Regarding the Classification of

Broadband Over Power Line Internet Access Serv. as an Info. Serv., 21 FCC Rcd 13281 (2006)

(memorandum opinion and order).

208

See In re Regulatory & Policy Problems Presented by the Interdependence of Computer & Commun.

Servs. & Facilities, 28 F.C.C.2d 267 (1971) (final decision and order) (“Computer I”); In re Amendment of

Section 64.702 of the Comm’n’s Rules & Regulations (Second Computer Inquiry), 77 F.C.C.2d 384 (1980)

(final decision) (“Computer II”); In re Computer III Further Remand Proceedings: Bell Operating Co.

Provision of Enhanced Servs., 14 FCC Rcd 4289 (1999) (report and order). Collectively, these matters are

known as the “Computer Inquiry Rules.”

209

See Press Release, FCC, Verizon Telephone Companies’ Petition for Forbearance from Title II and

Computer Inquiry Rules with Respect to Their Broadband Services Is Granted by Operation of Law (Mar.

20, 2006), available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-264436A1.pdf

(explaining

that a forbearance petition will be deemed granted if the FCC does not deny the petition within one year of

receipt, unless one-year period is extended by the FCC). Although the FCC did not explicitly grant such

relief, “the effect given to the petition by operation of law grants Verizon’s further broadband relief,

continuing our policy to encourage new investment.” In re Petition of the Verizon Tel. Cos. for

Forbearance under 47 U.S.C. § 160(c) from Title II & Computer Inquiry Rules with Respect to Their

Broadband Servs., WC Docket 04-440 (2006), 2006 FCC LEXIS 1333 (Chairman Martin & Comm’r Tate,

concurring).

210

Such services included: (1) packet-switched services capable of 200 Kbps in each direction and (2)

certain optical networking, hubbing, and transmission services. See In re Petition of the Verizon Tel. Cos.

for Forbearance under 47 U.S.C. § 160(c) from Title II & Computer Inquiry Rules with Respect to Their

Broadband Servs., WC Docket 04-440 (Feb. 7, 2006) (ex parte letter from Verizon Tel. Cos.), available at

http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6518324844

mobile wireless broadband Internet access service is not a ‘commercial mobile service’

under section 332 of the Act.”

211

Thus, over the past few years, the FCC has essentially unified the regulatory

status of cable, wireline, powerline, and wireless broadband Internet access services as

information services that are not subject to Title II common carrier requirements.

212

doing so, the FCC has focused on the abstract functional properties of ISPs as they

ranged across varying implementations or platforms. Underlying this unification has

been a significant degree of deregulation across broadband technologies, in keeping with

the statutory interest under the Communications Act in furthering competition and the

development of new technologies.

213

The FCC has nonetheless continued to demonstrate an interest in, and

commitment to, broadband Internet access. Certain policy statements have sought to

guide industry conduct to avoid both FCC enforcement actions and the “potentially

destructive” impact of overbroad and premature regulation of an “emerging market.”

214

In 2004, then-FCC Chairman Michael Powell challenged the industry to preserve four

“Internet Freedoms” to that end. They were:

(1) The “Freedom to Access Content . . . consumers should have access to their

choice of legal content” (within “reasonable limits” imposed by legitimate

network management needs);

211

In re Appropriate Regulatory Treatment for Broadband Access to the Internet Over Wireless Networks,

22 FCC Rcd 5901, 5901-02 (2007) (declaratory ruling).

212

See id. (“This approach is consistent with the framework that the Commission established for cable

modem Internet access service, wireline broadband Internet access service, and Broadband over Power

Line (BPL) – enabled Internet access service and it establishes a minimal regulatory environment for

wireless broadband Internet access service that promotes our goal of ubiquitous availability of broadband to

all Americans.”) (citations omitted).

213

See, e.g., Assessing the Communications Marketplace: A View from the FCC: Hearing Before the S.

Comm. on Commerce, Sci., & Transp., 110th Cong. 2 (2007) (statement of Kevin J. Martin, Chairman,

FCC), available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-270192A1.pdf

(“In 2005, the

Commission created a deregulatory environment that fueled private sector investment. . . . Broadband

deployment has been our top priority at the Commission, and we have begun to see some success as a result

of our efforts.”); see also, e.g., Thorne, Tr. II at 34 (“Over the past ten years, the policy of Congress and the

Federal Communications Commission has been to encourage investment and innovation in broadband

networks. This policy has been wildly successful.”). In addition, the FCC had undertaken to expand the

supply of broadband access services by, for example, promoting the use of unlicensed spectrum in rural

areas. See In re Implementation of the Commercial Spectrum Enhancement Act & Modernization of the

Comm’n’s Competitive Bidding Rules & Procedures, 20 FCC Rcd 11268 (2005) (declaratory ruling and

notice of proposed rulemaking) (implementing Enhance 911 Services Act, Pub. L. No. 108-494, 118 Stat.

3986, Title II (2004)). See infra Chapter VI.D for a more detailed discussion of federal spectrum policies.

214

Michael K. Powell, Chairman, FCC, Keynote Address at the Silicon Flatirons Symposium: Preserving

Internet Freedom: Guiding Principles for the Industry (Feb. 8, 2004), available at

http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-243556A1.pdf

(2) The “Freedom to Use Applications . . . consumers should be able to run the

applications of their choice” (within service plan limits and provided the

applications do not “harm the provider’s network”);

(3) The “Freedom to Attach Personal Devices . . . consumers should be permitted

to attach any devices they choose to the connection in their homes” (within

service plan limits, provided the devices do not “harm the provider’s network

or enable theft of service”); and

(4) The “Freedom to Obtain Service Plan Information . . . consumers should

receive meaningful information regarding their service plans” (so that

“broadband consumers can easily obtain the information they need to make

rational choices.”).

215

With some modification, those four Internet Freedoms were incorporated into an

FCC policy statement (“Broadband Policy Statement”), issued to accompany the

Wireline Order in 2005.

216

Recast as FCC principles, they included:

(1) The ability of consumers to “access the lawful Internet content of their

choice”;

(2) the ability of consumers to “run applications and use services of their choice,

subject to the needs of law enforcement”;

(3) the ability of consumers to “connect their choice of legal devices that do not

harm the network”; and

(4) the existence of “competition among network providers, application and

service providers, and content providers.”

217

In approving the AT&T/SBC and Verizon/MCI mergers in 2005, the FCC

required the companies to adhere to connectivity principles set forth in its Broadband

Policy Statement for a period of two years.

218

More recently, in approving the

215

Id. (italics included in published version of address).

216

See In re Appropriate Framework for Broadband Access to the Internet over Wireline Facilities, 20 FCC

Rcd 14986 (2005) (policy statement).

217

Id. Also in 2005 – prior to issuance of the Wireline Order – the FCC took enforcement action against

allegedly discriminatory behavior by an ISP. In re Madison River Communs., LLC, 20 FCC Rcd 4295,

4297 (2005). The resulting consent decree in that matter required a small North Carolina ISP to “not block

ports used for VoIP applications or otherwise prevent customers from using VoIP applications.” Id.

Because the FCC used its Title II authority in this case, under which it can regulate common carrier

services, this case may not be precedent for future enforcement authority over such services now

characterized as information services and regulated under the FCC’s Title I ancillary jurisdiction. See also

infra Chapters VII.B and IX.B for additional discussion of the Madison River matter.

218

See In re SBC Communs. Inc. & AT&T Corp. Applications for Approval of Transfer of Control, 20

FCC Rcd 18290 (2005) (memorandum opinion and order) (especially appendix F); In re Verizon

Communs. Inc. & MCI Inc. Applications for Approval of Transfer of Control, 20 FCC Rcd 18433 (2005)

(memorandum opinion and order) (especially appendix G).

The DOJ also examined the proposed mergers and successfully sought, under the Tunney Act, the

divestiture of certain assets as conditions to such mergers. See United States v. SBC Communs., Inc., Civ.

AT&T/BellSouth merger, the FCC required the combined company to agree not to

provide or sell (for a period of thirty months following the merger closing date) “any

service that privileges, degrades, or prioritizes any packet transmitted over

AT&T/BellSouth’s wireline broadband Internet access services based on its source,

ownership, or destination.”

219

Most recently, the FCC announced an inquiry “to better understand the behavior

of participants in the market for broadband services.”

220

Among other things, the FCC is

seeking information regarding the following:

• How broadband providers are managing Internet traffic on their

networks today;

• Whether providers charge different prices for different speeds or

capacities of service;

• Whether our policies should distinguish between content providers

that charge end users for access to content and those that do not;

and

• How consumers are affected by these practices.

221

In addition, the FCC has asked for comments “on whether the [Broadband] Policy

Statement should incorporate a new principle of nondiscrimination and, if so, how would

‘nondiscrimination’ be defined, and how would such a principle read.”

222

Action Nos. 05-2102 (EGS) & 05-2103 (EGS), 2007 WL 1020746 (D.D.C. Mar. 29, 2007). In particular,

the merging parties were required to divest themselves of long-term interests in certain local private line, or

special access, facilities. Id. at *5 (noting that “[a]part from the difference in geographic scope due to the

identities of the parties, the proposed final judgments are practically identical and require the same type of

divestitures.”). See infra Chapter VI.B for a discussion of special access facilities and their relationship

with broadband Internet services.

219

In re AT&T Inc. & BellSouth Corp. Application for Transfer of Control, 22 FCC Rcd 5662 (2006)

(memorandum opinion and order). Two FCC Commissioners issued a concurring statement expressing

their view that “[t]he conditions regarding net-neutrality have very little to do with the merger at hand and

very well may cause greater problems than the speculative problems they seek to address.” Id. at 5826

(Chairman Martin & Comm’r Tate, concurring).

The DOJ also reviewed the AT&T/BellSouth merger, examining, among other things, the merged

firm’s ability or incentive to favor its own Internet content over that of its rivals. See Press Release, DOJ,

Statement by Assistant Attorney General Thomas O. Barnett Regarding the Closing of the Investigation of

AT&T’s Acquisition of BellSouth 3 (Oct. 11, 2006), available at

http://www.usdoj.gov/atr/public/press_releases/2006/218904.pdf

. The DOJ concluded its investigation last

October, finding that “the merger would neither significantly increase concentration in markets for the

provision of broadband services to end users nor increase Internet backbone market shares significantly.”

Id.

220

Press Release, FCC, FCC Launches Inquiry into Broadband Market Practices (Mar. 22, 2007), available

at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-271687A1.pdf.

221

Id.

* * *

The legal and regulatory developments discussed above have prompted the

current debate over network neutrality regulation. In the next Chapter, we provide an

overview of the arguments in favor and against such regulation that have been put forth

to date.

222

Id.

III. OVERVIEW OF ARGUMENTS IN FAVOR OF AND AGAINST

NETWORK NEUTRALITY REGULATION

Technology experts have recognized since the Internet’s earliest days that

network resources are scarce and that traffic congestion may lead to reduced

performance.

223

Although such experts continued to explore different data-transmission

protocols and the viability of market-based pricing mechanisms through the 1980s and

1990s, the current debate over broadband connectivity policy did not accelerate until

more recently.

224

At about the same time that the FCC began its cable broadband

rulemaking proceedings in 2000,

225

data routing technologies advanced to the point

where some network operators began openly to consider using prioritization and other

active management practices to improve network management and provide certain

premium services for a fee.

226

Various interested parties, including some content and applications providers,

non-facilities-based providers of Internet services, and third-party commentators, have

expressed concern about network operators’ use of these routing technologies in an

environment that is not subject to common carrier regulation. Some of them, therefore,

have proposed that the transmission of data on the Internet be subject to some type of

“network neutrality” rules that forbid or place restraints on some types of data or price

discrimination by network operators.

227

This Chapter summarizes the major arguments in

favor of (in Section A) and against (in Section B) the enactment of some form of network

neutrality regulation put forth to date.

228

Arguments involving data discrimination and

prioritization, as well as competition and consumer protection issues, are addressed in

more detail below in Chapters IV through VIII of this Report.

223

See supra Chapter I.A.

224

See generally Vinton G. Cerf & David Farber, The Great Debate: What is Net Neutrality?, Hosted by

the Center for American Progress (July 17, 2006), available at

http://www.americanprogress.org/kf/060717%20net%20neutrality.pdf

; Tim Wu & Christopher Yoo,

Keeping the Internet Neutral?: Timothy Wu and Christopher Yoo Debate (Vand. Pub. Law, Research Paper

No. 0-27, 2006), available at http://ssrn.com/abstract=953989

225

See supra Chapter II.C for a discussion of relevant FCC proceedings.

226

See supra Chapter I.A.

227

See, e.g., Tim Wu, Network Neutrality, Broadband Discrimination, 2 J. ON TELECOMM. & HIGH TECH.

L. 141, 151 (2005) (“Over the history of communications regulation, the Government has employed both

common carriage requirements (similar to the neutrality regime discussed here) and limits on vertical

integration as [a] means of preventing unwanted discrimination.”). See also Cohen, Tr. II at 195 (arguing

that network neutrality regulation “is really a return to the status quo as where it was [in August 2005 and

before Brand X] so it’s not . . . a new set of regulations”).

228

This Chapter is not intended to be a comprehensive treatment of the many arguments put forth in favor

of and against network neutrality. Instead, this Chapter serves as a general survey of the types of

arguments raised by both sides of the network neutrality debate. Nor does this Chapter attribute every

single argument or variation thereon to every individual or entity that has made such arguments.

A. Arguments in Favor of Network Neutrality Regulation

Proponents of network neutrality regulation argue, among other things, that the

existing jurisdiction of the FCC, FTC, and DOJ, as well as oversight by Congress, are

insufficient to deal with what they predict will be inevitable and far-reaching harms from

so-called non-neutral practices. They suggest that after recent legal and regulatory

determinations, providers of certain broadband Internet services now have the legal

authority to act as gatekeepers of content and applications on their networks.

Principally, these advocates express concern about: (1) blockage, degradation,

and prioritization of content and applications; (2) vertical integration by network

operators into content and applications; (3) effects on innovation at the “edges” of the

network (i.e., by content and applications providers); (4) lack of competition in “last-

mile” broadband services; (5) legal and regulatory uncertainty in the area of Internet

access; and (6) diminution of political and other expression on the Internet. Net

neutrality proponents argue that various harms are likely to occur in the absence of

neutrality regulation and that it will be difficult or impossible to return to the status quo if

non-neutral practices are allowed to become commonplace. Proponents thus see an

immediate need to enact neutrality regulation.

229

1. Concerns about Blockage and Degradation of Non-Favored Content

and Applications

Network neutrality advocates suggest that, without neutrality rules, network

operators will use packet-inspection technologies to favor the transmission of their own

content and applications, or those of their affiliates, over those of other providers instead

of offering the unrestricted access generally available to end users today.

230

They

frequently suggest that end users’ access to the wider Internet will become balkanized

and restricted to what network operators choose to display in their own proprietary

“walled gardens.” Proponents believe such walled gardens will look more like the

original America Online dial-up service or even an Internet version of cable television,

with access to only a limited number of favored sites. Proponents further point to

preferential practices in other industries, such as cable television and telephony, as

indications of the likelihood that network operators will adopt comparable practices in the

absence of net neutrality regulation.

231

229

See, e.g., Cohen, Tr. II at 150 (“I can’t take the view that we should start from the premise of wait until

it’s all destroyed before we do anything about it.”).

230

See, e.g., Wu, supra note 227. See also EARL W. COMSTOCK, WHAT IS NET NEUTRALITY? (2006),

available at http://www.comptel.org/content.asp?contentid=658

; G. Sohn, Tr. I at 98; Farrell, Tr. I at 220.

231

See, e.g., Lawrence Lessig & Robert W. McChesney, No Tolls on the Internet, WASH. POST, June 8,

2006, at A23. Lessig and McChesney suggest that “[w]ithout net neutrality, the Internet would start to look

like cable TV. A handful of massive companies would control access and distribution of content, deciding

what you get to see and how much it costs.” Id. See also Tulipane, Tr. I at 259-66. In Tulipane’s view,

“prioritization based on source or content will result in a closed network, just like the cable system today.”

Id. at 266. Similarly, Sohn suggests: “[s]hort of outright blocking, ISPs could engage in various forms of

Advocates of net neutrality point to certain statements by ISP executives as

evidence of their intent to treat some content and applications differently than others.

232

They cite to the Madison River

233

matter as evidence that network operators do, in fact,

have the technological means and incentive to actively degrade or outright block certain

content and applications.

234

They also question whether end users will be able to

determine readily why certain content and applications might be unavailable or executing

more slowly or less reliably than others.

235

Some also suggest that the introduction of

specialized, virtual private networks (“VPNs”) that require users to purchase premium

service packages foreshadows the advent of a balkanized, non-neutral Internet.

236

In particular, these proponents warn that network operators might try to disfavor

some content and applications by inhibiting or forbidding users from attaching related

devices to their networks, such as the VoIP phone equipment of competing Internet

telephony providers or VoIP-enabled mobile phones.

237

They also state that cable

companies have, in fact, blocked streaming video applications to protect their own cable

television businesses and that wireless phone companies have placed limits on the types

of content and applications that can be accessed using their wireless Internet services.

238

Some network neutrality proponents also contend that network operator bans on

the use of basic residential packages to operate VPNs, open-access Wi-Fi antennas that

support multiple users, home networks, and computer servers all amount to violations of

neutrality principles.

239

Some, but not all, proponents, however, believe that such

discrimination, and the fears [sic] that could have the practical effect of driving innovators to really have

now a practical need to seek deals with each recipient’s ISP.” D. Sohn, Tr. II at 227-28.

232

See supra note 64.

233

In Madison River, an ISP allegedly blocked its customers from accessing a competing VoIP provider.

The ISP entered into a consent decree with the FCC that prohibited the ISP from blocking ports used for

VoIP traffic. The ISP also made a voluntary payment of $15,000 to the U.S. Treasury. In re Madison

River Communs., LLC, 20 F.C.C.R. 4295, 4297 (2005).

234

See, e.g., Davidson, Tr. I at 227-28. For Davidson, “prioritization in the last mile creates real concerns.

Particularly, we are concerned that prioritization through router-based discrimination in the last mile

degrades computing services, and creates incentives to relegate some of those computing services to a slow

lane.” Id.

235

See supra Chapter I.C.5.

236

See, e.g., Yokubaitis, Tr. II at 108.

237

See, e.g., Libertelli, Tr. I at 73 (“[F]or Skype, network neutrality is about protecting our users’ ability to

connect to each other, whenever and wherever they want. We support net neutrality[] because it embodies

a policy of decentralized innovation.”).

238

See, e.g., John Windhausen, Jr., Good Fences Make Bad Broadband: Preserving an Open Internet

Through Net Neutrality 16-23 (Public Knowledge White Paper, 2006), available at

http://www.publicknowledge.org/pdf/pk-net-neutrality-whitep-20060206.pdf

239

See, e.g., id.

restrictions may be justified because they are meant to solve situations in which a few

users generate costs that are imposed on other users.

240

2. Concerns about Charging Content and Applications Providers for

Prioritized Data Delivery

Net neutrality advocates also express concern that, short of outright blockage or

active degradation, network operators will present certain content and applications to

users in a preferential manner in exchange for payment. They express concern that

network operators may, for example, use packet-inspection technology to provide quicker

load times for certain providers’ Web pages or faster and more consistent connections for

favored VoIP or streaming video providers.

241

Some network operators have, in fact,

indicated that they would like to offer certain prioritized services or other kinds of

quality-of-service guarantees in exchange for a premium fee.

242

Some neutrality advocates object to the idea of a network offering prioritized data

transmission or quality-of-service guarantees in exchange for payment.

243

That is, they

object to a deviation from the long-standing first-in-first-out and best-efforts transmission

characteristics of the Internet. They are concerned about the potential for prioritization to

result in blocking or degradation of non-favored content and applications. These

advocates are concerned that content and applications from providers affiliated with the

network operator or having a greater ability to pay will be available in a “fast lane,” while

others will be relegated to a “slow lane,” discriminated against, or excluded altogether.

244

Further, creating priority fast lanes, according to some advocates, necessarily would

240

See, e.g., Wu, supra note 227, at 152.

241

See, e.g., Editorial, Open Net, THE NEW REPUBLIC, June 26, 2006, available at

http://www.tnr.com/doc.mhtml?pt=oy4NRC5%2Bfnu%2Fm585FtGwlC%3D%3D

242

See infra Chapter III.B.

243

See, e.g., Davidson, Tr. I at 228. In his view:

[W]hat we’re worried about is in that context, the power to prioritize in the last mile

effectively becomes the power to control the applications and content that customers can

effectively use.

So, imagine, for example, that a last mile provider with market power might be

able to use prioritization to, for example, relegate a competing Voice over IP provider to

a lower quality slow lane. It might prevent a competing video provider – prevent a

competing video service from accessing a higher tier of priority necessary to provide

good service, and preference its own services instead.

Id. See also Tulipane, Tr. I at 259-66.

244

See, e.g., Davidson, Tr. I at 229-30. According to Davidson, “[w]e are concerned about creating a fast

lane tier of traffic that is susceptible of exclusive dealings.” Id. at 229. In his view, “prioritization that

provides an incentive to create slow lanes so that you can charge people for the fast lanes is something that

we think is problematic.” Id. at 230.

result in (intentionally or effectively) degraded service in the remainder of the network.

245

Likewise, some advocates object to the creation of private networks that might provide

prioritized data transmission or other forms of quality of service to only a limited number

of customers, arguing that this will represent the “end” of the Internet as we know it.

246

Some advocates, therefore, argue that content and applications providers should

not be allowed to pay a premium fee for prioritized data transmission, even if they want

to do so. They object, for example, to a possible two-sided market model where content

and applications providers pay networks for prioritization in the same way that merchants

subsidize the purchase price of a newspaper by paying for the placement of

advertisements in return for greater consumer exposure to their advertisements.

247

Instead, in this view, networks should be required to derive revenues principally from

providing Internet access to residential and business customers.

248

Some advocates who

object to prioritized data transmission would, however, allow network operators to charge

end users more for the consumption of larger amounts of bandwidth.

249

Other advocates do not strictly object to prioritization or quality of service for a

fee.

250

They argue, however, that different levels of prioritization should be offered on

uniform terms to all “similar” content and applications providers and that all end users be

245

See, e.g., id. at 228-30 (“[P]rioritization . . . in the last mile degrades competing services, and creates

incentives to relegate some of those competing services to a slow lane . . . [given] that the only way that

you can have a fast lane that you can charge for, that is useful, is if there are also slow lanes that are less

useful, and less attractive.”).

246

See, e.g., Lessig & McChesney, supra note 231. Lessig and McChesney predict that, without neutrality

rules, network operators will use data prioritization “to sell access to the express lane to deep-pocketed

corporations and relegate everyone else to the digital equivalent of a winding dirt road.” In their view,

“[n]et neutrality means simply that all like Internet content must be treated alike and moves at the same

speed over the network.” Id.

247

See Pepper, Tr. I at 87 (“The last set of questions on net neutrality concern who can be charged for what

service on broadband connections. Should the Internet access be funded solely by consumers, or can the

cost be shared with content providers and application providers?”).

248

See, e.g., Editorial, supra note 241 (“Net neutrality would prohibit all of this. Telecoms could make

money they way they always have – by charging homes and businesses for an Internet connection – but

they couldn’t make money from the content providers themselves.”). See also Sidak, Tr. I at 107 (“In other

words, they don’t have a problem with network operators and end users contracting for prioritized delivery.

The problem they have is . . . with suppliers of content.”).

249

See, e.g., Davidson, Tr. I at 228 (“Not all network management is anti-competitive prioritization. And

there are a lot of things I think many of us agree that are not problematic in this context. So, charging end

users, whether it’s businesses or consumers, more for more bandwidth, not a problem here.”). See also

OMSTOCK, supra note 230.

250

See, e.g., D. Sohn, Tr. II at 230. In Sohn’s view, network neutrality regulation “wouldn’t need to

involve a complete ban on all prioritization, even on the Internet part. I think in particular, an ISP should

be free to offer prioritization capability that enables subscribers to choose what services to use it with.” Id.

See also Cohen, Tr. II at 150 (“There are and should remain many networks on which network providers

are free to discriminate based on the source, ownership or destination of data . . . .”).

guaranteed a minimum level of access to the entire universe of Internet content.

251

Another advocate suggests that network operators should be free to create specialized

service parameters and to provide prioritized data transmission, but with a requirement

that networks also maintain a basic level of best-efforts Internet service.

252

Some network neutrality proponents further suggest that, as the speed of the

Internet continues to increase with the deployment of faster technologies like fiber-optic

wirelines and improved wireless transmissions, the issue of prioritization may become

irrelevant.

253

They suggest that when Internet speeds of upwards of 100 megabits per

second (“Mbps”) are widely available, first-in-first-out and best-efforts delivery at these

rates should be sufficient to transmit all Internet traffic without any problems, even for

advanced and time-sensitive applications. These proponents suggest that all congestion

and bandwidth scarcity issues will effectively disappear at these speeds and the issue of

prioritization will eventually be moot. A neutrality regime, therefore, can be seen as a

temporary remedy for a problem that ultimately will be outgrown and an important

measure that will prevent network operators from creating artificial scarcity in their

networks in the meantime to derive additional revenues by charging content and

applications providers for new types of data transmission.

254

Thus, some of these

251

See, e.g., Wilkie, Tr. I at 170 (“The caveat might be that you might want to add that tiering and offering

higher levels of prioritization are allowable, but they would have to be offered on a non-discriminatory

basis, or what economists call ‘second degree price discrimination,’ that is, the prices are functions of the

level of functionality offered, not the identity of the customer.”). See also G. Sohn, Tr. I at 128 (advocating

that if one content or applications provider negotiates a particular service arrangement with a network

operator, a second competing content or applications provider should “absolutely” be provided with an

identical arrangement by the operator without having to engage in separate negotiations).

252

See, e.g., Press Release, USC Annenberg Center, Annenberg Center Releases Principles for Network

Neutrality (2006), available at http://www.annenberg.edu/news/news.php?id=13. See also D. Sohn, Tr. II

at 226 (suggesting that the optimum outcome is “to keep this neutral open Internet at an acceptable level of

service, to keep that in existence even as experimentation with other networks . . . proceeds”).

253

See, e.g., Network Neutrality: Competition, Innovation, and Nondiscriminatory Access: Hearing Before

the S. Comm. on Commerce, Sci., & Transp., 109th Cong. (2006) (testimony of Gary R. Bachula, Vice

President, Internet2) [hereinafter Bachula Senate Testimony], available at

http://commerce.senate.gov/pdf/bachula-020706.pdf

; Bachula, Tr. II at 164-74. See also Davidson, Tr. I at

231 (“In most cases, the best way to deal with any concerns about prioritization is to provide better

broadband, higher bandwidth offerings to consumers.”).

254

According to Bachula:

When we first began to deploy our Internet2 network some eight years ago, our

engineers started with the assumption that we would have to find technical ways of

prioritizing certain bits, such as streaming video or video conferencing, in order to ensure

that they arrived without delay.

For a number of years, we seriously explored various quality of service

techniques, conducted a number of workshops and even convened an ongoing quality of

service working group, but as it developed, all of our research and practical experience

supported the conclusion that it was far more cost effective to simply provide more

bandwidth. It was cheaper to provide more bandwidth than to install these sophisticated

quality of service prioritization techniques.

proponents believe that, instead of allowing network operators to engage in prioritization,

policy makers should focus on creating incentives for the deployment of next-generation,

high-speed networks.

255

3. Concerns about Vertical Integration

Net neutrality proponents also express concern about the prospect of network

operators integrating vertically into the provision of content and applications. Proponents

argue that network operators now have the legal and technological ability to control both

their own physical networks and the ability of content and applications providers to reach

end users. Proponents further suggest that vertically integrated network operators will

favor their own content and applications, or those of their affiliates, over others.

256

Some

of these proponents, therefore, argue that network operators’ ability to vertically integrate

should be legally restricted or forbidden altogether.

257

4. Concerns about Innovation at the “Edges” of the Internet

Proponents suggest that if so-called non-neutral practices are allowed to flourish

in the core of the networks that comprise the Internet, innovation by content and

applications developers that are connected to the Internet’s “edges” will suffer. Some

proponents, for example, are concerned about the complexity and cost that content and

applications providers would experience if they had to negotiate deals with numerous

network operators worldwide. They suggest that content and applications providers will

need to expend considerable resources to negotiate and enter into prioritization

agreements or other preferential arrangements with numerous networks and that many

(particularly, small) companies will not be able to pay the fees that operators will demand

to reach end users in a competitive manner.

258

Thus, they fear that innovators will be

With enough bandwidth in the network, there is no congestion, and video bits do

not need preferential treatment. All the bits arrive fast enough even if intermingled.

Bachula, Tr. II at 169.

255

Robert D. Atkinson & Philip J. Weiser, A “Third Way” on Network Neutrality, 13 THE NEW ATLANTIS

47, 58-59 (2006), available at http://www.thenewatlantis.com/archive/13/TNA13-AtkinsonWeiser.pdf

These commentators suggest that Congress should allow companies investing in broadband networks to

expense new broadband investments in the first year and also extend the moratorium on federal, state, and

local broadband-specific taxes, but make it contingent upon provision of an open, best-efforts level of

Internet service. Id. See also generally Lehr, Tr. I at 36 (“[Over time, network] penetration saturates. And

so, revenues growth slows. And the question is that if we want the industry to continue to meet the growth

in traffic, we have to figure [out] what the incentives are.”).

256

See, e.g., Joseph Farrell, Open Access Arguments: Why Confidence is Misplaced, in NET NEUTRALITY

NET NEUTERING: SHOULD BROADBAND SERVICES BE REGULATED?, supra note 42, at 195.

257

See, e.g., Christian Hogendorn, Regulating Vertical Integration in Broadband: Open Access Versus

Common Carriage, 4 R

EV. NETWORK ECON. 19, 30 (2005).

258

See, e.g., Davidson, Tr. I at 224-33. According to Davidson, “[a]s our founders have said, two graduate

students in a dorm room with a good idea would not have been able to create this service if the first thing

that they had to do was to hire an army of lawyers and try to reach carriage agreements with providers all

blocked, actively degraded, or provided with low-priority data transmissions, and the

development of the next revolutionary Internet site or application may be inhibited. They

predict that spontaneous innovation will be precluded or forced to proceed through

established businesses already having significant capital and favored relationships with

network operators.

259

Similarly, net neutrality proponents sometimes argue that non-

profit and educational entities may be at a disadvantage relative to highly capitalized

businesses.

260

5. Concerns about “Last-Mile” Competition in Broadband Service

Net neutrality proponents typically argue that a cable-telephone duopoly exists in

most markets for last-mile broadband connections and that competition from only two

broadband providers is not sufficient to check the harms that they envision. Net

neutrality proponents generally do not believe that one of these competitors will provide

users with an acceptable, alternative open service if the other decides to pursue exclusive

deals or data prioritization. Proponents also typically express doubt about the potential of

newer technologies like wireless Internet and broadband over powerlines to provide in

the near future a robust, competitive alternative to the access offered by the cable and

telephone companies.

261

A related concern expressed by some network neutrality proponents is that last-

mile ISPs might not disclose to end users the ISPs’ differential treatment of certain data

and that they will be able to get away with such non-disclosure due to a lack of viable

competitive alternatives in the marketplace or the difficulty of tracing problems to ISPs’

practices. Proponents also suggest that, to the extent that such disclosures are made by

ISPs, many end users will not be able to readily understand them, making such

around the world.” Id. at 226. See also Cohen, Tr. II at 152 (“[Historically, Internet start-ups] did not have

to negotiate. They did not have to persuade or cajole network providers for special treatment.”); Center for

Creative Voices in Media, Public Comment 6, at 2 (“Artists must have the freedom to distribute their works

over the broadband Internet, and the American public must have the freedom to choose from among those

works, rather than have the cable and telephone broadband providers who overwhelmingly control the

market for broadband deny those freedoms and make those choices for them.”).

259

See, e.g., Mark. A. Lemley & Lawrence Lessig, The End of End-to-End: Preserving the Architecture of

the Internet in the Broadband Era, 48 UCLA L. R

EV. 925 (2001). Lemley and Lessig suggest that, “[i]f

that strategic actor owns the transmission lines itself, it has the power to decide what can and cannot be

done on the Internet. The result is effectively to centralize Internet innovation within that company and its

licensees.” Id. at 932. See also Farrell, Tr. I at 154 (“[T]here is a concern if you allow last mile providers

to make charges on content providers, there is a concern about possible expropriation of successful content

providers.”).

260

See, e.g., Reconsidering Our Communications Laws: Ensuring Competition and Innovation: Hearing

Before the S. Comm. on the Judiciary, 109th Cong. (2006) (statement of Jeff C. Kuhns, Senior Director,

Consulting and Support Services, Information Technology Services, The Pennsylvania State University),

available at http://judiciary.senate.gov/testimony.cfm?id=1937&wit_id=5418

261

See, e.g., Feld, Tr. II at 18-19; Putala, Tr. II at 29 (“The much heralded independent alternatives are still

tiny.”); Wu, Tr. II at 255 (“I have been hearing that for ten years. I’ve never met anyone who has a

connection, broadband over power line, and it has been used a million times . . . .”).

disclosures ineffective in checking potential ISP misconduct.

262

Some network neutrality

proponents also argue that the use of data packet inspection and other traffic analysis

technologies by network operators may give rise to privacy concerns that end users might

not readily recognize.

263

6. Concerns about Legal and Regulatory Uncertainty

Net neutrality advocates suggest that the FCC’s recently issued broadband

principles, its ancillary jurisdiction over broadband providers under Title I of the

Communications Act of 1934, and the antitrust laws are insufficient to prevent or police

potentially harmful conduct by broadband providers.

264

In particular, they argue that the

FCC’s broadband principles are not legally enforceable, that the full scope of its Title I

authority has yet to be determined, and that any remedial action is likely to result in years

of litigation and appeals, leaving the status of the Internet in doubt.

265

Neutrality

advocates argue that more concrete examples of alleged harms, beyond Madison River,

do not exist primarily because network operators have been on their best behavior in the

short time since recent legal and regulatory determinations were handed down, to avoid

attracting further scrutiny. Proponents argue that without further regulation, however,

network operators will likely engage in such practices in the future and that there will be

no practical way to prevent or remedy the resulting harms without a comprehensive, ex

ante regulatory regime.

266

7. Concerns about Political and Other Expression on the Internet

Advocates suggest that, without a network neutrality rule, operators will likely

engage in practices that will reduce the variety and quality of content available to users,

generally. In particular, they suggest network operators may degrade or block content

that they find to be politically or otherwise objectionable or contrary to their own

262

See, e.g., Kenney, Tr. II at 103 (“I think these disclosure issues are important, but I don’t think that’s the

issue here today. In fact, the elephant in the room is whether or not disclosure of prioritization practices is

sufficient to remedy the harm.”).

263

See, e.g., id. (“I don’t think anyone has a full understanding of what sort of security and vulnerability

issues are at stake with deep packet inspection technologies.”).

264

See, e.g., Libertelli, Tr. I at 117 (“[W]e’re talking about a policy statement [(the FCC principles)]; we’re

not necessarily talking about a binding rule of decision.”); Farrell, Tr. I at 159 (“I am not convinced that

anti-trust, as currently enforced, is going to do a good job on those potential problems.”).

265

See, e.g., Network Neutrality: Competition, Innovation, and Nondiscriminatory Access: Hearing Before

the H. Comm. on the Judiciary, Task Force on Telecom & Antitrust, 109th Cong. 23, 35 (2006) (prepared

statement of Earl W. Comstock, President and CEO, COMPTEL) [hereinafter Comstock House

Testimony], available at http://judiciary.house.gov/media/pdfs/printers/109th/27225.pdf

266

See, e.g., Misener, Tr. II at 142 (“[W]e really believe that it would be in consumers and industry’s best

interest for certainty and for a national policy to be set by the Federal Government at the very highest level

. . . .”).

business interests.

267

Neutrality advocates suggest that other types of speech, such as

individuals’ Web logs, may also be disfavored or blocked as the incidental result of an

operator’s more general decisions about favoring certain content providers over others.

268

This argument appears to be a variation on the suggestion that, without a neutrality

regime, innovation (or, in this case, speech) at the edges of the network will be

inhibited.

269

B. Arguments against Network Neutrality Regulation

Opponents of network neutrality regulation include facilities-based wireline and

wireless network operators, certain hardware providers, and other commentators. These

parties maintain that imposing network neutrality regulation will impede investment in

upgrading Internet access and may actually hamper innovation. They also argue that,

apart from the Madison River case, the harms projected by net neutrality proponents are

merely hypothetical and do not merit a new, ex ante regulatory regime.

Principally, these opponents argue that: (1) the Internet is not neutral and never

truly has been, and a neutrality rule would effectively set in stone the status quo and

preclude further technical innovation; (2) effective network management practices

require some data to be prioritized and may also require certain content, applications, and

attached devices to be blocked altogether; (3) there are efficiencies and consumer benefits

from data prioritization; (4) new content and applications also require this kind of

network intelligence; (5) network operators should be allowed to innovate freely and

differentiate their networks as a form of competition that will lead to enhanced service

offerings for content and applications providers and other end users; (6) prohibiting

network operators from charging different prices for prioritized delivery and other types

of quality-of-service assurances will reduce incentives for network investment generally

267

See, e.g., Bill D. Herman, Opening Bottlenecks: On Behalf of Mandated Network Neutrality, 59 FED.

COMM. L.J. 107, 118 (2007) (submitted to FTC as Public Comment 26) (“A broadband provider should no

more be able to stop a customer’s email or blog post due to its political content than a telephone company

should be permitted to dictate the content of customers’ conversations.”). See also Peha, Tr. I at 26 (“There

could also be content filtering for other reasons. Perhaps for political reasons I will want to limit access to

advocacy groups for issues I oppose, or candidates I oppose.”).

268

See, e.g., Barbara A. Cherry, Misusing Network Neutrality to Eliminate Common Carriage Threatens

Free Speech and the Postal System, 33 N.

KY. L. REV. 483, 507 (2006) (submitted to FTC as Public

Comment 8) (“If antitrust principles are insufficient to substitute for the functions that common carriage

and public utility obligations have served in providing access, then free speech rights of individuals will be

sacrificed to serve economic interests of corporate owners of broadband facilities.”); Feld, Tr. II at 15

(“Goal number . . . two is the Internet is open and diverse as it exists today or better. . . . The First

Amendment cares about this stuff. Our democracy depends on this stuff, and Congress has told us to

protect it as part of the policy. Any policy that doesn’t protect that, even if it is more economically

efficient, is a failed policy.”). But compare Thomas B. Leary, The Significance of Variety in Antitrust

Analysis, 68 A

NTITRUST L.J. 1007, 1019 (2001) (raising the question of “whether an increase or decrease in

available variety, by itself, merits independent consideration in antitrust analysis”).

269

See, e.g., G. Sohn, Tr. I at 134 (“The Internet actually takes away the gate keepers, so people can engage

in democratic discourse, eCommerce, innovation. It’s been great. And at a certain point, we have to ask

ourselves, do we want it to remain that way?”).

and prevent networks from recouping their investments from a broader base of

customers, a practice which might, in turn, reduce prices for some end users; (7) vertical

integration by network operators into content and applications and certain bundling

practices may produce efficiencies that ultimately benefit consumers; and (8) there is

insufficient evidence of potential harm to justify an entirely new regulatory regime,

especially when competition in broadband services is robust and intensifying and the

market is generally characterized by rapid, evolutionary technological change.

1. Historical and Existing Non-Neutrality of the Internet

Opponents of network neutrality regulation argue that the Internet is not, and

never truly has been, “neutral.”

270

These opponents generally agree that the first-in-first-

out and best-efforts characteristics of the TCP/IP data-transmission protocol have played

a significant role in the development of the Internet.

271

They point out, however, that

since the earliest days of the Internet, computer scientists have recognized that data

congestion may lead to reduced network performance and have thus explored different

ways of dealing with this problem.

272

Net neutrality opponents point out that all network routers must make decisions

about transmitting data and argue that such decisions invariably have implications that

may not be strictly uniform or neutral. In particular, they note that networks have long

employed “hot potato” routing policies that hand off to other networks at the earliest

possible point data that is not destined for termination on their own networks. A

principal goal of hot potato routing is to reduce the usage of network resources.

273

Opponents note that, during periods of congestion, data packets may be rerouted along

another path or dropped altogether and that packets may need to be re-sent when

transmission errors occur.

Opponents of net neutrality regulation argue that the TCP/IP protocol itself may

have differential effects for various content and applications.

274

For example, static Web

page content like text and photos and applications like e-mail generally are not sensitive

to latency. Thus, users typically can access them via the TCP/IP protocol without

270

See, e.g., Ryan, Tr. I at 238 (“IP networks do prioritize. They have from the beginning of time. The

prioritization that they had in the network at its inception was basically a first in line prioritization, first

in/first out. So it’s prioritization based on time, and time alone.”). See also McTaggart, supra note 117.

271

See supra Chapter I.A for a discussion of the TCP/IP protocol.

272

See generally supra Chapter I. See also Peha, Tr. I at 17 (“Actually, the [TCP/IP] protocol for 35 years

has allowed priority. But, for the most part, people haven’t used it. Or even implemented it.”).

273

See, e.g., McTaggart, supra note 117, at 10-12.

274

See, e.g., Yoo, Tr. II at 219. According to Yoo, “every protocol inherently favors some applications

over others. TCP/IP, first come, first served, very good at some things, worse at others. In a sense, there is

no neutral way to go here, by choosing one protocol over the other, you will actually be choosing winners

and losers.” Id.

noticeable problems, even during periods of congestion. Applications like streaming

video and videoconferencing, however, may be sensitive to latency and jitter.

275

Net

neutrality opponents argue, therefore, that while first-in-first-out and best-efforts

principles may sound neutral in the abstract, their practical effect may be to disfavor

certain latency- and jitter-sensitive content and applications because prioritization cannot

be used to deliver the continuous, steady stream of data that users expect even during

periods of congestion.

276

Network neutrality critics also note that content providers increasingly are using

local caching techniques to copy their content to multiple computer servers distributed

around the world, and argue that this practice effectively bypasses the first-in-first-out

and best-efforts characteristics of the TCP/IP protocol.

277

Critics further observe that

network operators have preferential partnerships with Internet “portal” sites to provide

users with greeting homepages when they log on, as well as customized and exclusive

content and applications.

278

Similarly, they note that portals, search engines, and other

content providers often give premium placement to advertisers based on their willingness

to pay.

279

In their view, these practices all constitute additional indicia of existing non-

neutrality.

2. Prioritization, Blockage, and Network Management Requirements

Network neutrality opponents frequently argue that operators should be allowed

actively to restrict or block data that they believe may be harmful to the performance of

275

See, e.g., Pepper, Tr. I at 85-86 (“The problem with non-discrimination is that it does not recognize that

treating different packets differently is necessary for the effective delivery of many services. As more real-

time interactive services dominate Internet traffic, it’s going to be more important to differentiate among

packets.”). See also McTaggart, supra note 117, at 12-14.

276

Some network neutrality proponents, such as Wu, have concluded that, “[a]s the universe of applications

has grown, the original conception of [Internet Protocol] neutrality has [become] dated; for IP was only

neutral among data applications. Internet networks tend to favor, as a class, applications insensitive to

latency (delay) or jitter (signal distortion).” Wu, supra note 227, at 149. Expanding on this point, some

network neutrality opponents, such as Yoo, have concluded that, because “TCP/IP routes packets

anonymously on a ‘first come, first served’ and ‘best efforts’ basis . . . it is poorly suited to applications that

are less tolerant of variations in throughput rates, such as streaming media and VoIP, and is biased against

network-based security features that protect e-commerce and ward off viruses and spam.” Christopher S.

Yoo, Beyond Network Neutrality, 19 H

ARV. J.L. & TECH. 1, 8 (2005). Therefore, in his view, “[c]ontrary to

what the nomenclature might suggest, network neutrality is anything but neutral.” Id.

277

See, e.g., McTaggart, supra note 117, at 6-7 (discussing Google’s distributed computing network).

278

See, e.g., id. at 4-5 (discussing network partnerships with portals such as Yahoo!, Microsoft MSN, and

Lycos). See also Waz, Tr. II at 162 (discussing the premium placement of portals on mobile phones).

279

See, e.g., McCormick, Tr. I at 273 (“[I]f any of us want to kind of envision what prioritization on the

Internet might look like, I mean, I think the clearest understanding of what we know prioritization would be

is looking at a Google search page.”).

their networks,

280

citing reports that a relatively small number of users can potentially

overwhelm network resources through the use of bandwidth-intensive applications, such

as peer-to-peer file-sharing and streaming video.

281

They warn that active network

management, prioritization, and other types of quality-of-service assurances are needed to

prevent the Internet, or its individual parts, from slowing down or crashing altogether in a

high-tech “tragedy of the commons.”

282

In their view, merely expanding network

capacity is expensive and may not be the most cost-effective method of network

management, and future content and applications may be even more resource-intensive

than applications like BitTorrent are today.

283

3. Efficiencies and Consumer Benefits from Prioritization

Network neutrality opponents argue that market transactions for prioritization and

other forms of quality of service can, in many cases, allocate scarce network resources in

280

Network neutrality proponents generally allow that some active management is necessary to maintain

network performance, but typically maintain that it should be limited. See, e.g., P

UBLIC KNOWLEDGE,

PRINCIPLES FOR AN OPEN BROADBAND FUTURE: A PUBLIC KNOWLEDGE WHITE PAPER (2005), available at

http://www.publicknowledge.org/pdf/open-broadband-future.pdf. According to this group, “[s]ome have

maintained that network operators must have the ability to restrict access to the network for legitimate law

enforcement purposes, or for network management. While these examples may be valid, this authority can

be easily abused and should not be broadly permitted.” Id. at 10.

281

See supra Chapter I.C.1.

282

See, e.g., McCormick, Tr. I at 243. According to McCormick, “[a] better Internet doesn’t simply come

by adding capacity. Like road networks, rail networks, electrical networks, and traditional telephone

networks, the advanced networks that comprise the Internet cannot function efficiently and cost-effectively

without management. No network has ever been built without regard to prioritization of traffic, peak loads,

and capacity management.” Id. Wireless network operators, in particular, argue that because their

networks may not have as much bandwidth as other wireline providers, they must be allowed to limit or

block certain content and applications like BitTorrent and to otherwise actively manage the use of their

networks’ resources. Network neutrality opponents state that any unintended consequences produced by

neutrality rules may have particularly acute consequences for such networks. See, e.g., Altschul, Tr. II at

51 (maintaining that applying network neutrality regulations to wireless broadband networks “would have

unique effects and they would be negative effects”).

283

See, e.g., Thorne, Tr. II at 34-39 (discussing the costs of deploying broadband networks). According to

Thorne:

When Verizon puts its fiber down a street, it costs us, in round numbers, $800 per home.

It costs us again, in round numbers, another $840 to connect the home that actually takes

the service. We spend the money to pass the home, but we don’t know whether the

customer is going to buy broadband service at all, or buy it from us.

Id. at 39. See also Schwartz, Tr. I at 255 (“Economically, it doesn’t make sense that the solution is always

to build more. That’s going to involve carrying a lot of excess capacity, which is going to be expensive.”);

T. Randolph Beard et al., Why ADCo? Why Now? An Economic Exploration into the Future of Industry

Structure for the “Last Mile” in Local Telecommunications Markets, 54 F

ED. COMM. L.J. 421, 430 (2002)

(estimating the cost of fiber-optic wireline deployment in a metropolitan area at approximately $3 million

per mile).

a manner more consistent with the actual priorities of end users.

284

Opponents further

suggest that prioritizing streaming telemedicine video, for example, ahead of e-mail or

network gaming transmissions to reduce latency and jitter would be socially beneficial.

285

Net neutrality opponents thus argue that network operators should be allowed to

prioritize the transmission of certain data or provide quality-of-service assurances for a

fee in the same way that consumers pay for priority mail service. Some observers note

that many other types of paid prioritization arrangements such as first-class airline

seating, congestion pricing for automobile traffic and public transportation, and premium

advertisement placements are commonplace and generally considered to be socially

beneficial.

286

In addition, they dispute the notion that non-prioritized data will be

relegated to an unacceptable, antiquated slow lane. Rather, they argue that non-

prioritized data traffic will continue to receive an acceptable level of basic service that

will continue to improve over time along with more general advances in data

transmission methods.

287

4. New Content and Applications and the Need for Network

“Intelligence”

Network neutrality opponents argue that new types of specialized services and

premium content require sophisticated, “intelligent” data-traffic management at both the

core and edges of the Internet.

288

Principal examples include VoIP, streaming video for

movies and telemedicine, large video download files, interactive network video games,

and customized business applications. In their view, “dumb” networks based on the

original TCP/IP protocol’s first-in-first-out and best-efforts standards are becoming

284

See, e.g., Schwartz, Tr. I at 255-56 (“[I]t makes sense to use the price system as a signal of which things

merit priority.”).

285

See, e.g., McCormick, Tr. I at 244 (“A communication about your health, for example, is clearly more

important than how quickly your kid can download a video featuring the antics of someone’s pet

hamster.”).

286

See, e.g., Sidak, Tr. I at 112 (“Obviously, we observe price discrimination in competitive markets all the

time.”). See also Farrell, Tr. I at 157 (“Price discrimination, as you have probably all heard many

economists say in forums like this, is not necessarily harmful. And that’s correct, given the other

alternatives available.”).

287

See, e.g., J. Gregory Sidak, A Consumer-Welfare Approach to Network Neutrality Regulation of the

Internet, 2 J.

COMPETITION L. & ECON. 349, 355 (2006) (“Rather than being forced down Lessig’s ‘digital

equivalent of a winding dirt road,’ these content providers would be relegated to something more like a

business-class seat on a flight to Paris.”).

288

See, e.g., Verizon Communications Inc., Public Comment 60, at 6-8. Verizon, for example, suggests

that “[n]ew Internet content and applications require innovative new broadband delivery methods” and that

networks need to be able to prioritize data “to manage bandwidth and control traffic on their network – for

example, to offer different levels of service for content and applications providers to reach their

customers.” Id. at 7-8.

increasingly outdated for certain content and applications.

289

Opponents argue that many

of these newer applications are sensitive to different levels of speed, latency, jitter,

symmetry, bursting, and capacity. For example, virtual teleconferencing generally

requires high speed, low latency, and symmetry, while some one-time video downloads

might require only high speed. By contrast, VoIP does not require significant bandwidth,

but is sensitive to latency and jitter. Neutrality critics argue, therefore, that network

intelligence will be increasingly necessary to provide the optimal transmission climate for

each of these new types of content and applications and that both content and applications

providers and other end users should be allowed to purchase services appropriate to their

particular needs.

5. Network Innovation and Competition

Network neutrality opponents contend that network operators should be allowed

to innovate freely and differentiate their networks as a form of competition that will lead

to enhanced service offerings for content and applications providers and other end users.

This perspective has been described as an argument in favor of “network diversity.”

290

Thus, opponents believe that network operators should be able to experiment with new

data-transmission methods and a variety of business plans to better serve the evolving

demands of end users. If such experiments turn out to be failures, network operators will

learn from their mistakes and improve their offerings or simply return to the status quo,

consistent with the normal dynamics of the market process.

291

In their view, a ban on

prioritization would effectively restrict new types of competition, hinder innovation,

potentially preclude price reductions for consumers, hamper efficiencies, and lock in one

kind of business model.

292

They warn that in the nascent and evolving market for

broadband services, mandating a single business plan is likely to lead to inefficient and

unintended outcomes.

293

They also assert that allowing content and applications

289

See, e.g., Adam Thierer, Are “Dumb Pipe” Mandates Smart Public Policy? Vertical Integration, Net

Neutrality, and the Network Layers Model, in NET NEUTRALITY OR NET NEUTERING: SHOULD BROADBAND

INTERNET SERVICES BE REGULATED?, supra note 42, at 73. See also Pepper, Tr. I at 81-83.

290

See, e.g., Yoo, supra note 276, at 9 (“In other words, standardization of TCP/IP would have the effect of

narrowing the dimensions of competition, forcing networks to compete solely on the basis of price and

network size.”).

291

See, e.g., Yoo, Tr. II at 220 (“If we have four players and one wants to experiment with a different

architecture, if they are wrong, they will get hammered and they will come back to the fold. If they are

right, it’s precisely the kind of innovation we should tolerate and encourage.”).

292

See, e.g., American Bar Association Section of Antitrust Law, Public Comment 2, at 8 (“Ultimately, we

believe that the competitive process will drive investment and innovation in the Internet. That investment

and innovation will inure to the benefit of all consumers. We do not think that imposing non-

discrimination statutes, regulations or policies will offer any offsetting benefits economically.”).

293

See, e.g., Pepper, Tr. I at 88 (“[One] concern is really whether net neutrality regulation designed to

prevent anti-competitive conduct could limit, or prohibit consumer welfare-enhancing network

functionality and management, as well as discourage innovation. In other words, regulation is not

costless.”).

providers to purchase quality-of-service assurances and prioritization may allow new

content and applications providers to counteract the competitive advantages typically

enjoyed by incumbent providers, such as the ability to pay for large server farms or third-

party data caching services.

294

6. Network Investment and Potential Consumer Benefits

Opponents argue that prohibiting network operators from charging different

prices for prioritized delivery and other types of specialized services and premium

content will make it more difficult to recoup the costs of infrastructure investments and,

thereby, reduce incentives for network investment generally.

295

They argue that both end

users and content and applications providers should be free to select any level of service

provided by network operators under market-negotiated terms.

296

Network neutrality opponents also stress that, although the Internet began as a

research and government communications network, its explosive growth since the mid-

1990s has been fueled mainly by private, risk-bearing investment.

297

They emphasize

that the individual, decentralized networks that make up the Internet mostly are owned

and operated by private companies and, generally speaking, are private property, even

though they may be subject to certain legal requirements like rights of way

permissions.

298

They point out that deploying and upgrading broadband networks can

entail billions of dollars in up-front, sunk costs.

299

Thus, they argue, any regulation that

reduces network operators’ ability to recoup their investments also effectively increases

294

Similarly, some network neutrality opponents argue that efforts by current leading content providers to

codify the status quo under the guise of neutrality rules are really nothing more than a veiled strategy to

commoditize data transmission and, thereby, preserve their own existing competitive advantages against

possible competitive threats based on new data-transmission techniques. See, e.g., Yoo, supra note 276, at

9 (“[T]he commodification of bandwidth would foreclose one avenue for mitigating the advantages enjoyed

by the largest players.”). See also George S. Ford et al., Network Neutrality and Industry Structure 1

(Phoenix Center Policy Paper No. 24, 2006) (“[P]olicymakers should avoid Network Neutrality mandates

that have the intent or effect of ‘commoditizing’ broadband access services since such a policy approach is

likely to deter facilities-based competition, reduce the expansion and deployment of advanced networks,

and increase prices.”).

295

See, e.g., Lenard, Tr. I at 181 (arguing there is a “striking lack of concern about the effect on incentives

to invest and innovate”).

296

See, e.g., Sidak, Tr. I at 107 (“Well, why do you need to have a federal law prohibiting one kind of

transaction, when you’re perfectly happy with the other?”).

297

See, e.g., Waz, Tr. II at 155-61. Waz states that “[a]ll that competitive investment is what makes it

possible for a Google and Yahoo! and eBay and Amazon and others to be here today . . . .” Id. at 158.

298

See, e.g., Bruce Owen & Gregory L. Rosston, Local Broadband Access: Primum Non Nocere or

Primum Processi? A Property Rights Approach, in N

ET NEUTRALITY OR NET NEUTERING: SHOULD

BROADBAND INTERNET SERVICES BE REGULATED?, supra note 42, at 163.

299

See, e.g., Thorne, Participant Presentation, at 1 (identifying Verizon Communications capital

expenditures of approximately $45 billion during 2004-06).

their risk profile to investors and, accordingly, would prompt capital markets to demand

an adjusted, higher rate of return. They suggest such an increase in the cost of capital, in

turn, would decrease the likelihood that projects underway could be completed on their

planned scale.

300

In addition to reducing incentives for network investment generally, opponents

argue that banning network operators from selling prioritized data delivery services to

content and applications providers will prevent networks from recouping their

investments from a broader base of customers.

301

In particular, they suggest that

networks should be allowed to experiment with a model in which content and

applications providers pay networks for prioritization and other premium services in the

same way that merchants pay for the placement of advertisements in newspapers and

other publications.

302

They suggest that such a business model might reduce prices for

some end users, much as advertising subsidizes the subscription prices of ad-supported

publications, thereby allowing marginal customers to afford broadband service.

303

They

further suggest that such increased end-user penetration would also increase the effective

demand for content and applications, generally, and thereby benefit their providers.

304

7. Economies of Scope from Vertical Integration and Bundling

Net neutrality opponents argue that vertical integration by network operators into

content and applications, along with related bundling practices, may produce economies

300

Sidak, supra note 287, at 357. In addition, some commentators characterize neutrality rules as being a

kind of regulatory taking of private property that can no longer be justified under a theory of natural

monopoly or other similar grounds. See, e.g., Thomas W. Hazlett, Neutering the Net, F

IN. TIMES, Mar. 20,

2006; Richard A. Epstein, What We Need is Regulatory Bed Rest, FIN. TIMES, Mar. 20, 2006. Both articles

are available at http://www.ft.com/cms/s/392ad708-b837-11da-bfc5-0000779e2340.html.

301

See, e.g., Yoo, Tr. II at 217 (“[W]e need to allow more flexibility on the server side. . . . Part of those

costs should also vary based on who, which servers, which content and applications providers need those

services.”). See also Sidak, supra note 287, at 367-68.

302

See, e.g., Yoo, Tr. II at 217 (“[W]e have learned in fact, these are two-sided markets. Basically,

upgrades to the network have to be paid for either by consumers or by the server content application

side.”). See also Schwartz, Tr. I at 258-59 (“[N]obody knows what the right pricing structure is. I don’t

claim to know it; nobody does. There is no presumption that the right structure is to recover all of the cost

of consumer broadband networks from consumers alone.”). Other examples of two-sided or, more

generally, multi-sided markets include credit cards (involving merchants and cardholders); dating services

(men and women); video game platforms (developers and players); and telephone networks (callers and

receivers). See generally Jean-Charles Rochet & Jean Tirole, Two-Sided Markets: A Progress Report

(Institut d’Économie Industrielle (IDEI), Toulouse, Working Paper No. 275, 2005), available at

http://idei.fr/doc/wp/2005/2sided_markets.pdf

303

See, e.g., Schwartz, Tr. I at 259 (“What economics predicts–and it’s independent of a monopoly or–it’s

independent of the degree of competition in broadband access–the prediction is if you allow them to charge

content providers, in their own interest they will now reduce prices to consumers, and therefore, encourage

penetration.”).

304

See, e.g., id.; Sidak, supra note 287, at 367-68; Sidak, Tr. I at 114-15.

of scope and price reductions. They point out that many areas of telecommunications are

increasingly converging. For example, both cable and traditional telecommunications

companies increasingly are offering “triple-” and “quadruple-play” bundles of high-speed

data, telephony, television, and wireless services.

305

In addition, they state that the

vertical integration of distribution with other types of media content is already

commonplace because consumers typically do not want distribution alone, but, instead,

want the particular content enabled by that distribution.

306

Some opponents also suggest

that the prospect of additional revenue streams derived from vertical integration and

bundling could promote additional competition in last-mile broadband services and

provide other benefits to end users.

307

8. Insufficient Evidence of Harm to Justify New Regulation

Network neutrality opponents argue that there is insufficient evidence of harm to

justify an entirely new ex ante regime, particularly when, in their view, competition in

broadband services is robust and intensifying due, in large part, to de-regulation. They

state that, apart from the Madison River case, which was quickly resolved by the FCC,

the harms projected by network neutrality proponents are merely hypothetical and,

therefore, do not merit new rules.

308

Also, they note that a number of network operators

have publicly pledged not to block or degrade end users’ use of their services.

309

They

305

See generally Marguerite Reardon, Cable Goes for the Quadruple Play, CNET NEWS.COM, Nov. 7, 2005,

http://news.com.com/2100-1034_3-5933340.html. See also generally Your Television is Ringing,

ECONOMIST, Oct. 14, 2006, at 3 (special survey of telecommunications convergence).

306

See, e.g., Lenard, Tr. I at 177 (“So what may be needed for a successful business model may be a

bundled product offering that is sufficiently attractive to attract enough consumers to become subscribers at

prices that are going to pay off the costs of these very large investments.”). See also Thomas L. Lenard &

David T. Scheffman, Distribution, Vertical Integration and the Net Neutrality Debate, in N

ET NEUTRALITY

NET NEUTERING: SHOULD BROADBAND INTERNET SERVICES BE REGULATED?, supra note 42, at 1, 13.

307

See, e.g., Rosston, Tr. I at 164-65. According to Rosston, “some of these vertical relationships that

people are concerned about that may increase the profits of a new entrant may be the thing that is

necessary, in order to get a new entrant, in order to compete.” Id. See also Thorne, Tr. II at 57-58.

Verizon, for example, suggests that it would be interested in partnering with hospitals to develop

specialized medical applications that could be delivered over its fiber-optic wireline networks to allow the

remote treatment of patients. Id. Likewise, some observers have pointed to Google’s involvement in

advertisement-supported municipal wireless Internet systems as an example of how vertical integration

may enhance last-mile competition and benefit consumers. See, e.g., Sidak, Tr. I at 108-09; Thorne, Tr. II

at 37; Wallsten, Tr. II at 59.

308

See, e.g., Wolf, Tr. II at 143-44 (“[J]ust as a doctor would not prescribe needless medication for a

growing adolescent on the possibility that some day that adolescent might develop a condition, so, too, we

think Federal regulators are prudent to refrain from prescribing conditions that may in fact stifle or injure

needed growth.”). See also Kahn, Tr. I at 185 (“I think the lesson of history is be very, very careful that

you don’t meddle with a process that is clearly characterized by Schumpeterian [dynamic] competition.”).

309

See, e.g., Thorne, Tr. II at 40 (“[Verizon has] made clear [that] when consumers buy Internet access

capacity from us, they should be able to reach any lawful website they want to get to with that capacity, and

we do not and will not block, degrade, or interfere with consumers’ access to any website.”); Net

Neutrality: Hearing Before the S. Comm. on Commerce, Sci., & Transp., 109th Cong. 21 (2006) (statement

of Kyle McSlarrow, President & CEO, National Cable & Telecommunications Association), available at

argue that operators do not have sufficient power over the distribution of content and

applications

310

and, in fact, would alienate their end-user customers if they tried to

engage in such practices.

311

Furthermore, they question whether it would even be cost-

effective for network operators to search for and block specific kinds of content and

applications in an ever-expanding Internet universe, given that an increasing number of

proxy servers and encryption techniques are available to end users to counter any such

blocking.

312

Similarly, some observers suggest that if such practices are detected, end

users can quickly publicize them and thereby “embarrass” the relevant network operator

engaging in such conduct.

313

Finally, network neutrality opponents suggest that the existing jurisdiction of the

antitrust agencies and the FCC is sufficient to deal with any prospective problems

resulting from the use of new data-transmission methods.

314

Generally, network

neutrality opponents suggest that any such problems should be handled on a case-by-case

basis – not through ex ante legislation or regulation.

315

They express concern that any

such regime might be manipulated in order to achieve strategic, anticompetitive outcomes

or be subject to other forms of rent-seeking behavior and unintended consequences.

http://commerce.senate.gov/public/_files/30115.pdf (“NCTA’s members have not, and will not, block the

ability of their high speed Internet service customers to access any lawful content, application, or services

available over the public Internet.”).

310

See, e.g., Thorne, Tr. II at 42 (“Does Verizon have the ability to prevent Google or eBay or these others

from reaching end users, when the most we could do is temporarily shut off a couple percent of the end

users they can see? . . . There is no single broadband provider that has that kind of power.”).

311

Opponents argue that a shift away from the America Online-type walled-garden model has taken place

and predict, therefore, that customers would vigorously protest any attempt to return to it after becoming

accustomed to generally unrestricted Internet access. See, e.g., Pepper, Tr. I at 136-37.

312

See, e.g., Thorne, Tr. II at 43 (“What we are selling is precisely the capacity to reach all lawful content

and applications. Broadband providers are motivated to maximize the content and applications available to

our customers because doing that maximizes the value of our network and the sales we can make.”). See

also generally Cat and Mouse, On the Web, E

CONOMIST, Dec. 2, 2006, at 3 (The Economist Technology

Quarterly survey) (discussing the ability of networks to block end users’ access to desired content and

applications and methods that end users may employ to circumvent such practices).

313

See, e.g., Lehr, Tr. I at 44 (“So, if there is a particular behavior that a carrier is doing, some sort of

quality of service differentiation that really has no justification in cost, and looks really high-handed, it’s

very common for this to get, you know, blogged in real time, and for this to embarrass the carrier so that – I

mean, the carriers and the operators – and force them to change their behavior.”). See also Weiser, Tr. II at

92 (making the same point).

314

See, e.g., Muris, Tr. II at 122 (“If problems of the sort imagined by the advocates of regulation emerge,

the appropriate law enforcement authorities have the jurisdiction and expertise necessary to address

them.”).

315

See, e.g., Schwartz, Tr. I at 254 (“[I]f foreclosure does rise to the level of a serious competitive problem,

the right response is to address it at the time, on a case-by-case basis–at least that’s my view.”).

IV. DISCRIMINATION, BLOCKAGE, AND VERTICAL INTEGRATION

As discussed in the preceding Chapter, proponents of network neutrality

regulation have raised a variety of concerns about the effects of vertical integration in

broadband markets, as broadband Internet access providers have begun to offer online

content and applications in addition to their primary access services. In particular,

proponents are concerned that providers may block or discriminate against unaffiliated

content and applications, to the benefit of affiliated offerings. Because such concerns

may stem from diverse vertical arrangements, this Chapter will construe vertical

“integration” broadly to include any arrangement under which a broadband Internet

access provider may claim income generated by content or applications, such as joint

ventures and exclusive dealing arrangements, as well as outright ownership of content or

applications.

This is a particularly complicated issue because vertical integration into content

and applications provision can create both incentives to engage in procompetitive,

socially beneficial behavior and incentives to engage in anticompetitive, socially harmful

behavior. Vertical integration generally need not be anticompetitive or otherwise

pernicious

316

and is often driven by efficiency considerations.

317

For example, such

integration may facilitate further network or content and applications development, and it

may spur development of network, content, and applications more optimally suited to

each other. Both price and non-price dimensions of broadband Internet service may thus

improve. As a result, the notion that vertical integration tends generally to be

anticompetitive has been widely rejected in antitrust law and economics for several

decades.

318

Many net neutrality proponents argue that their concerns about vertical integration

arise only when there is insufficient competition in the underlying Internet access market.

In that case, a vertically integrated last-mile access provider might exercise its market

power to block access to competing content or applications, degrade the transmission of

competing content or applications, or reduce investment in best-efforts Internet access

services in favor of priority services that carry the access provider’s own or affiliated

content or applications. Other proponents, however, have concerns that are independent

of the degree of market power the access provider enjoys in the access market itself.

These include concerns about the so-called terminating access monopoly problem and the

potential “balkanization” of the Internet.

316

See, e.g., Farrell, Tr. I at 154 (concerns about vertical integration in broadband markets are substantial

but contingent, sometimes highly uncertain, and “very hard to observe, and pin down”).

317

See, e.g., Yoo, Tr. II at 213-14 (citing research by FTC Bureau of Economics Director Michael Salinger

regarding efficiencies in vertical integration in the telecommunications industry).

318

See, e.g., Joseph Farrell & Philip Weiser, Modularity, Vertical Integration, and Open Access Policies:

Towards a Convergence of Antitrust and Regulation in the Internet Age, 17 HARV. J.L. & TECH. 85, 87

(2003).

This Chapter of the Report discusses concerns that net neutrality proponents have

raised about vertical integration in broadband Internet services. Section A discusses

problems that are most likely to arise when a provider enjoys substantial market power in

the provision of last-mile Internet access; Section B discusses certain problems that may

arise independent of the degree of market power attributed to an access provider; Section

C discusses various benefits that may be derived from increased vertical integration in

these markets; and Section D provides a brief summary of the competing arguments and

remaining uncertainties.

Because several types of alleged problems with vertical integration are tied in

some way to price or data discrimination, and because both definitions and applications

of “discrimination” have been contentious in the broadband Internet access discussion,

319

this Chapter first briefly clarifies that the economic meaning of discrimination is that of

differentiation and is not intended to have any negative connotation.

320

Thus, this Report

– in particular, this Chapter and Chapter V – does not assume that price discrimination or

any form of product or service differentiation is necessarily anticompetitive or anti-

consumer.

321

Even where demand conditions allow a seller to price above marginal cost,

price discrimination can provide a means of increasing overall consumer welfare by, for

example, providing access to goods or services for some consumers who otherwise would

be priced out of the market.

322

319

See, e.g., Ford, Tr. II at 239 (criticizing imprecise usage of terms like “discrimination” in the broadband

policy discussion). Cf. Farrell, Tr. I at 204-05 (noting disagreement in price discrimination terminology

within Workshop, but suggesting semantic dispute is unproductive); Lehr, Tr. I at 37-38 (trying to “move

away from the loaded term” of “discrimination”); William H. Page & John R. Woodbury, Paper Trail:

Working Papers and Recent Scholarship, T

HE ANTITRUST SOURCE, Apr. 2007, at 6, available at

http://www.abanet.org/antitrust/at-source/07/04/Apr07-PTrail4=27f.pdf (criticizing Workshop participant

Sidak’s discussion of price discrimination and Ramsey pricing).

320

That is, we generally attach no negative connotation to “discrimination.” Plainly, however, as

mentioned above and discussed throughout this Chapter and Chapter V of this Report, concerns have been

raised about particular potential forms of discrimination, such as blocking or degradation of competing

content and applications.

321

Classical price discrimination can, depending on its form, involve a combination of differential pricing

and product differentiation. See generally A

RTHUR C. PIGOU, THE ECONOMICS OF WELFARE (Transaction

Publishers 2002) (1920) (articulating, among other things, a general theory of price discrimination). The

idealized model discussed by Pigou involves monopoly pricing; there is no suggestion here that any

particular entities in the broadband Internet access market enjoy monopoly power or its approximation. Cf.

William J. Baumol & Daniel G. Swanson, The New Economy and Ubiquitous Competitive Price

Discrimination: Identifying Defensible Criteria of Market Power, 70 A

NTITRUST L.J. 661, 662 (2003) (“[I]t

is competition, rather than its absence, that in many cases serves to impose discriminatory pricing.”);

Alfred E. Kahn, Telecommunications, the Transition from Regulation to Antitrust, 5 J.

ON TELECOMM. &

HIGH TECH. L. 159, 177 (2006) (emphasizing “the difference between price discriminations, such as might

be taken to reflect inadequacies of competition, and differentiations on the basis of differences in costs,

such as would unequivocally be reflective of effective competition”).

322

That is, by producing and selling additional units priced between the highest-priced good or service and

the marginal-cost good or service. Hal Varian demonstrated generally that an increase in output is

necessary for profit-maximizing price discrimination to increase welfare. See Hal R. Varian, Price

Discrimination and Social Welfare, 75 A

M. ECON. REV. 870, 875 (1985); see also generally JEAN TIROLE,

Product differentiation in its simplest form can be a means of offering different

versions of a good to different consumers, according to their demands. A common

example is airline travel. Although all passengers receive the same basic product

(transport from one airport to another), airlines offer different fares based on different

levels of service during the flight (first class or coach) and flexibility in making

arrangements (leisure travel advance fares or last-minute business fares). By linking

price and product differentiation, a seller may be able to capture profits that would have

been available under unitary pricing and yet serve segments of the market that otherwise

would be excluded.

323

A. Last-mile Access Concerns Contingent on Market Power

Some net neutrality proponents have argued that vertically integrated broadband

providers possessing market power in the provision of last-mile access could leverage

that power in ways ultimately harmful to consumers. There are two major related

concerns. First, such providers could have incentives to discriminate against competing

content or applications providers.

324

Second, such providers could have incentives to

underinvest in the facilities used to provide common, best-efforts Internet access services.

Because techniques such as deep packet inspection can reveal source or content

information, there is some concern that vertically integrated providers with sufficient

incentives to discriminate against competing content could do so.

325

Such blocking could

take several forms. A broadband provider with an interest in content or applications

could block competing content or applications outright. Less extreme forms of

discrimination could impose degraded or otherwise inferior transmission on competing

THE THEORY OF INDUSTRIAL ORGANIZATION 137-39 (1988). Several Workshop participants applied this

general point to the broadband competition discussion. See, e.g., Sidak, Tr. I at 114-15. Several others

focused on the particular variant of so-called Ramsey price discrimination, observing, for example, that

Ramsey pricing is “the most efficient way to recover fixed costs.” See Yoo, Tr. II at 217; Lehr, Tr. I at 38.

In a seminal paper based on then-current models of monopolist price discrimination, Frank Ramsey

considered how a proportionate tax system might be structured to raise a given amount of revenue while

imposing a minimum decrease in utility. See F.P. Ramsey, A Contribution to the Theory of Taxation, 37

CON. J. 47, 47 (1927). The most general answer – that, “the taxes should be such as to diminish in the

same proportion the production of each commodity taxed” – provided a foundation not just for models of

taxation, but for, among others, utility rate structures and constrained price discrimination. See id.

Ramsey’s model mirrors monopolist price discrimination, but does so subject to a profit constraint.

323

See PIGOU, supra note 321, at 279-80.

324

See, e.g., Farrell, Tr. I at 156.

325

See Michael Geist, ISP Must Come Clean on Traffic Shaping, TORONTO STAR, Apr. 16, 2007, at D5,

available at http://www.thestar.com/sciencetech/article/203408

. See also supra Chapter I for a discussion

of deep packet inspection and other traffic-shaping technologies.

content. For example, such content might be denied access to prioritized routing,

326

relegated instead to best-efforts or otherwise inferior routing.

327

1. Discrimination against Competing Content and Applications

Some net neutrality proponents have argued that, if a broadband provider had a

financial stake in particular content or applications, it could have an incentive to block its

competitors’ content or applications.

328

In broad economic terms, one Workshop

participant identified the potential incentives to block competing content or applications

as the incentives to “resist substitutes”

329

for complementary goods in which the

integrated entity has a stake.

330

The incentive to block competitors could, for example, be to protect the primary

(broadband Internet access) market from future competition, especially from content or

applications providers that might themselves seek a presence in the access market;

331

the access provider could seek to facilitate price discrimination in the primary market.

332

326

In the alternative, the broadband provider could charge a very high price to competing content providers

to access priority routing.

327

See, e.g., CENTER FOR DIGITAL DEMOCRACY, LIFE IN THE SLOW LANE: A GUIDE TO THE UN-NEUTRAL

NET (2006), available at http://www.democraticmedia.org/issues/UNN.html.

328

See, e.g., G. Sohn, Tr. I at 116 (regarding “the possibility” that a provider would “favor certain

applications, content, and services”); cf. Libertelli, Tr. I at 76 (alleging actual applications discrimination or

blocking in wireless broadband 3G markets).

329

Farrell, Tr. I at 156. Farrell points out that if the broadband provider were allowed to charge competing

content providers a price for access equal to profits the broadband provider would lose by customers

buying the competing content instead of his own content, then there would be no incentive to block access.

However, this would lead to a very high price for the content – even monopoly levels. See also Rosston,

Tr. I at 163.

330

Some cable companies providing broadband service are currently integrated into IP telephony (in

addition to cable services, including video on demand). Conversely, some telephone companies providing

broadband service are currently integrated into cable-type video services (in addition to telephone services).

For example, AT&T through its affiliation with Akimbo Systems will branch out into other Internet content

as well. See Laurie Sullivan, AT&T Aims for Internet Television, T

ECHWEB TECH. NEWS, Apr. 18, 2006,

http://www.techweb.com/wire/networking/185303601

. IP telephony faces competition from third-party

providers such as Vonage, while video on demand services are now beginning to see competition from

third-party sources. See, e.g., Saul Hansell, Smaller Video Producers Seek Audiences on Net, N.Y.

TIMES,

Oct. 6, 2005, at C1, available at

http://www.nytimes.com/2005/10/06/technology/06video.html?ei=5090&en=042ceaad45ac8536&ex=1286

251200 (smaller producers trying to bypass traditional TV networks and sell directly to consumers over

Internet).

331

See Farrell & Weiser, supra note 318, at 109-10.

332

See id. at 107 (“Participating in, or dominating, the applications market can help a platform monopolist

to price discriminate; this objective may make even inefficient vertical leveraging profitable.”).

The assumptions underlying these concerns are controversial. First, to the extent

that such concerns about vertical integration depend on the vertically integrated entity

having significant market power in a relevant broadband Internet access market, there is

considerable disagreement as to whether such market power exists.

333

Even if an access

provider has sufficient market power to discriminate against competitors in

complementary content or applications markets, there remains the question of whether it

has sufficient incentive to do so. In an oft-cited article suggesting that there are

legitimate concerns about vertical integration in broadband markets, Farrell and Weiser

(both of whom participated in the Workshop) observed that an access provider,

depending on various contingencies, might or might not have sufficient incentives to

block competition in content or applications markets.

334

In that article, Farrell and

Weiser argue that “[p]rice discrimination need not in itself be inefficient or anticonsumer,

but the platform monopolist’s desire to price discriminate can . . . lead it to exclude

efficient competition or price competition in complementary products.”

335

They further

argue, however, that “platform monopolists” will balance the fact that the platform

business is more valuable when complements are supplied efficiently against the

possibility that “competition in the complement can sometimes threaten the primary

monopoly.”

336

Others argue that countervailing incentives are dominant and that discrimination

problems are merely hypothetical.

337

Specifically, they assert that a broadband access

provider’s chief incentive is to maximize the value of its core business – its network – to

present and potential customers.

338

Because that value depends centrally on the content

and applications to which the network provides access, several Workshop participants

maintained that providers would not have an adequate incentive “to limit their end users’

experience on the public internet.”

339

333

Chapter VI of this Report, infra, discusses more fully the present and (likely) future state of competition

in broadband access markets.

334

See Farrell & Weiser, supra note 318, at 100-01.

335

Id. at 108.

336

Id. at 109.

337

See, e.g., Lenard, Tr. I at 195. See also U.S. INTERNET INDUS. ASS’N, NETWORK NEUTRALITY AND

TIERED BROADBAND (2006), available at http://www.usiia.org/pubs/neutrality.doc.

338

See Lenard & Scheffman, supra note 306, at 18-19 (“[U]nder any market structure, the platform

provider has a strong incentive to maximize the value of the platform to consumers . . . . Broadband

providers benefit from having applications and content markets that maximize value to their customers.

Anything that detracts from user value will also reduce the demand (and hence the price that can be

charged) for the platform.”).

339

Thorne, Tr. II at 42-43; see also Sidak, Tr. I at 104 (“Network operators provide a complementary

service to Internet content. They do not have an interest in reducing the supply of a complement.”).

Thus, the degree to which a last-mile broadband access provider has a sufficient

incentive to discriminate against competing content and applications is an empirical

question. The broadband provider must weigh potential profits from additional revenue

from additional sales of its own content, against potential losses stemming from the

diminution of content or applications that consumers view as essential complements to

the access service. Certain net neutrality proponents have cited the Madison River matter

as evidence that the incentive to discriminate is, or could be, sufficient to prompt an ISP

to block a rival’s application.