Before the


Washington, D.C. 20554


In the Matter of

Inquiry Concerning the Deployment of
Advanced Telecommunications
Capability to All Americans in a Reasonable
and Timely Fashion, and Possible Steps
to Accelerate Such Deployment
Pursuant to Section 706 of the
Telecommunications Act of 1996


CC Docket 98-146






Jeffrey A. Eisenach
1301 K Street N.W.
Suite 550 E.
Washington, D.C. 20005
(202) 289-8928
(202) 289-6079 Facsimile

Charles A. Eldering
Adjunct Fellow
Telecom Partners Ltd.
900 Town Center
New Britain, PA 18901


 September 14, 1998


Table of Contents

I.          Introduction and summary

A.            Intermodal competition vs. regulation

B.            The slippery slope of regulating advanced services

C.            Containment of regulation: differentiating historically regulated services from advanced telecommunications services

II.         Defining "broadband," "advanced telecommunications services," and "reasonable and timely fashion"

A.            "Broadband" and "advanced telecommunications capability"

B.            Overview of technologies for "advanced telecommunications capability"

C.            "Reasonable and timely fashion"

III.       The demand for advanced telecommunications services

A.            Broadband Demand in Residential Markets

B.            Broadband Demand in Business Markets

C.            Supply and Demand

IV.       Technologies for providing advanced telecommunications services

A.            Cable television based ("cable modem") technologies

B.            Upgrading telephony networks via digital subscriber loop ("xDSL") technologies

C.            Other broadband technologies

V.        Removing barriers to competition and Deploying advanced telecommunications capabilities on a "reasonable and timely basis"

A.            The history of cable regulation gives a clear indication as to the effect of regulation on growth

B.            Deployment of broadband capable networks requires substantial investment which will only occur in a deregulated environment

C.            Removing barriers to broadband investment can be accomplished through a containment policy for regulation

D.            Universal service issues

VI.       Conclusions

A.            Intermodal competition, not regulation, is the key to encouraging the deployment of advanced telecommunications capability

B.            Harmonization of public policy towards advanced telecommunications capability networks can only be accomplished through deregulation

C.            Anti-trust vigilance, not prophylactic regulation, is the key to encouraging deployment of advanced telecommunications capability

APPENDIX A: A proposal for packet based definition of broadband services and containment of regulation

Overview of Packet Technology

Why Packet Technology is Different From Circuit Switched Technology

Deregulating Packet-Switched Technologies

Defining Packet Based Services

Interconnection and Unbundling in a Broadband Environment

Cost Allocation


APPENDIX B: Digital Broadband Working Group Briefing Attendee


I.          Introduction and summary

The Progress & Freedom Foundation ("PFF") is a private, non-profit, non-partisan idea center which was established in 1993 and studies the impact of the digital revolution and its implications for public policy. PFF makes these comments in an effort to educate the Commission and the general public as to what public policies will best encourage the deployment of broadband services, enable continued expansion of the Internet, and foster the growth of electronic commerce. Much of what is presented in these comments has been formed from contributions from a number of industry leaders in the telecommunications, networking and computer industries, based on meetings held in the last 18 months as part of its "Digital Broadband Working Group.1"

Through this activity and its other research, PFF has accumulated a substantial body of information, expertise and analysis relating to the issues addressed in this proceeding. This research suggests that competition, not regulation, is the appropriate model for the market for broadband digital communications services. While continued regulation of certain traditional services (i.e. Plain Old Telephony Service, "POTS") arguably is necessary, at least during a transitional period, the threat of regulatory spillover from the traditional telecommunications world into the digital broadband world represents a clear and present danger to investment in and deployment of digital broadband services. This analysis leads, in turn, to the suggestion that the Federal government adopt a "containment philosophy" for regulation, in which the narrowband infrastructure is deregulated according to the Telecommunications Act of 1996 ("the Act") and entrants are provided access to the basic network elements including unbundled loops, but in which digital broadband services are left wholly unregulated.

PFF offers herein its comments on CC Docket 98-146, the Notice of Inquiry Concerning the Deployment of Advanced Telecommunications Capability to All Americans in a Reasonable and Timely Fashion, and Possible Steps to Accelerate Such Deployment Pursuant to Section 706 of the Telecommunications Act of 1996 (hereafter referred to as the "NOI"). In the separate Memorandum Opinion/Order and Notice of Proposed Rulemaking, Petition of Bell Atlantic Corporation for Relief From Barriers to Deployment of Advanced Telecommunications Services et. Al. (CC Docket Nos. 98-11 et al., hereafter referred to as the "NRPM") the Commission proposes the establishment of separate affiliates by incumbent Local Exchange Carriers ("LECs") as a means of providing advanced telecommunications capabilities on a "deregulated" basis. PFF is not making separate comments on the Opinion/Order but makes references to the Opinion/Order and the concept of a separate subsidiary for data in the context of the Notice of Inquiry.

    1. Intermodal competition vs. regulation

The information gathered in PFF’s studies clearly indicates that intermodal competition – the competition which can occur between the wired local loop network operators (cable TV providers, incumbent Local Exchange Carriers, and competitive Local Exchange Carriers) and between the wired local loop network operators and their wireless counterparts (fixed wireless, mobile, and satellite) – has the potential to provide a rich and varied array of advanced telecommunications services to Americans at significant penetration rates. However, intermodal competition will be stifled by attempts to "level the playing field," which have the effect of raising entry barriers for network operators desiring to provide advanced services. By reducing the incentive to invest on the part of many participants, such efforts delay deployment of the infrastructure required to support advanced services at high penetration rates. The ability to serve all Americans, particularly those in rural areas, depends on the ability of network operators to obtain the appropriate economies of scale and scope in their deployments. Regulation of those deployments is likely to reduce their chances of obtaining either of these economic factors and could severely retard the deployment of advanced infrastructures.

PFF’s analysis of the technologies available for advanced telecommunications services indicates that there are subtle differences between competing technologies that will result in different service types, features and pricing, based on the architecture and business model each network operator/service provider adopts. The beauty of the Internet and the tremendous growth in electronic commerce that it is fostering is that it provides a common thread under which all of these architectures and services can operate. The previous models for regulation, based on telephone services and technology, broadcast services, or cable television, are simply inadequate, do not reflect the nature of advanced telecommunications capabilities and the Internet, and cannot possibly keep pace with the growth of the Internet and electronic commerce. Only by allowing competition to flourish can the Commission assure that advanced telecommunications services are deployed in a "reasonable and timely fashion" as required by Section 706 of the Act.

    1. The slippery slope of regulating advanced services

The Commission is tasked with implementing the Act and accordingly must ensure that all of the Act’s requirements with respect to unbundling, resale, interconnection and ILEC entry into long distance are met. Similarly, the Commission must, as specified in Section 706 of the Act,2 determine how best to encourage the deployment "on a reasonable and timely basis of advanced telecommunications capabilities to all Americans…." Although the Act specifies that one of the mechanisms for encouraging the deployment of advanced telecommunications capability to all Americans is regulatory forbearance, the Commission has determined in the NPRM3 that it does not have the statutory authority to forbear from Sections 251(c) and/or 271. This conclusion rests, in turn, on the Commission’s conclusion that "advanced services offered by incumbent LECs are either ‘telephone exchange service’ or ‘exchange service.’"4 This definition, which the Commission justifies based on a singular reference to the Federal-State Joint Board on Universal Service, threatens to drag all advanced telecommunications services into a complex system of regulation which is in no way prepared to deal with the Internet, electronic commerce, and the emerging dependence of the economy on the transport of digital information.

Having made such a definition, the Commission would seem to have no choice but to examine each and every advanced telecommunication service to determine how best to regulate that offering. Even if incumbent LECs adopt the option of creating separate subsidiaries, the role of the Commission in the regulation of advanced telecommunication services is bound to increase rather than decrease. Many will now ask, for example, why Title VI cable operators are not required to establish separate subsidiaries for their data services, or be forced to comply with interconnection, unbundling and resale requirements. Others will insist – and, indeed, the NPRM implies – that the rules for separate subsidiaries should eliminate any competitive advantages incumbent LECs have in the offering of advanced telecommunications services (including advantages associated with economic efficiencies). Such an approach raises the obvious question of why firms would enter markets in which they are prohibited from having a competitive advantage. It also ensures extensive, intrusive, and complex regulation (accompanied by lengthy regulatory delays and litigation) with respect to how separate subsidiaries are established and operated. Rather than limiting the Commission’s role ensuring that competitive LECs have adequate access to unbundled twisted wire pairs and the ability to collocate in a cost-effective manner, the NPRM puts the Commission on the road to regulating the deployment and marketing of all advanced telecommunications capabilities.

    1. Containment of regulation: differentiating historically regulated services from advanced telecommunications services

The alternative is to separate historically regulated services from advanced services. Specifically, as PFF’s research demonstrates, the Commission could establish a "containment policy" for regulation in which packet based services are distinguished from circuit switched services, with intrusive regulation limited to circuit switched services.

This delineation raises issues with respect to the erosion of universal service subsidization by a migration of services to unregulated packet based systems, and also could require in frequency unbundling of twisted wire pairs. While this approach has, like any other public policy, implementation details which must be addressed, it is certainly preferable to "harmonization through regulation" in which each and every aspect of advanced telecommunications services are regulated.

In summary, the PFF research indicates that a "containment philosophy" with respect to telecommunications regulation can be developed in which packet based services are left entirely unregulated. For local loop applications, a data rate requirement of 1.5 Mb/s in the network-subscriber direction and 384 kb/s in the subscriber-network direction would be applied to ensure that the telecommunications capability can support advanced services. The specifics of this proposal are discussed further in Appendix A.

PFF recognizes that such a proposal reflects only one means in which the Commission can promote the timely deployment of advanced telecommunications capability to all Americans, and urges the Commission to explore in depth any and all steps to prevent the unnecessary regulation of advanced services.

  1. Defining "broadband," "advanced telecommunications services," and "reasonable and timely fashion"

    1. "Broadband" and "advanced telecommunications capability"

In the NOI the Commission asks commentators to define what is meant by "advanced telecommunications services"5 and "broadband."6 As discussed at greater length in Appendix A, sound public policy for the deregulation of advanced services can be based on a definitions for these terms which encompass all emerging packet based services and which excludes all presently regulated circuit switched and presently regulated packet services (e.g. Frame Relay). Furthermore, consistent with the clear intent of Congress, as expressed in Sec. 706 and elsewhere in the Act, any such definition should be as expansive as practicable, permitting the Internet to continue to grow in a marketplace environment, not a regulated one.

    1. Overview of technologies for "advanced telecommunications capability"

As will be discussed in Section IV, there are a number of technologies capable of providing advanced telecommunications services. Each of these technologies has subtleties associated with the architecture which prevents a precise "apples to apples" comparison, but such differences are precisely why intermodal competition provides an alternative to regulation. If growth can occur free of regulatory barriers, consumers will have the ability to choose from a variety of services offered via a number of delivery mechanisms.

The telephone industry technology which is envisioned as being "broadband capable" is the family of digital subscriber loop technologies, referred to in general as "xDSL."7 The advantage of the xDSL technologies is that they utilize a significant portion of the existing copper telephone wires which already connect each home and business to the telephone network to provide broadband services. The ability to utilize the existing infrastructure means that advanced telecommunications services could be deployed without creating an entirely new plant. However, many existing twisted wire pair connections require conditioning in order to provide xDSL services, and if penetration rates and data rates to and from the homes are to become significant, fiber optic access systems (field based terminals connected to the central office or point of presence via a fiber optic cable) will need to be deployed in order to support xDSL networks.

Cable networks, which are broadband by definition, can support advanced telecommunications services via "cable modem" technology. Cable modems allow subscribers to use the existing cable system to receive data which is sent on downstream frequencies, and to transmit data from an individual subscriber on an upstream frequency. The advantage of cable modem technology is that it can be deployed on existing cable TV infrastructures. However, this is also the weakness of cable modem technology, cable TV’s architecture requires that bandwidth be shared amongst all of the subscribers on the node, leading to bandwidth limitations in the downstream, and "noise funneling" problems in the upstream which can make two-way cable systems unreliable.8

Comparing telephone and cable based technologies from an economic perspective indicates that while cable systems are inherently less costly than telephone systems, upgrading the broadcast based cable architecture for two-way services can result in per-subscriber costs which are higher than the costs for a "switched" telephone type architecture in which each subscriber has access to a specific drop cable.9 However, the crossover point in terms of cost – the point at which penetration rates are high enough to warrant a switched telephone company type architecture – depends strongly on the data rate and price of the service.

The two conclusions that can be drawn from the previous studies of these architectures are that i} to reach significant penetration rates all architectures require significant investments, ii} intermodal competition between telephone and cable, if it is permitted to exist, is extremely powerful based on the differences in technology.

Wireless technologies, including fixed terrestrial, mobile wireless, and satellite, all offer alternatives to wired technologies for advanced telecommunications services. Although these technologies probably are not the most economical means of providing advanced telecommunications services to the majority of residential subscribers due to limitations of spectrum and technical limitations associated with mobility, they provide the ability to serve specialized markets which account for a significant percentage of the market for advanced services. This market segment is likely to include businesses of all sizes, rural subscribers of broadband services, and subscribers willing to pay a premium for mobility. In addition, these technologies can serve as a "cap" on the exercise of market power that may exist in particular markets. For example, emerging fixed broadband wireless services based on LMDS licenses will be able to serve segments of the residential market, and LMDS operators will be able deploy services quickly, flexibly and at relatively low costs by erecting a small number of antennas to provide services to, for example, suburban subscribers (unlike mobile architectures which require an extensive network of antennas to provide continuous services).

Further discussion of the technologies for providing broadband services is presented in Section IV.

    1. "Reasonable and timely fashion"

The Commission has requested comments on what might define a "reasonable and timely fashion."10 As vague as such statutory terms may be, there is no doubt that the marketplace has already given a definition to this term: reasonable and timely is being defined by the growing dependence of the economy on electronic commerce and the fact that growth of electronic commerce will ultimately be slowed if broadband connections cannot be established to residences and small businesses. As will be discussed in Section III, there is evidence that the demand for data transport services is increasing at a rate which is many times that for traditional telecommunication services. No existing infrastructure is capable of meeting this demand. Given this unprecedented growth, the rapidly changing nature of the marketplace and technological environment and the potentially damaging impact that over-regulation of advanced telecommunications services could have on the economy, the definition of "reasonable and timely" must be determined by the marketplace and not the Commission or any other regulatory body. The past history of telecommunications regulation provides sound evidence that no regulatory infrastructure will be able to track the demand for advanced telecommunications services and adequately regulate these services.

  1. The demand for advanced telecommunications services

Predicting the demand for digital broadband services is a little bit like predicting the demand for sandbags in Missouri as the first raindrops begin to fall in Minnesota: About all you can be sure of is that you’ll need a lot of them. In the case of bandwidth, there is no doubt that the rain has begun to fall, and that a flood of significant magnitude is coming.

The Internet and the vast array of services and applications it enables will be the primary driver of demand for digital broadband for the foreseeable future. Set up nearly three decades ago for a small community of university and government researchers, the Internet has been widely accessible for nearly a decade. Only in the last few years, however, has demand for Internet access begun to rise very sharply.

The explosive growth in bandwidth demand associated with the Internet is the product of two underlying rates of growth, in (1) the number of computers connected to the Internet, and (2) the computing power of those machines.

The computers connected to the Internet consist of two types: servers or "hosts," and users or "clients." The numbers of host and clients are both growing at exponential rates. As of January 1998, the Internet had nearly 30 million host computers, over 20 times the number five years earlier.11 The number of Internet servers is expected to quintuple, to 100 million servers in the next two years.12 The implication of these growth rates is that the amount of material available to Internet users is growing explosively, making the Internet a much more attractive product to consumers.

It is not surprising, then, that the number of Internet users is also growing very rapidly. About 2.5 million U.S. users were connected to the Internet in 1990,13 compared with more than 56 million U.S. Internet users today.14 According to the U.S. Department of Commerce, "fewer than 40 million people around the world were connected to the Internet during 1996. By the end of 1997, more than 100 million people were using the Internet."15 Forecasts of future worldwide growth range from 130 million to 1 billion by the year 2000,16 and one industry analyst estimates that private user demand for Internet services is growing 85 percent a year.17

The impact of this explosive growth in the number of Internet users is multiplied by the rapid growth in the abilities of computers to create, store, transmit, and receive digital traffic. Since 1993, microprocessor speeds on the average $3,000 personal computer have increased by a factor of ten – from 33 MHz Intel 486s to 300 MHz (or greater) Pentium IIs;18 modem speeds have increased by a factor of twenty – from 2,400 bits per second to 56,000;19 and data storage capabilities have increased by a factor of thirty – from 240 megabyte hard disk drives to 8 gigabytes.20 By all indications, the power of digital machines will continue to increase at comparable rates indefinitely into the future. These advances enable information and data of all kinds to be processed more efficiently, in much greater quantities and at much faster speeds. The rise in processing power makes possible a wide variety of new electronic information services and computer-based communications applications.

The result of these two phenomenon is that the overall demand for digital bandwidth is growing at a staggering rate. According to one analyst, the total amount of digital traffic will grow from 32 quadrillion bits per day in 1996 to over 400 quadrillion bits per day in 2006, an average annual growth rate of nearly 70 percent.21 Bellcore predicts that demand for bandwidth will increase at least 20-fold in the next five to ten years.22 John Sidgmore, the Chairman of WorldCom, recently stated that demand for bandwidth on his company's backbone increased 1,000 percent in 1997 and that it is doubling every 100 days.23


Indeed, data traffic generated by computers and other electronic devices will soon exceed all other types of traffic on communications networks. Internet pioneer Vint Cerf recently predicted that by the year 2010, 90 percent of all communication could be data, and only 10 percent voice.24 Other studies have predicted that by 2001, voice will make up only 4 percent of total traffic on public and private telecommunications networks, whereas data will make up 95 percent and video and fax the remaining 1 percent.25

    1. Broadband Demand in Residential Markets

Most households already have access to large amounts of bandwidth, but most of it is analog and one-way. Virtually all households (97 percent) have telephone wires that supply more than enough bandwidth for voice communications. Most households (65 percent) also have cable wires that, on average, supply 50 channels of video programming.26 Almost all households that don't receive cable do receive over-the-air television, which delivers at least 12 channels of video – and will presently, through digital television technology, deliver much more.

Demand for analog bandwidth has increased only modestly in the last decade. While the number of lines in use has grown steadily, and faster than the rate of population growth, usage of voice lines has held roughly constant at 25 minutes per day per line over the past 15 years or more.27 Television viewing has likewise held constant at about 7 hours per day over the past decade, slightly higher for households with cable.

Residential demand for digital bandwidth, by contrast, is exploding. At least one-quarter of U.S. households now have a personal computer (PC) equipped with a modem that can be used to access data-based information services.28 Demand for modem-equipped PCs is continuing to expand at a rapid rate.29 The new generation of sub-$1,000 entry-level PCs have attracted many first-time buyers;30 prices for entry level computers had dropped to as low as $500 by the middle of 1998.31 Indeed, in May 1998, Gateway2000 introduced its "Your:)Ware" program, which allows consumers to purchase a fully-equipped PentiumII computer, along with access to the Internet, for a monthly fee of $49.95 per month.32

Computer software has also spurred residential demand for computer-based services. User-friendly software programs such as Netscape Navigator and Microsoft Internet Explorer have contributed greatly to the success of the World Wide Web. According to one study, the demand for some six million residential second telephone lines in 1995 – almost half of all second residential lines – can be attributed principally to on-line access.33

Most major consumer on-line service providers have embraced open standards and now provide access to the Internet, in addition to specialized services such as bulletin boards, chat rooms, live conferencing, software, home shopping and database services specializing in business, legal, medical and other technical information. On-line services still continue to attract increasing number of subscribers willing to pay for specialized services. It is estimated that the number of subscribers to consumer on-line services will continue to increase at a rate of 70 to 100 percent a year through 2000.34 America Online's subscriber base has increased from 303,000 in 1993 to over 11 million today.35

The available evidence suggests that demand for bandwidth in the residential market is highly elastic. A recent study by Robert Crandall and Chuck Jackson, for example, estimated that only four million consumers would be willing to pay $70 per month for an upgrade from 56.6 kb/s access to 1.1 Mb/s, but 20 million would pay $25.36 As Crandall and Jackson are careful to note, such estimates are based on a relative paucity of data. However, they are consistent with other data, such as a recent study by the Yankee Group. That study reported that nearly two-thirds of online households are interested in high-speed Internet access, but only 27 percent might be willing to pay $40 per month (about the same as the average cable bill) for unlimited high-speed access.

    1. Broadband Demand in Business Markets

Businesses have been using digital bandwidth much longer than residences. Electronic Data Interchange, a protocol for computer-to-computer transaction of billing, purchasing, invoicing, and other business functions, pre-dates the Internet.37 Businesses have also been big users of electronic funds transfers and electronic mail. In recent years, business usage of Internet services has of course grown very rapidly as well.

Until the rise of the Internet, many businesses were served by expensive private or leased facilities, obtained from value-added network (VAN) providers such as GE Information Services (GEIS) and IBM, as well as from local telephone companies, competitive access providers and long-distance carriers.

As in the residential market, business demand for digital bandwidth is being propelled by the personal computer. Nearly 70 percent of all computers are owned by businesses. Eighty percent of those machines are networked,39 and 44 percent are connected to the Internet.40 According to Forrester Research, by the year 2000, 363,000 businesses – a full 100 percent of the nation's large businesses, 50 percent of mid-sized businesses and 30 percent of small businesses – will be connected to the Internet.41 It is estimated that over 90 percent of Fortune 1000 companies have either established or plan to establish a corporate Intranet.42 From 1995 to 1996 the percentage of Fortune 500 companies with Web sites grew from 34 to 80 percent.43 Leased T1 lines are increasingly used to access the Internet and public switched networks. According to analyst Dataquest, the number of installed T1 lines will surge by about 23 percent per year during the next two years.44

While much of the bandwidth required by businesses in the past has been used for intra-corporate communications, future demand will largely be shaped by how much business and commerce is actually conducted over the Internet. Electronic commerce involves the sale and purchase of goods electronically and "any kind of communication or collaboration with customers, suppliers, and other business partners via computers linked to a network."45 A great deal of the commerce conducted on private or leased facilities can readily be shifted to the Internet. Many companies are creating methods to make electronic commerce easier and more secure.

A great deal of business-to-business commerce is transacted today via fax, phone, or through private electronic links. At present, businesses do about ten times as much on-line business with other businesses as they do with consumers.46 Essentially all business-to-business traffic can migrate to digital networks, and it is rapidly doing so. A recent survey by The Economist estimated that the value of business-to-business electronic commerce conducted over the Internet would increase to between $50 and $150 billion by 2000.47 Forrester Research has found that business-to-business commercial transactions over the Internet will hit $8 billion this year, and projects a $327 billion dollar annual market in 2002.

Video conferencing is another application that may generate significant new demand for digital bandwidth in business markets. According to a 1997 study, 76 percent of businesses are either currently using video conferencing technology, undergoing trials, or otherwise planning for its future use.48 The video conferencing industry has experienced an annual growth rate of 100 percent in the last year. Video conferencing technology is rapidly becoming easier to set-up and cheaper to purchase.

On-line retail commerce presents by far the most important prospect for the longer term, affecting both business and residential demand. Estimates of the volume of electronic commerce (measured in terms of the value of the underlying transactions) on the Internet by the year 2000 range from $10 billion to $230 billion.49 These, in turn, would generate between $3 billion and $30 billion in revenues for electronic commerce service facilitators. According to a survey by CommerceNet, an industry consortium, and Nielson, a media-research firm, 73 percent of Internet users are using the Web for commerce, and 10 million Web users in the U.S. and Canada have actually made on-line purchases.50 According to projections by International Data, a Massachusetts consultancy, by 2000 46 million consumers in America alone will be buying online.   

    1. Supply and Demand

Estimating demand for digital bandwidth is particularly difficult because of the mutually reinforcing impacts of growth in the constituent parts.51 Because networks become more valuable to each as the overall number of users (and quality of service) expands, innovation and growth in individual market segments leads to faster growth in others.52 Computers boost demand for Internet services; Internet services boost demand for computers. U.S. computer sales reached $16.7 billion in 1994; they are forecast to reach $73 billion in 2002.53 The number of publicly-traded U.S. computer companies has risen from a little over 400 at the beginning of 1994 to over 1,000 today. These companies have a total market value of nearly $1.4 trillion, a 400 percent increase from 1994. In 1996 and 1997, 391 Internet and computer industry companies held initial public offerings, generating $19.5 billion54 in market capitalizations. The high-tech sector is now the largest industry in the U.S., with sales of $866 billion, 57 percent higher than in 1990.55 High-tech employment is also booming, employing 4.3 million people in 1996, with over a quarter of a million added between 1995 and 1996.56 Growth in each one of these sectors helps propel growth and demand in all the others.

Increases in the actual supply of bandwidth also propel broadband demand.57 Competitive access providers have deployed high-capacity fiber optic networks in most major cities. CAP revenues from private line and data services have risen from $216 million in 1993 to $695 million in 1996.58 The number of CAPs nearly doubled, from 41 to 78, in that same period.59 Four companies – IXC Communications, Qwest Communications, William Communications Group, and Level 3 Communications – are currently spending $5.9 billion to build long-haul fiber networks extending 63,000 miles, to be completed by the end of 1999.60 Several companies and consortia are also investing billions to deploy large constellations of satellites to provide broadband access. Teledesic and Celestri will invest $9 billion and $13 billion, respectively, to develop world-wide broadband satellite backbone networks. Traditional local exchange and long-distance carriers have been merging at a rapid pace, in part to achieve the efficiencies necessary to deploy broadband facilities more expeditiously.61

Increases in computing power and bandwidth further impel demand for Internet and other information services. Internet Service Providers have emerged as an entirely new class of telecommunications carriers who are just beginning to generate tremendous revenues.62 There are now over 4,300 Internet Service Providers (ISPs) in the United States, more than triple the number at the beginning of 1996.63 One estimate places the size of the ISP market for access services at nearly $6.5 billion in 1997, up from just $880 million in 1995.64 The market is expected to grow to $50 billion by 2000.65 Online Service Providers (OSPs) like America Online, CompuServe, and the Microsoft Network have likewise seen their growth accelerate in the 1990s as consumer demand for online services skyrocketed. The revenues of the largest OSP, AOL, have increased from $38 million in 1993 to over $1.4 billion today.66

It is impossible to estimate with precision the demand curves for broadband services. Such estimates would require the availability of time-series or cross-sectional data which simply do not exist. And, in any case, such estimates would rapidly be overcome by new technological developments, which fundamentally affect the nature of services available through the Internet and hence the underlying consumer demand curves.

However, it is quite clear that the demand for more capable and capacious data delivery services is growing rapidly. Consumers report their desire to purchase broadband services, and when offered the opportunity, they are increasingly doing so. Businesses increasingly depend on broadband services to offer an ever-widening array of on-line products, market other products, and exchange data needed for business-to-business electronic commerce.

  1. Technologies for providing advanced telecommunications services

Broadband services to homes and business have heretofore been provided by two,

telephone-based technologies; leased lines (e.g. T-1 and fractional T-1 lines) and Integrated

Services Digital Network (ISDN) services. T-1 lines, typically used by businesses, universities and other large organizations, provide 1.544 Mb/s bi-directional connections, but at costs of up to $2,000 per month.

ISDN technology uses digital modems on both ends of an ordinary copper wire that utilize much more of copper’s potential bandwidth and boost capacity by a factor of 4 – to a maximum of 128 kb/s.67 ISDN is now available to 85 percent of residential access lines and 100 percent of business lines.68 Although upgrading the network for ISDN is quite expensive,69 most of the costs have already been borne.

ISDN is, at best, only an interim solution to increasing bandwidth demand, for two reasons. First, ISDN does not offer enough bandwidth for certain applications, like delivering video.70 Second, ISDN relies entirely on the circuit switches designed for voice traffic, and these switches will become increasingly congested with data traffic as a result of increased Internet usage.71 Approximately one million subscribers, nearly 90 percent of which are businesses, are currently using ISDN.72 The absence of standards, installation problems and high prices73 have slowed deployment of ISDN and limited initial customer acceptance. By 1998, the initial customer set-up cost for ISDN had decreased considerably, and monthly rates for ISDN service were falling,74 but these improvements appear to have come too late for a technology that is now seen as offering only a limited increase in functionality.

Two new technologies are fundamentally changing the market for digital broadband services: Cable television based ("cable modem") technologies and telephone based xDSL technologies. Other technologies, including terrestrial and satellite wireless, and electric utility based services, have the potential to serve particular market segments and/or to compete eventually for the entire marketplace.

    1. Cable television based ("cable modem") technologies

Virtually all U.S. households have access to cable television.75 Most cable networks were built for at most 54 video channels on coaxial cables with 350 MHz of usable bandwidth.76 Between 80 and 90 percent of all cable networks do not allocate any significant amounts of bandwidth to upstream data; the closest these networks generally come is a sliver of bandwidth for pay-per-view requests.77

During the past few years, cable operators have invested heavily to upgrade their networks to hybrid fiber-coax (HFC), which greatly expands bandwidth and provides the ability to activate a "return path" for two-way communications.78 On HFC networks, signals are sent from the cable head-end via fiber-optic cable to neighborhood "drop sites" or "nodes" where it is shunted to coaxial cables and makes its way to subscribers’ premises. According to industry analysts, between one-third and one-half of all cable systems have undergone or are now undergoing HFC upgrades; by the year 2000, several analysts predict that 60 percent of cable plant will be upgraded to HFC.79 Relative to ADSL upgrades and deploying fiber to the curb, HFC is relatively cheap, although at high penetration rates the costs for a fully interactive HFC system can exceed those for a switched digital infrastructure.80

A HFC architecture is required to support the most promising access technology on cable networks, the cable modem.81 Cable modems transmit data over cable networks at high speeds – up to 30 Mb/s.82 Cable modem technology avoids many of the problems associated with ADSL. Whereas ADSL is limited by distance and by the condition of the copper lines, virtually the entire cable infrastructure is capable of accommodating the bandwidth needed for cable modem services. On the other hand, some observers have expressed concerns about the scalability of cable-based technology. Even with HFC infrastructure in place, cable networks use a tree-and-branch architecture, with all users ultimately sharing a common transmission path.83 As a result, each new user limits the bandwidth available for other users, possibly resulting in congestion problems. Many analysts, however, believe that scalability problems can be resolved.84

Cable modems have passed the trial stage and several companies have begun deploying data services through offerings such as @Home and Roadrunner.85 Cable companies that are currently offering high-speed Internet access charge residential subscribers between $30 to $50 a month for service, including ISP service.86

Initial reports suggest that consumer acceptance of cable modem technologies is meeting or exceeding expectations. One recent report places the total number of cable modem subscribers at 450,000 as of mid-1998, with providers reporting growth rates in excess of 100 percent annually.87 Roadrunner and Media One Express recently announced much higher than expected consumer acceptance in their Portland, Maine rollout, claiming that subscribership in that market had reached seven percent of all eligible customers and that the number of subscribers rivaled that of America On-Line.88 One study projects that cable modem services will have 1.6 million users by the year 2000.89 Other studies project over three million cable modem users by 2002.90

    1. Upgrading telephony networks via digital subscriber loop ("xDSL") technologies

The local telephone network currently provides ubiquitous voice service and mass-market access to the Internet. Although the telephone network was not designed to carry digital data, it is rapidly being upgraded to do so more quickly and efficiently.

Telephone companies are converting the existing analog copper loop to support high-speed digital transmissions. For data, the local loop is more than just the twisted-pair copper wire that runs between a subscriber’s premises and the telephone company’s central office; it also includes two modems at each end of the wire, one of which is controlled by the subscriber. Today, almost all Internet users use analog voice grade modems to connect their computers to the telephone network.91

The maximum speed of analog modems is lower than the maximum capacity that the copper wire will allow. Analog signals – whether from a telephone or a modem – are limited to using only certain frequencies within the copper wire. These usable frequencies represent 4 kHz of bandwidth, which in digital terms is equivalent to 64 kb/s. Most analog modems today operate at 28.8 or 33.6 kb/s.92 The newest generation of analog modems is capable of 56 kb/s, the practical limit for analog modems.93 But the copper wire used in local telephone networks has a theoretical bandwidth of 1 MHz.94 The local loop can be upgraded to support much of the near-term demand for bandwidth without having to replace the embedded base of copper wire. The key to this upgrade is replacing the analog modems on the ends of the loop with digital ones.

The major telephone companies are now upgrading their networks with digital subscriber line (xDSL) technology.95 Like ISDN, xDSL expands the capacity of the existing copper loop by replacing analog modems with digital adapters that use much more of copper’s potential bandwidth and by dividing each existing copper wire into multiple channels. Unlike ISDN, however, xDSL enables the channels carrying data traffic to be diverted around the central office switch and onto packet-based networks. This frees up the circuit switches for voice traffic, and means that telephone companies can handle the increased demand for traffic by purchasing relatively inexpensive and efficient packet-routing equipment, instead of much more expensive circuit switched central office equipment.96    

There are many varieties of xDSL technology, each of which provides different transmission rates to end users. Asymmetric DSL (ADSL) is at the low end of the range in terms of speed, but is also technologically closest to wide-scale deployment. As its name implies, ADSL provides asymmetric capacity, with very high bandwidth downstream and lower bandwidth upstream. The high-speed ADSL channel ranges from 1.5 to 8.44 Mb/s, and the upstream channel ranges from 16 to 640 kb/s. Other xDSL technologies under development promise bandwidths up to 52 Mb/s.97

Telephone companies have recently begun deploying ADSL,98 although most other DSL technologies are still in the trial stage. Other xDSL technologies are expected to attract considerably fewer subscribers.99 There have however, been recent problems with choosing a universal standard for ADSL, a factor which may slow its deployment.100 A bigger concern with xDSL is that it has severe distance limitations: it cannot be used on very long loops (18,000 feet or more), which represent approximately 20 percent of all loops.101 Moreover, the available bandwidth from xDSL begins to fall-off considerably at distances greater than 9,000 feet from the central office.102 Furthermore, even many short copper wires are deployed in ways that would require significant reconditioning before they could withstand the demands of ADSL.103 ADSL modems are currently expensive – estimates range from $1,000 to over $3,000104 – but these prices can be expected to drop fairly quickly once the technology matures.105

Recently, an industrial consortium of leading PC industry, networking and telecommunications companies joined together to create a subset of the existing ADSL standard to encourage the rapid deployment of ADSL technology. The new standard being developed, titled "Universal ADSL" or "UADSL" would permit a rapid deployment of ADSL supporting rates of up to 1.5 Mb/s in the downstream and 384 kb/s in the upstream. One of the claimed advantages of UADSL is that it does not require the installation of a "splitter" at the residence, and thus could offer network operators the ability to upgrade for high-speed data simply by installing a UADSL modem at the central office/point of presence. Computer manufactures would supply the residential UADSL equipment as a plug and play card in a PC, or as part of a modem shipped with the PC, so that consumers only need to connect their UADSL modem to an UADSL enabled line to obtain service. The Universal ADSL Working Group ("UAWG") recently announced the completion of an initial specification and its intention to continue to pursue adoption of a standard by the International Telecommunications Union ("ITU") by October 1998.106

It appears to be too early to predict the degree of consumer acceptance of ADSL offerings. In Phoenix, US West launched its DSL service in October 1997 and within a week 1,700 customers requested service, twice what the company expected.107 Some analysts predict that ADSL will have as many as two million residential and small business users by the turn of the century.108 On the other hand, ADSL currently lags far behind cable modems in terms of deployment, with approximately 25,000 users, compared with 450,000 cable modem subscribers.109

    1. Other broadband technologies

      1. Fixed wireless (MMDS, LMDS) technologies

Fixed wireless technologies are expected to play an increasingly important role in the local loop, particularly for data services.110 WinStar, a rapidly growing provider of local services, recently launched its fixed-wireless service, Wireless Fiber, using the 38 GHz band of spectrum. The service provides local, long distance and Internet access services over a network of roof-mounted access antennas that terminate on a landline switch.111 Other competitors, including AT&T, are expected to offer fixed wireless services in the near future.112

LMDS. Local Multipoint Distribution Service (LMDS) is a digital microwave broadband service that operates at very high frequencies with capacious bandwidth.113 LMDS licensees may control up to 1.3 GHz of wireless spectrum, which can be used to carry data at speeds in excess of 1 Gbps.114 LMDS can provide customers with multichannel video programming, telephony, video communications, and two-way data services.115 LMDS has true duplex capability, unlike most competing broadcast technologies. This allows for deployment of effective two-way voice and data services, and paves the way for future interactive video services to be added, with little or no additional infrastructure cost.116 LMDS uses a small-cell configuration, and is able to polarize and reuse spectrum very effectively over a small area. CellularVision currently operates the only commercial LMDS site in the United States.117 Although the deployment schedule for LMDS is highly uncertain, many major industry players – including AT&T, MCI, Sprint, GTE Corp. and Ameritech – have expressed interest in deploying the technology.118

LMDS requires a clear line of sight and is subject to fading due both to rain and foliage, thus has technical limitations which could be a substantial impediment to high penetration rate deployments. However, LMDS may be a crucial technology for broadband services provided to small businesses and high-end residential customers if competition in the wired arena is slow in coming.

MMDS. Another form of fixed wireless, Multipoint Multichannel Distribution Service (MMDS), also shows promise as a broadband distribution system. MMDS is now used for analog wireless cable-TV services, but is being upgraded to provide residential and business Internet access at speeds up to 27 Mb/s.119 Operators have been experimenting with both one-way service (using spectrum for the downstream link, and telephone lines for the uplink) and two-way service (using spectrum in both directions).120 A key limitation of MMDS, like LMDS, is that it requires a clear line of sight between the transmitter and receiving antenna, which makes the service unavailable in many geographic areas and in certain weather conditions.

      1. Land based mobile wireless technologies

Mobile wireless data services have been built based on existing cellular telephone platforms, beginning with receive-only paging services, and now extending to two-way wide area packet based networks. Cellular digital packet data ("CDPD") technology has been developed to transmit data over the cellular network in the idle times between voice transmissions, and supports data rates of 19.2 kb/s. In the US the ARDIS cellular data network has been operating for more than 10 years and has over 40,000 subscribers which send and receive data at 19.2 kb/s. RAM mobile data operates a wireless network that provides service to more than 92 percent of the US urban business population.121

The Ricochet wireless modem system and service offered by Metricom has proven to be a great success in the limited markets presently served which include Washington D.C., the San Francisco Bay Area, and Seattle, Washington, along with a number of major airports.122 Ricochet subscribers are provided with data rates of 28.8 kb/s – 40 kb/s. The Ricochet system uses unlicensed spectrum in the 902-928 MHz spectrum and closely spaced transmitters to provide coverage.

Wireless Local Area Networks (LANs) which can support data rates of 1 Mb/s or above are typically deployed in situations where limited roaming (within a department, building or campus) is sufficient. These networks use unlicensed spectrum and are based on spread-spectrum technology.

Although it is unlikely that the use of mobile wireless technologies will support the deployment of advanced telecommunications capabilities to "all Americans," the ability to provide mobile connectivity to a significant number of business subscribers plays an important role in the deployment of advanced services. Advances in compression and network technologies will ultimately permit these systems to transport video at 384 kb/s or above.

      1. Satellite technologies

Since 1994, DBS has been offering broadband video services nationwide at prices competitive with cable operators. More recently, DBS operators have introduced two-way services that use the telephone line as a return-path.123 Hughes offers its DirecDuo service, which provides both high-speed (400 kb/s) Internet access and DBS video programming through the same satellite dish. There are currently 4.4 million DBS homes,124 although it is uncertain how many of these also subscribe to DBS data services.

Although the two-way capabilities of DBS are limited, there are currently numerous efforts underway to deploy broadband satellite networks with much larger upstream bandwidth. The largest of these projects are Teledesic125 and Celestri,126 which will rely on low-earth orbit (LEO) satellites to build international satellite backbone networks.127 In addition, there are at least five proposals to deploy narrow-to-mid-band satellite networks that would provide personal communications services to end-users, as opposed to backbone services to businesses and ISPs. The largest and most advanced of the narrow-to-mid-band LEO projects is Iridium.128 Iridium, which initially planned to offer services by the end of 1998, but has recently experienced delays and will distribute two thousand handsets to individual, corporate and government customers in late 1998 to test services.129


  1. removing barriers to competition and Deploying advanced telecommunications capabilities on a "reasonable and timely basis".

In this NOI the Commission raises the issue of how to remove barriers to competition for the deployment of advanced telecommunications capabilities. A related question is that of how market forces will propel the development of broadband facilities, and how this can happen in a regulated vs. unregulated environment. The least regulated segments of the telecommunications industry – long-distance, wireless and information services – already have attracted large investments in broadband facilities. An enormous amount of additional investment is still needed, however, to supply broadband local transport to all Americans. But the carriers that supply most local transport today – local exchange carriers and cable operators – are the most heavily regulated entities in the telecommunications industry. The economic incentives of these carriers are in large part still determined by regulatory policy, not market forces. Numerous regulatory impediments greatly reduce the incentives that incumbent local exchange carriers and cable operators have to deploy broadband facilities. In many cases, these same regulatory restrictions also deter would-be competitors from deploying local broadband access facilities.

    1. The history of cable regulation gives a clear indication as to the effect of regulation on growth

The history of cable television regulation, since its invention in 1948 and development as a widespread commercial product in the 1970s, is a tale of massive inconsistency. Regulators have had a difficult time categorizing cable television, which has characteristics of both common carriage (i.e. telephony) and broadcasting (i.e. television).130 The history of cable regulation has led some to conclude that "from the 1960s to the 1990s the authorities took the worst of both regulatory models and piled them high on top of cable, to the point where they almost killed it."131

Cable prices, which previously were regulated by local franchising authorities, were deregulated by act of Congress in 1984. The 1984 Act established that cable rates would be deregulated wherever there was "effective competition" for the provision of television. The Commission determined that the effective competition test would be met by the presence of three broadcast television signals – a test which was easily met and which resulted in the de facto deregulation of cable television prices virtually everywhere.132

By 1992, however, rising cable TV prices led members of Congress to re-impose price controls. The Cable Act of 1992 required that the FCC ensure that cable rates were "reasonable," and the Commission reacted by imposing first a price freeze, then a 10 percent across the board rollback, and then a further seven percent across the board rollback. Not surprisingly, these government-imposed price cuts substantially reduced the market valuation of cable television companies, hindering their ability to raise capital for network upgrades and other needed modernization. The price controls did not, however, benefit consumers, who suffered from reduced service offerings, lower quality of service and an overall price-quality combination they found less attractive than before re-regulation.133

By 1996, Congress appeared to have recognized its mistake. Motivated by complaints from a number of sources, and by the presence of increasing competition to cable from the satellite broadcasting industry, it used the Act as a vehicle to effectively end price regulation of cable television effective March 31, 1999.

Deregulation of prices, per se, does not by any means end Federal or local regulation of the cable television business. Cable remains subject to local franchising fees (up to five percent of revenues), "must-carry" rules (which force cable operators to carry programming for which they receive no compensation), restrictions on cross-ownership of stations and on ownership of programming – plus an array of local franchising requirements.

Perhaps of even greater concern, recent statements from leading members of Congress have suggested that cable deregulation is not yet a "done deal" – that rate "re-regulation" may once again lie in cable's future. Such an action, which undoubtedly would be aimed at once again freezing or rolling back cable television prices, could deliver a body blow to the ability of cable television operators to undertake the kinds of investments in broadband services required to extend advanced telecommunications capabilities to all Americans..

Further, the status of cable under the Act is anything but secure. As cable companies increasingly compete in the markets for local telephone service and Internet access, there is a very real possibility that the Commission will feel obliged to bring cable – including the data services it increasingly provides – under a new, more intrusive regulatory regime. The establishment of separate subsidiary requirements for incumbent LECs who wish to provide advanced services on an unregulated basis will no doubt led to petitions intended to bring cable data services into that regulatory regime.134

    1. Deployment of broadband capable networks requires substantial investment which will only occur in a deregulated environment

Cable companies and incumbent LECs are subject to a wide-ranging and highly intrusive regulatory apparatus that affects virtually every aspect of their operations. The regulations that have the greatest effect on their ability and incentives to deploy broadband capacity fall into two main categories; rate and entry regulation, and resale and unbundling requirements. In addition, both sets of firms are subject to a variety of other impediments, such as "must carry" rules for cable, and the Inter-LATA restriction for telephone companies.

Rate and Entry Regulation: Incumbent local exchange carriers and cable operators are subject to a variety of restrictions regarding the types of services they may offer over their networks and the rates, terms and conditions at which these services must be provided. Section 214 of the 1934 Communications Act prevents phone companies from constructing new facilities or discontinuing existing service without advance permission.135 Section 201 of the Act enables the Commission to regulate what sort of devices can be connected to the telephone network, and thereby what kind of add-on services can be offered over telephone lines.136 Local telephone companies must file tariffs and cost support information for each of their common carrier offerings with either state or federal regulators, or both. Any new service that a local telephone company seeks to offer must be added to its tariff.

Title VI of the Communications Act extends franchise137 and rate regulations138 to the provision of "cable services."139 As discussed above, cable MSOs remain subject to rate regulation for their "basic" services through March 31, 1999. The current regulations resulted in dramatic reductions in the ability of cable companies to attract capital, without any commensurate benefits for consumers. Indeed, most observers credit the failure of the proposed Bell Atlantic/TCI merger in 1994 to the imposition of the second round of mandated price rollbacks by the FCC in February of that year.140 Delay or cancellation of the rate deregulation scheduled for next year may result in another round of canceled mergers and delayed entry by cable operators into the broadband marketplace, insuring further delay in the deployment of advanced services by cable operators.

Existing rate and entry regulations reflect the conclusion that regulators can allocate goods and resources and set prices more efficiently than market forces. In some instances (declining in number as new technology and the globalization of most markets spur competition) this may be true. Unregulated natural monopolies, for example, may price too high, and produce too little. But the beneficial effects of regulation (such as they are) can only be realized if regulators perform their functions efficiently, on schedule, on the basis of up-to-date information, and to protect the public, not industry incumbents. The Commission must by definition take a long and careful look at each problem that it confronts: few would argue that a fast but capricious government agency would serve the public well. However, the result is that regulatory decisions have an extremely long time factor built in: as an example, routine licensing decisions take far longer than they should. The resulting delay often solidifies the economic status quo, protects incumbents against would-be competitors, and deprives the public of new services at lower prices.141

Meanwhile, ratemaking proceedings provide a ready forum for critics that continuously try to tug incumbents in two diametrically opposed directions. One group of critics claims that broadband investment is not responsive to consumer demand, and thus a form of wasteful goldplating that harms ratepayers. Another group claims that broadband investment is behind consumer demand, and blames incumbents for being insufficiently responsive to subscribers' needs. In truth, there is only one way to determine the market's true need for broadband investment: let market forces, not regulators or special interest groups, drive investment decisions.

Carriers' decisions to offer new services are also affected by the inherent uncertainty involved in whether authorizations and tariff approvals are needed in the first instance. A carrier seeking to provide a new broadband service must determine what kinds of regulation apply to its proposed service. Will the new service be regulated or unregulated? Basic or enhanced? Intrastate or interstate? Local or long-distance? IntraLATA or interLATA? Is a separate affiliate required? In many instances, however, a carrier will not know how a proposed service will be regulated and whether the carrier can offer it all. Many broadband services simply do not conform to established regulatory categories. On the Internet, for example, it is almost impossible to track where calls originate, what communications paths are traveled, or where information is ultimately delivered or consumed. Yet every telephone service must be categorized as intra- or interstate and intra- or interLATA. The Commission has declined to adopt objective criteria to determine whether Internet services related services are interLATA in nature, but has instead stated that it will make this determination on a case-by-case basis.142

The need to squeeze new services into regulatory categories to which they simply do not fit has a dilatory effect on the ability of carriers to quickly introduce products to market. Regulatory uncertainty is further compounded by the lack of uniformity among the states – each of which has its own unique regulatory requirements – and between the states and federal regulators. Moreover, with technologies quickly evolving, carriers face the constant risk that regulatory commissions will alter their treatment of new services after they have already been introduced. For example, ISDN is sold by local exchange carriers as a single, high bandwidth line. Each individual ISDN line contains three separate channels that can be used to send and receive information in any configuration that the subscriber wishes. In 1995, the Commission held that only one subscriber line charge (SLC) should be applied to each ISDN line.143 In its 1997 Access Charge Reform Order, the Commission changed course again: it ordered one, newly-calculated, ISDN-only SLC to be charged per ISDN line, but changed the amount of the SLC.144 There is no guarantee that these issues will not be revisited, and further reversals are possible. State regulation complicates things further. Delaware regulators, for example, invalidated Bell Atlantic's ISDN tariff because they thought that Bell Atlantic had not adequately explained why ISDN should cost more than regular analog service.145 A Delaware court overturned the ruling – more than nine months later.146 Regulatory uncertainty of this kind clearly acts as a strong disincentive to investment in ISDN facilities in the first place.

Numerous disincentives to invest have been created through specific pricing mechanisms. One example is the level at which the Commission has set the "productivity offset" or "X-factor." This is the annual adjustment that LECs are required to make to their rates based on anticipated increases in productivity as a result of more advanced technology or other efficiency-enhancing factors. The higher the "X-factor" is set, the more efficient LECs are expected to become, and the lower rates should be as a result (since rates are based on costs). The Commission has conceded that its latest Price Cap Order sets the X-Factor well above what is "reasonable."147 As a result, incumbent local exchange carriers have a disincentive to upgrade to advanced technology that would increase the level of the X-factor even further, since it would only reduce the rates that such companies could charge for the new advanced services they developed.

Finally, regulators have used their ratemaking authority to load costs on to certain services and not others. These kind of subsidy programs have the effect of encouraging carriers to enter markets and provide services based on artificial regulatory incentives rather than true economic incentives. In 1983, for example, the Commission exempted Enhanced Service Providers (ESPs) from paying access charges for their use of the local telephone network. In 1987, the Commission recognized that this distinction made little sense since ESPs use precisely the same local lines as are used for voice calls, and indeed use them (on a per-customer basis) substantially more.148 The absence of access charges for some providers, but not others, creates an uneven playing field that is a disincentive for many kinds of potential competitive entrants.

Unbundling and Resale Rules: Under the Act, incumbent LECs are required to unbundle and sell to competitors whatever new capabilities and services incumbent local exchange carriers add to their networks.149 The Act also grants the states the authority to set the prices for new network elements "based on the costs of providing" them.150 As currently formulated, the pricing standards developed in virtually all states require local exchange carriers to give competitors access to network elements at prices at or near the incremental cost of providing them, and below the actual book cost including capital and depreciation.

These pricing standards create strong disincentives for incumbent local exchange carriers to invest. On the one hand, if incumbent LECs make large investments in new facilities and services, competitors will be able to take these elements at cost.151 On the other hand, if no competitors take these elements, the investment made by the incumbent LEC will be largely unrecoverable. Under price cap regulation, local exchange carriers are not entitled to automatically recover their costs, but must bear the full risk of their investments.152

The unbundling rules also give competitors very little incentive to deploy broadband technology themselves. Competitors may buy unbundled pieces of the existing network below cost – as well as any successful new technologies at cost – with substantially less risk of losing unsuccessful investments. No economically rational new competitor will build anything that it can buy below cost.

Finally, the resale obligations of the Act require that carriers that develop new services provide these services to their competitors at wholesale rates.153 This places incumbent LECs in the position of being unable to differentiate their broadband services from those of their competitors. As with the unbundling requirements, the resale requirements force incumbent LECs to bear the entire risk of their investments in broadband services, without providing any assurance that customers (or competitors) will take the new service.

Cable operators are also subject to burdensome unbundling regulations of sorts. Cable operators must devote large portions of their systems' capacity to other programmers.154 Cable is required to devote one third of its channels to carry local TV stations, and is required to set aside additional channels for lease and "public access." Under these rules, the more capacity that a cable operator builds, the more it must give away to others. As a result, cable operators' incentives to expand their systems are more limited than they might otherwise be.

Many of these pseudo-carriage obligations also apply to new entrants that seek to compete against incumbent cable operators. For example, the Act establishes a new regulatory regime for providing video services – an Open Video System (OVS).155 The Commission subjects OVS operators to many of the same obligations that cable operators face regarding the carriage of local broadcast signals. The Act directs the Commission to extend its rules concerning sports exclusivity, network nonduplication, and syndicated exclusivity to operators of open systems. OVS operators are also required to comply with the obligations regarding "must-carry" of broadcast stations;156 leased-access channels for commercial use;157 carriage of channels for public, educational or governmental use,158 and carriage of noncommercial educational channels.159

Other Regulatory Impediments: The principal providers of broadband services today are large integrated carriers that operate national or international networks. For example, the major Internet backbones are controlled by the long-distance carriers, WorldCom, MCI, AT&T and Sprint.

Incumbent LECs are generally prohibited from providing interLATA services over their networks.160 Yet most of the services that use advanced broadband facilities are inherently interLATA in nature. Accordingly, this ban effectively extends to most types of "information services," including all types of Internet services except local Internet access.161 ISDN and ADSL services, for example, are rarely used to make local telephone calls, but principally to send and receive data on the Internet, corporate intranets, or on-line services like CompuServe and America Online. But since these data transmissions travel across LATA boundaries, incumbent LECs are restricted in their ability to provide online or Internet information services,162 and reducing their incentive to invest in the facilities that are used to provide local access to them.

The interLATA restriction likewise bars incumbent LECs from building Internet backbones or any other kind of regional or national broadband network. Internet backbones are, by definition, regional or national in scope. They necessarily cross LATA boundaries. It is regulatory restrictions – and not economics – that keep the incumbent LECs from participating in these markets.

The interLATA restriction also has been interpreted to impose equal access obligations on incumbent LECs’ local Internet access services.163 At present, incumbent LECs offering Internet access cannot provide or resell interLATA Internet carriage. Rather, an incumbent LEC typically asks its Internet customers to select a preferred interLATA Internet provider, in much the same way as the customer selects an interexchange carrier by way of a PIC. Other Internet Service Providers, however, are permitted to bundle local Internet access with interLATA Internet carriage. It is not surprising that most are doing so; it is a far more efficient arrangement as it enables the ISP to purchase interLATA carriage in bulk, often at steep discounts.

Additional economies of scope arise from the fact that facilities or employees can simultaneously perform tasks that are helpful to a firm's efforts in more than one activity. Being able to take advantage of these economies of scope is often an incentive for firms to make large incentives. At present, incumbent LECs face a wide array of separate affiliate requirements that prevent many economies of scope from being realized.

Under the Act, even after interLATA relief, Bell companies must provide all interLATA services – including interLATA information services – through separate subsidiaries.164 This requirement significantly increases the costs of designing and operating a broadband network by requiring the deployment of duplicative operations and personnel, and eliminating many of the efficiencies of integrated operations. Non-BOC incumbent local exchange carriers also must provide long-distance services through separate affiliates.165 All incumbent LECs (except for rural telephone companies) are required to provide all in-region broadband wireless services through a separate affiliate.166

    1. Removing barriers to broadband investment can be accomplished through a containment policy for regulation

PFF’s research indicates that the investment for broadband infrastructures, independent of the technology, will only be forthcoming in a "reasonable and timely fashion" if there is a containment policy for regulation. This containment policy should ensure that "advanced telecommunications services," which can be defined as all presently unregulated packet based services, should be free from regulation.

In essence this proposal distinguishes traditional "telephone exchange service" or "exchange service" from "advanced telecommunications services" based on the format and the data rate over the wired local loop: packet based services having data rates of 1.544 Mb/s or above in the downstream direction (central office/point of presence towards the subscriber) and 384 kb/s or above in the upstream direction (subscriber towards central office/point of presence). Packet based services are defined as those services in which information (of any type) is routed based on information contained within a header/addressing field as opposed to a position in a time, as occurs in circuit switched networks. Existing Title II services which meet these criteria (e.g. Frame Relay) would remain regulated under Title II.

The interconnection, unbundling and resale requirements of the Act in the local loop portion of the network would be met for Title II services via separation of packet based vs. narrowband circuit switched services at the twisted wire pair level (frequency unbundling) or at a switch/terminal which permits separation of packet based information from circuit switched information.

Pure price cap regulation – applied to regulated services only – must be adopted both at the Federal and State level in order to permit the use of integrated transport systems which will provide carriers with the ability to achieve economies of scope.

PFF’s research indicates that this delineation would allow the Commission to separate the historically regulated telecommunications equipment and services from advanced and emerging equipment and protocols which will provide the basis for an advanced telecommunications infrastructure. Failure to boldly deregulate advanced telecommunications services and attempts to adapt the complex system of telephone service regulation with its corresponding web of cross-subsidies will only serve to severely delay the deployment of advanced telecommunications capabilities to all Americans. In order to meet the statutory requirements of Section 706 the Commission must deregulate, not further regulate, advanced telecommunications services.

    1. Universal service issues

Taking the bold step of delineating traditionally regulated services from advanced telecommunications services and forbearing from regulating advanced services with the knowledge that the majority of services (including traditional voice services) may eventually migrate to unregulated platforms raises the issue of universal service funding. Specifically, the proposal forces one to consider how the Commission can knowingly take steps that might limit the funding base for the existing universal service program. However, the situation is not as dire as it might seem, given the high penetration rate of telephone services (93.9 % of US residences) and the fact that the migration of voice services from circuit switched to the proposed unregulated packet based networks will take upwards of a decade. The present reform of the universal service system will need to continue, but should be isolated from advanced telecommunications services.

What is clearly required for advanced telecommunications services is examination of the penetration rates for advanced telecommunications services which will be achieved through market forces, as well as the penetration rates over time which form the basis for sound public policy, minimize "redlining" and support rural areas. Once the deployment of advanced telecommunications capability begins en masse, it will be possible for the Commission to determine what explicit mechanisms can be used, if necessary, to ensure that "all Americans" have access to those services.

The present path of the Commission will lead to the mixing of narrowband subsidization policies with broadband issues, and can only serve to further promulgate a system in which prices for local loop services do not way reflect costs. Should this system be applied in any way to advanced services it will only serve to deter investment and slow competition.


  1. Conclusions

The NOI which initiated this proceeding asks a number of profound questions, but none more important than those contained in paragraph 80.

Looking into the future, we ask what, if any, system of regulation might best fit the market for advanced telecommunications capability. Enacting such a system might require major amendments to the Act. For example, it is reasonable to question a policy of regulating several competitors in a market differently . . . . [W]e ask parties to consider the Internet industry as a model of what a maturing market for advanced telecommunications capability and advanced services might be.

PFF’s research suggests that the Internet model is precisely the correct one for the market for advanced telecommunications services. However, the Commission does not have to, and indeed must not, await the passage of new legislation before embarking on the path towards implementing such a system.

As this filing shows, developments in the broadband digital marketplace are occurring at breakneck speed. Cable modem subscribership is growing at triple-digit annual rates; demand for broadband services is growing, arguably, even faster. Steps taken in the direction of additional regulation are difficult to undo, and, indeed, set in place their own "feedback loops" – more regulation begets still more regulation. If this cycle is permitted to continue, achieving the Internet model will become far more difficult, if not impossible. Thus, while the Commission may well need new direction from Congress before fully implementing the Internet model suggested in the NOI, it can take steps today to prevent the further spread of regulation.

    1. Intermodal competition, not regulation, is the key to encouraging the deployment of advanced telecommunications capability

The wide range of technologies which will be used to support advanced telecommunications services, coupled with the demand for bandwidth both at the business and the residential level will drive the deployment of infrastructures for these services – but only if left to grow unregulated. Once the path of regulation, as opposed to deregulation, is chosen, it will be difficult to escape the cycle of regulatory rulemaking, litigation, and regulatory reform. Intermodal competition, in particular the competition which could occur between cable operators and LECs, can provide the basis for the deployment of advanced telecommunications capability.

    1. Harmonization of public policy towards advanced telecommunications capability networks can only be accomplished through deregulation

      1. The dangers of "harmonization" through taking a Title II approach to packet networks and Title VI networks carrying data.

The NPRM associated with this NOI, which proposes separate subsidiaries for data services, is one step towards "harmonization through regulation" in which data services delivered by telephone companies are regulated either directly (as in the case of incumbent LECs) or indirectly by forcing separate subsidiary requirements on incumbent LECs. Embarking on such a path will necessitate reexamination of all of the existing rules regarding data and network operators including Title VI cable operators.167

Another example of how "harmonization through regulation" is bound to occur if the Commission does not adopt a containment policy towards regulation is in the area of Customer Premises Equipment ("CPE") and network interfaces on the subscriber side. Part 68 definitions for the narrowband telephone interface have been successful in creating a competitive market for telephone CPE, but it is frequently forgotten that the technology which enabled this interface to be developed was based on decades of private investment and what was essentially a closed architecture. Extending the methodology of Part 68 to the advanced telecommunications services industry which has not yet deployed significant amounts of equipment will only have a chilling effect on making new services available to all Americans. The NPRM168 raises the issues of standards for the advanced telecommunications services interfaces and a "broadband" part 68. The fact that incumbent LECs could choose to offer advanced telecommunications services as a regulated Title II common carrier will necessitate the Commission’s involvement in defining advanced services interfaces and may lead to intrusive regulation in this area.

      1. Bundling of services and CPE must be permitted to achieve economies of scope.

One of the powerful aspects of intermodal competition is that it provides subscribers with a choice of services including the end-user or Customer Premises Equipment. Initially, these services will not be such that "mix and match" between any two elements is possible, but due to the nature of the Internet and the availability of alternate technologies interoperability is not necessary. Cable modems may only allow connection to the Internet via a single ISP but provide high speed connectivity to the Internet with no restrictions. Telephone company (incumbent or competitive LEC) provided data services offer an alternative to cable, as will wireless service providers.

Because of the technologically advanced nature of these networks, service providers will base their deployments on architectures which have interfaces in the residence that may be specific to their equipment, but which end in standards-based interfaces (e.g. Ethernet or Universal Serial Bus) and provide connectivity with any number of devices in the residence.

Bundling of services and CPE should be seen less as a requirement for regulation and more as an incentive for investment. The wireless industry, where bundling of services and CPE is the industry norm, has not been harmed by these practices. The Commission should pay close attention to its successes in deregulation and apply these lessons to advanced telecommunications services.

    1. Anti-trust vigilance, not prophylactic regulation, is the key to encouraging deployment of advanced telecommunications capability

In the combined statement of Commissioner Michael K. Powell on the Order/NPRM,169 the Commissioner accurately points out that communications policy has historically emphasized prospective, prophylactic regulation which tends to stifle innovation and innovation and impede the beneficial operation of market forces. We note that in addition to stifling innovation and impeding market forces, regulation is self-perpetuating. Bold steps must be taken in order to break the vicious cycle of regulation.

Such a step could be taken by the Commission by defining advanced telecommunications capability in a manner that precludes its regulation, and exempts such services from any such regulation under Title II. Providing such a definition would allow incumbent LECs to provide advanced services on an integrated basis without being subject to interconnection, unbundling, and wholesale requirements, and would allow them to obtain the economies of scope needed to provide advanced services in a profitable manner at high penetration rates. There are valid concerns with respect to the market power of incumbent LECs, but such concerns should be dealt with as anti-trust issues, rather than removing incentives to deploy advanced networks and services.

Excluding advanced telecommunications capability from regulation would not undermine the Commission’s authority or ability to ensure that collocation and unbundling of twisted wire pairs is accomplished in a manner which allows competitive LECs to enter the market. Aggressive and timely enforcement of this aspect of the Act would ensure competition for narrowband services as well as the entry of competitive LECs into advanced services.

By providing a deregulated environment for advanced telecommunications services, the Commission can encourage investment by the groups able to provide an infrastructure for advanced telecommunications services: incumbent and competitive Local Exchange Carriers, cable operators, and wireless service providers. Any regulation which significantly stifles investment from any one of these sectors threatens the competitive marketplace and will lead to a slow, highly regulated rollout of advanced services which may have grave economic consequences.


APPENDIX A: A proposal for packet based definition of broadband services and containment of regulation

Although there are many details and nuances to the technology surrounding the Internet, there exists a clear distinction between the circuit switched technology on which the telephone industry has been built, and the packet based system upon which the Internet is based. This distinction can be used to segregate the historically heavily regulated circuit switched telephone network from emerging packet based networks, which can subsequently be left unregulated to allow unbridled growth.

It is not a trivial task to separate technologies and establish the appropriate regulatory (or perhaps "unregulatory") mechanisms for insuring that the historical circuit switched network is appropriately deregulated while isolating the packet based services. In addition to the basic technical matter of distinguishing the technologies, there are complex issues related to the of unbundling of twisted wire pairs which can simultaneously transport circuit switched narrowband services and packet based broadband services, and avoiding inappropriate cross-subsidization of the packet based network by revenues from the circuit switched network. In addition, many parts of the network will simultaneously carry circuit based services and packet services, or transport circuit based services encapsulated in packets or visa versa. The goal here is not to separate all network elements and services, but rather to allow deployment of new local loop packet services in an unregulated mode. The present regulation of circuit based networks, including those that are used to transport packets, does not have to be discarded entirely, and the slow and painful process of deregulation of those networks will continue separate from the deployment of packet based networks and services.

This proposal discusses how local loop packet based services can be isolated from the burdensome regulatory history associated with circuit switched services. This isolation or "containment" of regulation will enable deployment of advanced networks based on demand and permit true competition in the area of broadband networks. This proposal does not attempt to look at all of the network and backbone issues related to packet services and deregulation, but instead focuses on the technical definition of packet based services which would allow separation of packet and circuit switched services in the local loop and the data rates required to ensure that the network is indeed broadband.

Technical issues related to unbundling are discussed, and it is shown that narrowband services can be separated from packet based services on both twisted wire pair drops, as well as in integrated transport systems in which the circuit switched data may be carried in a packet or cell format through much of local loop. Finally, cost allocation issues are addressed, and it is proposed that pure price cap regulatory schemes will need to be applied to narrowband circuit switched services while simultaneously deregulating broadband packet based services in order to provide appropriate incentives to deploy technologically innovative solutions in the local loop.

This proposal does not suggest that the existing plans for deregulation of the narrowband network are unworkable nor that the careful monitoring of the unbundling of network elements of existing Local Exchange Carriers and their entry into the provisioning of long distance services is unnecessary, but suggests that a bolder deregulatory approach to the deployment of packet based services is required.

Overview of Packet Technology

In order to understand distinctions between packet based and circuit based technologies it is necessary to review the basics of circuit switched telecommunications systems. Although the use of human operators to establish connections between subscribers in the earliest telephone systems was already a type of circuit switching, it was the advent of the Strowager mechanical switch170 which allowed for automated connections between subscribers based on the number dialed, and was the first instance of automated circuit switching.

In the initial analog telephone system, circuits were initially switched based on simply establishing a mechanical connection between subscribers. Later, Frequency Division Multiplexing (FDM) technology was used to aggregate calls in the long haul network, with each subscriber being assigned a bandwidth (in the form of a range of frequencies) within a frequency multiplex, and calls could be removed and added from the frequency multiplex. The advent of digital switching allowed for calls to be aggregated based on the concatenation of voice samples generated by digital sampling of the voice signal received at the central office. Digital switching, in which the binary representation of the voice signal is manipulated and individual bytes of information are routed to perform the switching, forms the basis of the existing telecommunications network in the US.

The sampling of the voice signal at a rate of 64 kb/s produces an acceptable quality signal, and in order to reconstruct the signal properly (without gaps in speech or excessive delay) the signal is transported through the telecommunications network as one byte of user information every 125 ms. In order to ensure that the information is transported at a constant rate, each call or circuit is reserved time, also known as a "time slot" in the network for signal transport.

The basis for telecommunications services in the US is illustrated in Fig. 1, which shows a DS-1 signal frame, composed of a framing bit followed by 24 channels of voice data, each channel composed of one byte of voice information (with some bits occasionally used for signaling). The resulting composite of 24 channels, each having one byte transmitted every 125 ms, plus the framing bit, is a signal of 1.544 Mb/s. Numerous other transmission rates and formats are possible, but all are fundamentally based on the concept of transmitting voice channels at a data rate of 64 kb/s with minimal delay.171 Additionally, the routing of the information is based upon the temporal location of the information: a subscriber calling from location A to location B may be assigned channel number 2, the second "time slot" in a DS-1 and thus all bytes appearing in channel 2 of a DS-1 signal from location A will be routed to location B.

Figure 1. DS-1 circuit switched frame structure.

Why Packet Technology is Different From Circuit Switched Technology

Packet technology has fundamentally arisen from the interconnection of computers, where large amounts of data needed to be sent between machines, with little regard for a constant (smooth) data rate, but with the need for rapid transport of large blocks of data. Hence the computer based packet protocols which have been developed traditionally work on the basis of seizing the bandwidth when available, and transporting large amounts of information in a short period of time. This simplification does not discount the large variants on the design parameters of Internetworking, Wide Area Network (WAN), Local Area Network (LAN) and Metropolitan Area Network (MAN) systems,172 but demonstrates that the development of packet based transport technology was motivated by interconnection of computers rather than subscribers using voice communications.

Figure 2 illustrates basic packet structures including (a) the IEEE 802.3 frame format, which is related to the original baseband Ethernet specification; (b) the Internet Protocol (IP) datagram,173 which forms the basic unit for the transport of data across the Internet; and; (c) the Asynchronous Transport Mode (ATM) cell structure.174

Figure 2. Frame structures for packet based communications including (a) IEEE 802.3 (Ethernet) frame structure; (b) TCP/IP datagram; and (c) ATM cell.


The common thread between these structures is that they are all designed to transport multiple bytes of data, and the routing of the data is dependent upon self-contained information at the front (header) of the packet or cell. In no way can this structure be confused with a circuit switched signal in which a "time slot" is reserved and the destination of the information is based on the temporal location of the information.

Within packet based systems there are numerous differences. As an example the packets of Fig. 2 (a) and (b) are variable in length, while the ATM cell in (c) is of fixed length. In addition, some packet based systems are connectionless (e.g. the Internet) and packets can take different routes to the final destination, while ATM systems are connection oriented, meaning that cells are routed to their destination based on pre-established connections. There is considerable discussion as to whether IP or ATM will be the ultimate transport scheme, and hybrid schemes using IP routing to establish ATM connections are actively being explored.175 It is unlikely that these disputes will be resolved in the near future, nor is it clear that any of the pure ATM, pure IP, or hybrid ATM/IP solutions will be adopted by the marketplace. It is entirely possible that new solutions will be developed which may not be based on the existing packet formats and protocols, but it is almost certain that they will be packet based, rather than circuit switched.

Voice services can be transported within all of these packet based systems, with the fundamental problem being one of delay: if one waits long enough to fill the packet or cell with voice samples, the time expended results in problems of noticeable echo or delay in the voice signal.176 Additionally, the packet protocols have different (or in some cases lack) procedures for dealing with lost, delayed, or out-of-sequence packets. Voice services can be transported over any of the packet protocols, but with different levels of success, and it is fair to say that the quality generally does not match that of the existing circuit switched voice network.

This is not to say that these problems cannot be overcome: in fact it is likely that packet based technology will ultimately be a replacement for circuit switched technology. However, there are numerous technological hurdles to be overcome, and a fair amount of time will pass before packet technology becomes the predominant transport mechanism for traditional voice services. Given the embedded base of circuit switched technology for voice services, it is likely to be several decades before the majority of voice traffic migrates to packet based infrastructures.

Deregulating Packet-Switched Technologies

Given that distinctions can be made between the circuit switched and packet networks, one approach to ensuring rapid and unencumbered growth of packet based services is to eliminate regulation of the packet based network.

In the following discussion the concept of service separation rather than network separation is used because it the packet transport service, or "broadband packet pipe" which is being distinguished from the circuit switched services. The end service application being provided over the packet network is inconsequential – once the packet based transport service is deregulated any applications which utilize the broadband packet pipe are deregulated- regardless of the application. This situation is quite different from the historical circuit switched network, where the service and network were indistinguishable. The initial phone network was only used for voice transport, and was a private network during its initial stages of deployment. Subscribers did not have the ability to separate the service from the transport. Packet based transport, and in particular the Internet Protocol ("IP"), is a public standard upon which a multitude of applications have been built, and in which applications can be run independent of the specifics of the physical and logical layers of the network.177 The interfaces for the packet based transport services are defined by public protocols (e.g. IP) and standard physical interfaces (e.g. Ethernet) and although it is likely that these protocols and interfaces will evolve or be replaced by new ones, there is no doubt that the tremendous advantages (network externalities) generated by global interconnectivity will ensure that the ensuing protocols and standards will be open rather than proprietary.

Despite the significant differences between the telephone network and the Internet, critics will rapidly point out that a}the success and growth of the Internet was based upon the use of the extensive and heavily regulated switched telephone network, and b}deregulating broadband packet based services will result in private networks and new broadband monopolies in the local loop. While point a} is certainly true, future growth of the Internet will depend heavily on the deployment of new packet based infrastructure, both backbone and local loop, since the existing narrowband infrastructure is entirely inadequate for high-speed services. As for the possibility of private broadband networks and local loop monopolies, it is clear that extensive private investment is required to deploy broadband networks, but that those networks will necessarily be open at the application layer in order for the investors to realize the benefits of network externalities.178

Defining Packet Based Services

A proposed definition of packet based services for the purposes of deregulation is:

Packet based local loop services shall mean all advanced telecommunications services provided using packet or cell based transmissions, over any transmission media, between a central office, head end, point of presence, or equivalent central facility, and packet or cell based terminating and origination equipment at a subscriber location where the packet based nature of the transmission permits routing of the packet or cell based on addressing information contained within said packet or cell, and wherein the peak data rate supported is no less than x Mb/s in the central facility to subscriber direction and y Mb/s in the subscriber to central facility direction.

The variables x and y in this definition would be set to ensure that the deregulated network is of sufficient capacity to ensure that the services offered are indeed broadband, and not just packet substitutes of the narrowband services (e.g. POTs over low-speed Internet connections). As an initial proposal, the downstream direction rate (x) is envisioned to be 1.5 Mb/s and the upstream direction (y) 384 kb/s.

It is likely to be necessary to put lower limits on the average data rates to preclude systems which can support high peak data rates to an individual subscriber at the expense of the rest of the subscribers on the network. Such systems could qualify for the deregulatory advantage but would not provide high bandwidth to all subscribers simultaneously. As an example, it is entirely possible to have a wireless or CATV system in which it is theoretically possible to give 6 Mb/s to one individual subscriber at the expense of the rest of the subscribers. At an extreme this would mean that a broadcast type infrastructure which serves 500 subscribers with one 6 Mb/s downstream channel and one 1.5 Mb/s upstream channel could be entirely unregulated. The additional restriction on the average data rate would ensure that the system is sufficiently broadband, or in economic terms, that there has been a substantial investment per subscriber.179 The additional restrictions would be:

…and wherein the average data rate supported to all subscribers connected to the network is no less than 384 kb/s in the central facility to subscriber direction and 128 kb/s in the subscriber to central facility direction.

Further restrictions on existing services would keep existing regulated packet services (e.g. frame relay) from becoming deregulated:

Such provisions shall not apply to existing Title II data services.

To see how such a definition would work in practice it is necessary to look at some practical situations in which both packet based and circuit switched equipment is employed. Fig. 3(a) illustrates a system in which services are provided in the local loop using a cell based format, which in this case is an ATM transport system. Even when the service is routed to a Time Division Multiplexed (TDM) circuit switch the local loop service itself remains packet based. If such a service met the given definition it would be exempt from regulation. Since the deregulatory effect is intended for the local loop, the format of the routing of the information on the backbone network has not effect on the definition of the local loop packet service.

Fig. 3(b) illustrates cases in which packet service providers lease or own local loop facilities, as well as the case in which the packet service provider leases bandwidth through a TDM circuit switch. In the case of leasing bandwidth, the service is exempt from regulation when the specified data rates are met. Similarly, if the service is provided by direct interconnection through leased or owned local loop facilities it would also be exempt from regulation when the specified data rates are met.

The conclusion that can be drawn is that when a containment philosophy is applied to the local loop, separation of packet based services from circuit switched services is possible, even if there are circuit switched elements in the rest of the network. There are clearly numerous other issues related to deregulation of the backbone network which are not dealt with here, but which can be addressed separately.




Figure 3. Local loop packet based transport architectures including (a) use of a circuit (Time Division Multiplex) switch at the network side of the packet based transport; and (b) delivery of packet based services over separate or leased facilities.





Interconnection and Unbundling in a Broadband Environment

Because circuit switched local loop services will co-exist with emerging packet services for a considerable time period, it is necessary to consider how the regulated circuit switched infrastructure can be unbundled and interconnected to meet the Section 251 requirements of the Act.180

Fig. 4(a) illustrates the simplest case where loops are unbundled, and POTs services reside on loops adjacent to some form of Digital Subscriber Loop (xDSL) transmission. Many xDSL transmission technologies were developed to be POTs compatible, and thus present no significant problems with respect to having high-speed digital signals in the same bundle of wires as an analog voice signal. Additionally, all DSL technologies were developed to permit multiple pairs to transmit digital signals without interfering with each other- this in fact is the principal challenge in DSL system design. The conclusion is that DSL technologies are compatible with POTs signals, and DSL signals are by design compatible with other DSL signals in a bundle,181 so and unbundling of the pairs presents no technological difficulties.


Figure 4. Unbundling and interconnection for narrowband services based on (a) loop unbundling; (b) narrowband service unbundling; and (c) narrowband service unbundling in an integrated transport system.


Fig. 4(b) illustrates the case in which a twisted wire pair path is used to provide both narrowband services (typically analog POTs) as well as broadband packet services. In this case splitters (separation filters) are used at both ends to separate the packet based broadband signals from the narrowband signals. There are newer "splitterless" technologies which do not require a physical splitter, but in any case the ability to separate the services is clear.

Finally, Fig. 4(c) illustrates the case in which an integrated transport system is used which carries the circuit switched and packet based services over unified platform: there are numerous examples of such systems including Fiber-to-the-Curb (FTTC) and Fiber-to-the-Home (FTTH) systems, as well as the Hybrid Fiber Coaxial (HFC) systems used by cable operators. In these systems, the circuit switched data may be carried within packets or cells directly from the central office or head-end to the residence. For traditional POTs services, the circuit switched signal is reconstructed at or near the residence, and an analog telephone signal is presented at an interface (e.g. RJ-11 jack). Similarly, the circuit switched information is made available at the network side in either digital or analog form. From a technical perspective, there are no problems related to the removal of the circuit switched information from the packet based services at either end of the network.

Cost Allocation

One of the most complex and perhaps most ominous barrier to broadband deployment is the issue of cost allocation and what is perceived by many as the "necessity to avoid cross-subsidization of new broadband services by regulated services." Upon closer examination it appears that most of the arguments forwarded in this area are related to fears regarding the dominant role incumbent LECs would play in the broadband services market. While these fears are certainly not unfounded, it is important to note that building new, high penetration wired broadband networks may have economies of scale and scope which requires significant investment for these new services, at least in the short term, until penetration rates are significant and the demand (and correspondingly the price) results in significant intermodal competition. If significant penetration rates are achieved, the demand for bandwidth will be stimulated to the extent that other facility based competitors can enter the market and stable competition can exist.

The ability to move forward on broadband deployment and resolve the issue on cost allocation is dependent on establishing pure price cap methodologies for narrowband services, both at the federal and state levels. On one hand, the move towards price cap regulation at both levels is already taking place, and the adoption of price cap models for long distance service by the Commission, and for local exchange service by many of the states, is encouraging.182 On the other hand, the contentious issue of cross-subsidization of video services by regulated narrowband services has been addressed but never resolved by the Commission.183 The initial proposals to trigger decreases in price cap indices based on "exogenous" changes such as the offering of video programming or other unregulated activities184 would create a huge disincentive for the deployment of broadband infrastructure and completely undermine the logic behind price cap regulation.

Even for traditional narrowband services cost allocation issues are almost impossible to resolve.185 In the face of rapid technological change and the growing demand for bandwidth, it will be an impossible task to adequately monitor and separate broadband and narrowband service costs and determine prices. By adopting pure price cap methodologies for narrowband circuit switched services and allowing packet based broadband services to be deployed on an truly unregulated basis (e.g. no exogenous adjustments) it will be possible to achieve significant penetration rates for new services. Achieving significant degrees of penetration will result in market sizes which will result in competition, and prevent long term natural monopoly situations from existing.


The proposal offered here attempts to illustrate that there are fundamental differences between the existing circuit switched telephone network and emerging broadband packet based services. These differences can be used to establish a barrier between regulated narrowband services and new broadband services which will need to remain unregulated in order to foster private investment in network infrastructure and establish significant penetration rates for these new services.

The issues of unbundling of narrowband network elements to comply with the Telecommunications Act of 1996 in scenarios where packet based and circuit switched services are carried over the same network elements are complex. Nevertheless, there exists the ability to separate the services at the end points (subscriber location and central office or point-of-presence). This will permit further deregulation of the narrowband network without burdening the packet based network with regulation.

Finally, the issue of cost allocation is addressed. A movement to pure price cap methodologies for narrowband services, both at the state and federal levels, is essential to permit deployment of broadband networks



APPENDIX B: Digital Broadband Working Group Briefing Attendees

Maggie Barrington


US WEST, Advanced Technologies

Thomas Barry

Senior Vice President, Federal Relations

SBC Communications, Inc.

Marc Berejka

Federal Regulator Affairs Manager


Robert Blau

Vice President of Executive & Federal Regulatory


Debra Brunton

State Affairs Representative


Philip Burgess

President and Chief Executive Officer

Center for the New West

Jeffrey Campbell

Manager, Federal Government Affairs


Dave Charlton

Business Development Manager


John Charters

Vice President, Internet Services & Application Development

U S WEST Communications

Larry Clinton

Assoc. Vice President of Large Company Affairs


Scott Cooper

Government Affairs Manager


Ophyll D’Costa

Executive Director, Strategic Development

U S WEST Communications

David Dorman

Chief Executive Officer

PointCast, Inc.

Jeffrey Eisenach


The Progress & Freedom Foundation

Charles Eldering


Telecom Partners Ltd.

Robert Frankenberg

President and Chief Executive Officer

Encanto Networks, Inc.

Paul Fuglie

Assistant Vice President, Regulatory

GTE Corporation

Dick Green

President & Chief Executive Officer


Don Gips

Chief Domestic Policy Advisor

The Office of the Vice President

Dennis Glaves

Assistant Vice President, Congressional Affairs

GTE Corporation

Gita Gopal

Department Manager, H-P Labs


Rob Griffen

Regulatory Counsel, Information Services Group

Bell Atlantic

Mike Grubbs

Director, Convergence Products

Gateway 2000

Tim Hackman

Director of Public Affairs


Christine Hemrick

Vice President/General Manager, Internet Appliances and Applications Business Unit

Cisco Systems, Inc.

Tony Hennon

Senior Vice President


Grace Hinchman

Manager, Public Affairs

Digital Equipment Corporation

Link Hoewing

Director, Issues Analysis

Bell Atlantic

Laura Ipsen

Manager, Government Affairs

Cisco Systems, Inc.

Ted Jenkins

Vice President & Director of Corporate Licensing


George A. (Jay) Keyworth


The Progress & Freedom Foundation

Robert Kirkwood

Director of Government Affairs


Kal Krishnan

Sr. Vice President & Chief Technology Officer

Encanto Networks, Inc.

David Krone

Vice President of Government Relations

TCI, Inc.

Mary McManus

Director of Federal Public Affairs

Lucent Technologies

Donald McClellan

Senior Fellow

The Progress & Freedom Foundation

Garland McCoy

Vice President for Development

The Progress & Freedom Foundation

Jill Murphy

Director of Communications

The Progress & Freedom Foundation

Alex Netchvolodoff

Vice President for Public Policy

Cox Enterprises, Inc.

Donna Northington


US WEST, Advanced Technologies

Michele Obermeier


US WEST, Advanced Technologies

Robert Pepper

Chief, Office of Plans and Policy

Federal Communications Commission

Jeffrey Peters

Vice President, Digital & Applied Imaging

Eastman Kodak

Mike Pettit


Spence, Fane, Britt & Browne

Lewis Platt

Chief Executive Officer


David Porter

Vice President Government Affairs


Bruce Posey

Vice President, Federal Relations

U S WEST Communications

Tim Regan

Vice President & Director, Federal Government Affairs

Corning, Inc.

John Robinson

Vice President, Strategic Planning


Daniel Scheinman

Vice President, Legal & Government Affairs

Cisco Systems, Inc.

Paul Shumate

Executive Director of Broadband Local Access


Michael Schwartz

Senior Vice President, Communications


Steven Stewart

Program Manager, Telecommunications Policy, Government Programs


Tim Stone

Vice President, Business Development


Raymond Strassburger

Director, Government Relations-Telecommunications Policy

Northern Telecom

Lynn Streeter


US WEST, Advanced Technologies

Howard Symons


Mintz, Levin, Cohn, Ferris, Glovsky & Popeo, P.C.

Thomas Tauke

Senior Vice President, Gov. Relations

Bell Atlantic

Solomon Trujillo

President & Chief Executive Officer

U S WEST Communications

Jan Wadsworth

West Coast Counsel

America Online

Timothy Waters

Vice President, Data Product Management


Thomas Wheeler

President and Chief Executive Officer


Bud Wonsiewicz


US WEST, Advanced Technologies

Joseph Zell

President, !NTERPRISE Networking Services

U S WEST Communications


1.        The Digital Broadband Working Group was convened by The Progress & Freedom Foundation at the request of Hewlett-Packard Chairman, President and Chief Executive Officer Lewis E. Platt and US West President and Chief Executive Officer Solomon J. Trujillo. Participants in the group's several meetings included representatives of all major segments of the computing and communications industries, and are listed in Appendix B of this filing. The Progress & Freedom Foundation is grateful to all of those who participated. The views presented here, however, are those of the authors and do not necessarily reflect the views of the Digital Broadband Working Group as a whole or of any of its individual participants.


2.        Pub. L. 104-104, Title VII, § 706, Feb. 8, 1996, 110 Stat. 153, reproduced in the notes under 47 U.S.C § 157.


  1. NPRM, paras. 18, 69-79.


4.        NPRM, paras. 40-42.


5.        NOI, para. 13.


6.        NOI, para. 14.


7.        xDSL includes High Speed Digital Subscriber Line (HDSL) technologies for providing up to 2 Mb/s of bi-directional data service over two twisted wire pairs, Asymmetric Digital Subscriber Line (ADSL) technology which can provide data rates of up to 6 Mb/s in the downstream and 640 kb/s in the upstream direction, and Very High Speed Digital Subscriber Line (VDSL) technology, which can support data rates of up to 25 Mb/s in the downstream over a limited distance.


8.        C. Eldering, N. Himayat and F.M. Gardner, "CATV return path characterization for reliable communications," IEEE Communications Magazine, vol. 33, no. 8, pp. 62-69 (August 1995).


9.        N. Omoigui, M. Sirbu, C. Eldering, and N. Himayat, "Comparing Integrated Broadband Architectures from an Economic and Public Policy Perspective," in The Internet and Telecommunications Policy Research, G.W. Brock and G.L. Rosston, eds. (Lawrence Erlbaum Associates, Mahwah, NJ, 1996).


10.     NOI, para. 59.


11.     Network Wizards, Internet Domain Name Survey, January 1998, WWW/report.html. A host used to be a single machine on the Net. Today, a single computer may host multiple systems (with multiple domain names and Web addresses). The January 1998 data reflect a new methodology for counting domain names, and the January 1998 data therefore are not strictly comparable to previous figures.


12.     A. Rutkowski, General Magic, Internet Trends, Feb. 1997,


13.     Breakthroughs, U.S. News and World Report, Dec. 25 1996/Jan. 1 1996, at 101-104, 106-108.


14.     Intelliquest Press Release, Latest Intelliquest Survey Reports 56 Million American Adults Access the Internet/Online Services, Nov. 18, 1997,


15.     Lynn Margherio, et al, The Emerging Digital Economy (Washington, DC: U.S. Department of Commerce, 1998), p. 2. Hereafter Emerging Digital Economy. See also Bob Woods, "Net Users Break 100 Million Barrier – Study," Newsbytes, June 30, 1998, reporting on a study by Matrix Information and Directory Services. The study projects 707 million Internet consumer users by January 2001.


16.   See P. Rolfes, Novell CEO: Networks Put a Face on Internet, Columbus Dispatch, Nov. 20, 1997, at 1F (130 million by 2000); J. Welsh, AFather@ of Internet Expects Bright Future for Technology, Wisconsin State Journal, Nov. 18, 1997, at 1C (300 million by 2000, according to Vint Cerf); B. Metcalfe, CamCon 97 Draws Out the Digerati to Ponder the Future of the Internet, InfoWorld, Nov. 17, 1997, at 187 (1 billion by 2000, according to Nicholas Negroponte).


17.     D. Molony, The Big Squeeze, Communications Week International, February 3, 1997 (citing Carl Howe, senior analyst at Forrester Research Inc, Cambridge, Massachusetts).


18.     Compare M. Miller, Looking Back – 1990-94: The Windows Era, PC Magazine, Mar. 1997, with Sneak Peek: Dell Dimension XPS D266 vs. Gateway G6-300XL, Oct. 1997,


19.     See A. Goldstein and J. Files, Chip Industry Needs Helping Hand, Dallas Morning News, June 16, 1997, at 2D.


20.     Compare R. Manning, All Systems Go, Sacramento Bee, Feb. 6, 1996, at D1 with Sneak Peek: Dell Dimension XPS D266 vs. Gateway G6-300XL, Oct. 1997,


21.     More Traffic on The I'way, Industries in Transition (Jan. 1997).


22.     See M. Janah and M. Thyfault, Networks: Telecommuters Find Data-moving a Snap, InformationWeek, Apr. 7, 1997.


23.     L. Bowman, WorldCom Sounds the Bandwidth Alarm, PC Week Online, Jan. 29, 1998,; M. MacLachlan, WorldCom Makes Megadeals to Develop Network Infrastructure, Internet Week, Oct. 6, 1997. See also Emerging Digital Economy, p. 2.


24.     S. Murray, Quieter Future Forecast, Baltimore Sun, Jan. 23, 1998, at 1C.


25.     More Traffic on The I'way, Industries in Transition (Jan. 1997).


26.     See Duesterberg and Pitsch, The Role of Competition and Regulation in Today's Cable TV Market (Washington, DC: The Hudson Institute, 1998), p. 25. Nearly 60 percent of cable subscribers have access to 54 channels of programming, or more.


27.     See FCC Trends in Telephone Service 33 (Mar. 1997) (The level of local calling has remained relatively constant over a long period of time despite the introduction of facsimile machines, computer modems and other devices that use telephone lines. Increases in long distance calling have caused the total usage per line to increase from 46 minutes in 1980 to 52 minutes in 1995.?). The "minutes" counted by the FCC actually double-count the total usage of the line.


28.     According to Jupiter Communications, at the end of 1996 nearly 40 million households in the U.S. have PCs, for a penetration rate of 39 percent. Over 70 percent – 27 million – of these households have modems connected to their PCs, and over half of them – 15 million – use the Internet or other online services. Jupiter Communications Press Release, New Devices and Technologies Will Drive Net Into 36 Million Homes by 2000, Jan. 6, 1997.


29.     U.S. sales of modems increased from 3.3 million units in 1990 to 19 million in 1995. Forecast Shipments of Modems, Computer Industry Forecasts, Jan. 15, 1997, citing High-Performance Modems, Information Week, Oct. 21, 1996, at 26. The MMTA attributes virtually all of this growth to the rise in Internet use. MultiMedia Telecommunications Association, 1996 MultiMedia Telecommunications Market Review and Forecast 13 (1996). The Data Analysis Group forecasts sales of 47 million modems in three years.


30.     As IDC Research observed, declining PC prices and the introduction of the sub-$1,000 PC have helped to drive consumer adoption of the PC and therefore, online services in the home. IDC Research Press Release, Web Has Reached Mass Market Proportions (1997) (quoting IDC senior analyst Jill Frankle).


31.     J. Carlton, PCs Under $1,000 Attract New Kinds of Customers, Wall St. J., Jan. 26, 1998, at B8.


32.     See:


33.     Lee L. Selwyn and Joseph W. Laszlo, Economics and Technology, Inc, The Effect of Internet Use on the Nation's Telephone Network (Jan. 22, 1997) (prepared for the Internet Access Coalition). According to Bell Atlantic, between one-third and two-thirds of second lines are purchased solely for data services, such as Internet access and fax machines. Joint Comments of Bell Atlantic and NYNEX on Notice of Inquiry at 11, Usage of the Public Switched Network by Information Service and Internet Access Providers CC Dkt No. 96-263. (F.C.C. Mar. 24, 1997).


34.     U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 26-8 (1997).


35.     1996 and 1997 America Online Annual Reports; J. Kornblum, AOL Reaches 11 Million Market, CNet, Jan. 20, 1998,,4,18294,00.html.


36.     See Robert W. Crandall and Charles L. Jackson, Eliminating Barriers to DSL Service, unpublished manuscript, May 1998.


37.     More than 100,000 U.S. companies use EDI daily to send electronic purchase orders, invoices, and other standardized documents to suppliers and customers.


38.     38 percent of all computers are owned by large businesses, and 31 percent by small and medium businesses.


39.     According to IDC, over half of the business PCs in the United States are now connected via a LAN and 80 percent of all organizations with more than 100 employees have a LAN installed. U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 27-14 (1997).


40.     New York and the Future of Office Technology, Westchester County Business Journal, May 5, 1997, at 16 (citing IKON study).


41.     See U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 28-24 (1997).


42.     U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 27-14 (1997).


43.     C. Anderson, In Search of a Perfect Market, Survey – Electronic Commerce, The Economist, May 10, 1997, (hereinafter Economist Survey).


  1. S. Salamone, It's Not Your Mother's T1 Anymore, TechWire, Apr. 23, 1997. That translates into an additional 300,000 T1 lines this year and another 380,000 next year, bringing the total number of installed lines to 1.98 million by the end of 1998. Id. Companies that want to establish a presence on the Internet need more connectivity, said industry analyst David Strom of David Strom Inc. in Port Washington, N. Y. The proliferation of Internet service providers has made T1 a cost- effective choice. The monthly charge for a T1 is a function of distance. The more ISPs, the more POPs, Strom said, noting that most corporations will find they have an ISP POP in their vicinity. Id.


45.     U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 28-22 (1997).


46.     Economist Survey, supra.


47.     Economist Survey, supra (citing Forester, Yankee Group, IDC and Cowies/SIMBA).


48.     J. Linehan, A Bit Too Cautious, Perhaps?, Communications Week, July 21, 1997, at 36.


49.     Economist Survey, supra (citing Forrester, Yankee Group, IDC, Cowies/SIMBA, Jupiter, Multimedia Research Group). Another research firm, Mountain View, California-based INPUT, predicted in a February 1996 report that worldwide sales of goods and services traded over the Internet would grow from $70 million in 1995 to $255 billion in 2000, a compounded growth rate of more than 400 percent a year.@ U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 28-24 (1997).


50.     Emerging Digital Economy, p. 35.


51.     The various factors driving the demand for digital bandwidth have had a profound impact on the U.S. economy. The Department of Commerce reports that the Information Technology sector of the economy has accounted for 25 percent of U.S. GDP growth in recent years and that "IT and electronic commerce can be expected to drive economic growth for many years to come." Emerging Digital Economy, p.2.


52.     See Kevin Werbach, "Digital Tornado: The Internet and Telecommunications Policy," Federal Communications Commission, PFF Working Paper Series #29, March 1997, p. 3. ("[I]t is valuable to understand the Internet as a feedback loop. A feedback loop occurs when the output of a system is directed back into the system as an input. Because the system constantly produces fuel for its own further expansion, a feedback loop can generate explosive growth.")


53.     U.S. Department of Commerce, U.S. Industry & Trade Outlook >98, at 26-4 (1997).


54.     S. Galante, 1997 IPOs Sputter, Then Pick Up, CNet, Dec. 22, 1997,,4,17554,00.html.


55.     S. Lohr, Information Technology Field Is Rated Largest U.S. Industry, N.Y. Times on the Web, Nov. 18, 1997, (citing American Electronics Association study).


56.     S. Lohr, Information Technology Field Is Rated Largest U.S. Industry, N.Y. Times on the Web, Nov. 18, 1997, (citing American Electronics Association study).


57.     The FCC's Kevin Werbach noted that the Internet, however, is driven by a particularly powerful set of self-reinforcing conditions. Some supply factors (such as the availability of higher-capacity networks) permit an expansion of demand (for example, by allowing bandwidth-intensive services such as high-resolution video transmission). Like a digital tornado, the vortex continues, as the new level of demand creates the need for additional capacity, and so forth. See K. Werbach, FCC Office of Plans and Policy, OPP Working Paper No. 29, Digital Tornado: The Internet and Telecommunications Policy 4, 5 (Mar. 1997).


58.     Connecticut Research, 1994 Local Telecommunications Competition at II-15 (6th ed. 1994); New Paradigm Resources Group & Connecticut Research, 1997 Annual Report on Local Telecommunications Competition 28 (8th ed. 1997).


59.     Connecticut Research, 1994 Local Telecommunications Competition at II-1 (6th ed. 1994); New Paradigm Resources Group & Connecticut Research, 1997 Annual Report on Local Telecommunications Competition 7 (8th ed. 1997).


60.     J. Keller, Ex-MFS Managers Plan to Build Global Network Based on Internet, Wall St. J, Jan. 20, 1998; S. Schiesel, From a Supplier of Gas Comes a Digital Pipeline, N.Y. Times, Jan. 12, 1998, at D10.


61.     In 1996, telecommunications mergers topped $100 billion in value; 1997 followed with more than $90 billion in deals. P. Truell, Buoyant Stock Market Keeps Mergers in Pipeline, N.Y. Times, Jan. 5, 1998, at D3.


62.     Comments of the United States Internet Providers Association, Usage of the Public Switched Network by Information Service and Internet Access Providers, CC Dkt. No. 96-263 (F.C.C. filed Mar. 24, 1997).


63.     J. Rickard, Boardwatch Directory of Internet Service Providers (Fall 1997),


64.     P. Elstrom, New Boss, New Plan, Business Week, Feb. 2, 1998, at 122; Terrestrial Services Market Reaches $218.3 Billion, According to IDC, PR Newswire, Jan. 21, 1997.


65.     P. Vadlamudi, Amid the Churn and Change, ISP Market Keeps on Growing, Investor's Business Daily, Nov. 13, 1997, at A8 (citing Maloff Group estimates).


66.     1996 and 1997 America Online Annual Reports.


67.     ISDN supplies two 64 kb/s channels that can be used for either voice or data, together with one 16 kb/s network signaling channel.


68.     Robin Gareiss, Mapping a High-Speed Strategy, Data Communications, Apr. 1997, at 62. As of the end of 1995 (the most recent date for which accurate FCC data is available), ISDN was available on 62 percent of access lines (65 percent of Bell Companies' lines) and 20 percent of all switches were ISDN capable (29 percent of Bell Companies' switches).


69.     Because ISDN "cannot be done on a per-user basis," it can cost upwards of $500,000 to upgrade a switch to offer ISDN. R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL (June 20, 1997). That is the cost of upgrading the software in the switch; some central offices with older switches need to replace the switch altogether to be able to offer ISDN. See J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global *19 (Mar. 13, 1997).


70.     See J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global Table 3 (Mar. 13, 1997) ("key hurdles" for ISDN include "inadequate for multimedia apps (videoconferencing)."); R. M. Maclellan, Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL *7 (June 20, 1997) (stating that "decent videoconferencing" requires "at least 128-384 kb/s"); id. ("Digital TV requires at least 1.5 Mb/s downstream (using MPEG compression) with HDTVs requiring will above that.").


71.     See R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL (June 20, 1997) ("[A]n ISDN connection is really no different than one for a basic POTS line or a modem-based data call. Each ISDN user chews up a dedicated line card when on line and utilizes bandwidth on the trunks leaving the CO switch.").


72.     It has been estimated that there are currently about 1.1 million ISDN lines in service. About 80,000 of these are PRI ISDN lines. Residential ISDN lines now comprise a very small portion of the total installed base, at about 12 percent. T. Rudisill, et al., Raymond James & Associates, Ind. Rpt. No. 1816648, Local Access Market (Nov. 21, 1996).


73.     Because the duration of an ISDN call is directly proportional to the strain it places on the voice network, ISDN pricing is usage-based, which is difficult for many residential users to accept. See J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global (Mar. 13, 1997) ("[T]he cost to heavy end users can reach or exceed $100 per month, not very economical, especially in the face of competition from cable operators which promise much greater bandwidth at a cheaper price.").


74.     See R. Kahan, ISDN: "Not Dead Yet", Teleconnect, Apr. 1997, at 115; Bell Atlantic Press Release, Bell Atlantic Cuts ISDN Prices Again, July 24, 1997.


75.     Today cable networks passes about 95 percent of U.S. households, two thirds of which subscribe. National Cable Television Association, Cable Television Developments 1 (Fall 1997).


76.     Third Annual Report, Annual Assessment of the Status of Competition in the Market for the Delivery of Video Programming, CS Dkt. No. 96-133, paras. 16-17 (F.C.C.1997).


77.     T. Arnold and P. Hyzer, Building New On-Ramps for the Information Superhighway: A Look at New Cable Modem and ADSL Technologies, Dec. 13, 1996,


78.     Cox Communications, for example, is investing nearly $4 billion to upgrade its networks. See Duesterberg and Pitsch, p. 25.


79.     Industry analysts report that HFC network architecture currently exists in approximately 35 percent of all cable systems, and that over one-third of all cable subscribers are served by systems employing HFC architecture. Third Annual Report, Annual Assessment of the Status of Competition in the Market for the Deliver of Video Programming, CS Dkt No. 96-133, para 172 (F.C.C. 1997). See also D. Shapiro, et al., Deutsche Morgan Grenfell Inc., Ind. Rpt. No. 1964154, Modems *3 (Aug. 27, 1997) ("Although it is expensive to complete the hybrid-fiber coax (HFC) rebuild or upgrade that is generally a precursor to deploying a two-way cable modem service, what is often overlooked is that several operators have been upgrading their networks diligently for the past three, four, and five years, and a great deal of this money has already been spent."). Allied Business Intelligence estimates that 69 percent of the cable industry's total plant mileage will be within HFC networks by 2000. Allied Business Intelligence Press Release, Morgan Stanley, however, does not see this happening, even after 10 years. See Meeker, Morgan Stanley, Dean Witter, Co. Rpt. No. 2578415, @Home *4 (Aug. 20, 1997) ("Some individual MSOs will likely exceed the industry averages, but it is important to note that one-third of the US is served by smaller private cable companies that are for the most part capital-constrained. Thus, even after ten years, it is unlikely that more than 60% of the US will be served by 2-way HFC.").


80.     The cost of upgrading to HFC is approximately $200 per home passed, and only an additional $25 to add bi-directional capabilities. K. Maxwell, ADSL Forum, Cable Modems and ADSL, See also Shapiro, et al., Deutsche Morgan Grenfell Inc., Ind. Rpt. No. 1964154, Modems - Industry Report *8 (Aug. 27, 1997) ("After the upgrade is complete, operators must then place reverse modules in their amplifiers to activate the return path, but at an estimated $10-$15 per home passed, this process is relatively trivial compared to the time and expense (about $175-$225 per home passed) of the initial upgrade or rebuild."). There are however, difficulties with return path operation, and even in the absence of noise issues (return path noise funelling) the costs for an HFC architecture which provides guaranteed bandwidth to subscribers at a high penetration rate can be higher than for switched digital infrastructures including FTTC or VDSL. See C. Eldering, N. Himayat and F.M. Gardner, "CATV return path characterization for reliable communications," IEEE Communications Magazine, vol. 33, no. 8, pp. 62-69 (August 1995) and N. Omoigui, M. Sirbu, C. Eldering, and N. Himayat, "Comparing Integrated Broadband Architectures from an Economic and Public Policy Perspective," in The Internet and Telecommunications Policy Research, G.W. Brock and G.L. Rosston, eds. (Lawrence Erlbaum Associates, Mahwah, NJ, 1996).


81.     Unlike ADSL and ISDN, which offer broadband service using the existing infrastructure, cable modems require the upgrade of present uni-directional, coaxial cable-based networks to HFC. Cable modems run data from the user's coaxial cable into a standard port in the user's computer. One television channel (in the 50-750 MHz range) is allocated for downstream data, and another channel (in the heretofore unused 5-42 MHz band) is reserved for upstream data. The downstream channel offers speeds of up to 36 Mb/s. Upstream throughput is estimated at 500 kb/s to 10 Mb/s. It is estimated that individual users will experience access speeds of 500 kb/s to 1 Mb/s, but possibly more, depending on the transmission technology used. In addition to significantly higher throughput rates, cable modem subscribers benefit from almost zero latency: like a LAN, cable modem users are constantly connected to the network, so there are no dial-up procedures or busy signals.


82.     Kim Maxwell, Cable Modems and ADSL, ADSL Forum,


83.     It is estimated that cable modem users will have to share bandwidth with between 500 to 5,000 other data subscribers. Cable Datacom News, Overview of Cable Modem Technology and Services, See also R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL (June 20, 1997).


84.     These analysts feel that most implementations of HFC take into account the effects of multiple users accessing the shared bandwidth and that the addition of more equipment at the head-end will solve any demand problems that would arise. See D. Shapiro, et al., Deutsche Morgan Grenfell Inc., Ind. Rpt. No. 1964154, Modems - Industry Report (Aug. 27, 1997) ("[A]n operator with a recently upgraded system would likely have plenty of capacity to devote two (yielding 60 Mb/s), three (yielding 90 Mb/s), or even four (yielding 120 Mb/s) 6MHz channels if demand proved to be sufficient. Moreover, the scalability of the architecture provides an even more robust solution. A node in a modern HFC plant contains not only two lit fibers (one downstream and one upstream), but also several dark fiber pairs. As a result, an operator could put another laser at the headend and another optical node unit (ONU) at the termination of one of the dark fiber pairs and effectively split the node in two: instead of 750 MHz serving one 1,000 home nodes, there would now be 750 MHz serving each of two 500 home nodes, thereby offering twice the capacity per home . . . With two more dark pairs left, it could be done yet again and again, if necessary, to create four 250 home nodes with 750 MHz of capacity each."). Cable modem defenders also contend that the "bursty" usage patterns of the Internet (and, presumably, other broadband applications) will result in very few instances of all users needing the entire bandwidth at the same time. See ibid. ("[T]he likelihood of all modem subscribers logging on simultaneously is slight, and owing to the bursty nature of data traffic, the probability that all on-line users will demand data at the same time is far slighter.").


85.     More than 4.5 million homes have been upgraded to receive cable modem service. Jeff Peline, Cable Modems Fight for Lead, Cnet,


86.     Ira Brodsky, Cable Modems to the Rescue?, Telephony, June 23, 1997. Bresnan Testing Cable Modems in Marquette, Michigan, Broadband Bob Report, Apr. 7, 1997, Most customers are also charged a one-time installation fee that ranges from $99 to $175. Sun Country Cable Chooses Internet Ventures' PeRKinet Service, Broadband Bob Report, June 19, 1997,; Comcast Launches @Home Service in Middlesex County, N.J., Broadband Bob Report, May 6, 1997,


87.     David Bowermaster, "Cable Modems Outpace ADSL," ZDNet (July 31, 1998), at 0,3440,2125359,00.html.


88.     Press Release: Road Runner/Mediaone® Express Reach Customer Milestone, April 15, 1998, at The press released claimed 75,000 subscribers for these two services. Contrast this with a November 1997 report that only 60,000 subscribers had signed up on all cable modem services combined. P. Farhi, Slow Start for a Fast Connection, Washington Post, Tuesday December 23, 1997.


89.     Bill Pietrucha, Cable Modem Buildout Slower Than Expected, Newsbytes, Mar. 28, 1997. The Data Analysis Group predicts that 8 million cable modems will be sold in the US in 2001. Forecast Market for Cable Modems in Users Homes, US, Computer Industry Forecasts, Jan. 15, 1997, citing Cable Modems Bump into Network Roadblocks, Lightwave, Sept. 1, 1996 at 7.


90.     Cable Datacom News, Cable Modem Market Stats and Projections, http://cabledatacomnews. com/cmic16.htm.


91.     Analog modems convert the digital information generated by computers into a format that is identical to the analog voice signals produced by a telephone. At the central office, analog signals are received by another modem that digitizes the signal in order to send it through a digital switch. Approximately one-quarter of U.S. households have a personal computer (PC) equipped with an analog modem. According to Jupiter Communications, at the end of 1996 nearly 40 million households in the U.S. have PCs, for a penetration rate of 39 percent. Over 70 percent – 27 million – of these households have modems connected to their PCs, and over half of them – 15 million – use the Internet or other online services. Jupiter Communications Press Release, New Devices and Technologies Will Drive Net Into 36 Million Homes by 2000, Jan. 6, 1997.


92.     Note that 54 percent of households connecting to the Internet use 28.8 or 33.6 kb/s modems, 23 percent 56 kb/s, and 15 percent 14.4 kb/s or slower. The remainder of households (8 percent total) divide between ISDN, cable modems, ADSL, and satellite or wireless. Jupiter Communications Press Release, 56 Kb/s "Midband" Solution Will Dominate Home Internet Access, Oct. 17, 1996.


93.    We note that 56 kb/s is the practical limit for 56k modems. One observer estimates that between 25 and 40 percent of telephone lines will not support any connection speed faster than 33.6 kb/s, no matter what modem is connected to the line. Les Freed, Fast Connections for All?, PC Magazine, Oct. 21, 1996, at 83 (noting that lines connected to older "subscriber line concentrators" cannot transmit faster than 33.6 kb/s). Further, even with properly conditioned lines, 56k modems achieve speeds near 56 kb/s "only under laboratory conditions. Real-world speeds will average between 40 and 50 kb/s." Id. Standards disputes have also slowed the deployment of 56k modems. U.S Robotics (recently merged with 3Com) builds such modems to its own "x2" standard, while Rockwell and Lucent build to the incompatible AK56flex@ standard. The International Telecommunications Union is expected to ratify a standard for 56 kb/s modems this year, but there is no guarantee that current modems will be compatible with the standard. Id.


94.     Unshielded twisted pair copper wire has a current maximum bandwidth of 100 MHz, depending on wire gauge, insulation, and length of the wire. In the 1990s the Electronics Industry Association and Telecommunications Industry Association (EIA/TIA) developed a set of standards to define the maximum bandwidth of certain types of copper wire. See American National Standards Institute/Telecommunications Industry Association/Electronics Industry Association, Standard 568 A 95, Commercial Building Telecommunications Cabling Standard (Oct. 1995). The highest bandwidth is provided Category 5, capable of transmitting 100 MHz over 100 meters. Category 5 is the current standard for running high-speed local area networks within offices. Copper telephone outside plant is equivalent to Category 2, supporting 1 MHz of bandwidth.


95.     The industry refers to the alphabet-soup of DSL technologies, described below, by the generic acronym xDSL.


96.     See J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global *19 (Mar. 13, 1997) ("[S]ervice providers would need to add data communications equipment, such as Ethernet hubs and routers, rather than the far more expensive central office switching equipment.").


97.     High data rate Digital Subscriber Line (HDSL) uses two lines and achieves rates of 1.544 Mb/s, equivalent to a T1 trunk. Single line DSL (SDSL) is similar to HDSL but uses only one line. SDSL can achieve the same throughput as HDSL with half the lines, but at shorter distances – 10,000 feet compared to 12,000 feet for HDSL. Very high data rate Digital Subscriber Line (VDSL) is used for the very shortest distances, and can achieve speeds of 13 Mb/s under 4,000 feet and up to 52 Mb/s at 1,000 feet. See generally ADSL Forum, General Introduction to Copper Access Technologies, general_tutorial.html; ADSL Forum, ADSL Tutorial,


98.     All the major U.S. local telephone companies are well into ADSL technical and market trials, and most expect widespread deployment of ADSL in their regions during 1998. GTE, Pacific Bell and U S West recently launched commercial ADSL service in certain cities in their regions, and Bell Atlantic and Bell South are in the process of doing so. See ADSL Forum, ADSL Trials and Service Deployments, Some analysts fear that telcos might resist rapid ADSL deployment because they view it as a threat to their profit-producing installed base of T1 and other leased lines to business customers. See D. Shapiro, et al., Deutsche Morgan Grenfell Inc., Ind. Rpt. No. 1964154, Modems - Industry Report (Aug. 27, 1997) ("RBOCs are afraid of cannibalizing their lucrative T1 business by undercutting it with cheaper ADSL pricing"); Everen Securities, Inc., Ind. Rpt. No. 1926482, Convergence of LAN, WAN, & Internet (July 23, 1997) ("[B]efore xDSL will be widely deployed,] telcos must overcome the fear that those technologies will erode their bread-and-butter leased-line access business.").


99.     See T. Rudisill, et al., Raymond James & Associates, Ind. Rpt. No. 1816648, Local Access Market (Nov. 21, 1996) ("HDSL will continue to be driven by strong demand for 1.5 Mb/s digital service from the Internet Service Providers (ISPs), the transition from old repeatered T-1 lines to HDSL standards in the commercial market, and a move for fractional T-1 and full-duplex HDSL into the residential community."); Merrill Lynch Capital Markets, Ind. Rpt. No. 1910047, ADC Telecommunications (June 17, 1997) ("The HDSL market has begun to take off. Industry sources indicate about 488,000 lines will be deployed in 1997. That's almost 100% growth over 1996. Looking forward, very healthy growth rates are expected to continue.").


100.  AT&T/Lucent developed one standard, Carrierless Amplitude/Phase Modulation (CAP). As of December 1996, 90 percent of all deployed ADSL hardware was based on that standard, and "most trials have used CAP technology." Alan Stewart, The Battle for Bandwidth, Communications News, May 1997, at 36. But several standards bodies favor a different one, Discrete Multitone (DMT). ADSL, Edge, May 19, 1997; Anne Knowles, Incompatible ADSL Standards Duke It Out, InfoWorld, Dec. 23/30, 1996, at TW1. As independent observers note, "a standards war may prevent you from boarding this data communications bullet train." James Powell, Speedy ADSL Slow to Arrive, Windows Magazine, Aug. 1997, at 238.


101.  The addition of a digital remote terminal can help matters somewhat, by extending fiber closer to the home and reducing copper loop lengths below 12,000 feet. See generally ADSL Forum, General Introduction to Copper Access Technologies, general_tutorial.html; ADSL Forum, ADSL Tutorial,; Reohr Dot Com, Intelligence, ADSL: Turning Copper Into Gold,


102.  To achieve its maximum speeds, ADSL requires nearly the entire available bandwidth (about 1 MHz) of copper; transmission speeds slow quickly as line noise and signal loss (attenuation) rise. For long loops, telephone companies have installed loading coils that improve the signal quality within the 4 kHz used by analog devices, but reduce the available total bandwidth. Thus, while an 8.44 Mb/s downstream rate can ideally be achieved if the customers premises are less than 9,000 feet from the telco's switch, this rate goes down drastically if the distance increases. At a distance of 18,000, the downstream rate goes down to about 1.5 Mb/s. ADSL is not feasible if the distance is beyond 18,000 feet. Reohr Dot Com, Intelligence, ADSL: Turning Copper Into Gold,


103.  Even shorter loops can be bundled with many other loops when they leave the central office; splicing individual loops out of the bundle degrades their condition and increases line noise, reducing the maximum speed of ADSL. Reohr Dot Com, Intelligence, ADSL: Turning Copper Into Gold,


104.  See D. Shapiro, et al., Deutsche Morgan Grenfell Inc., Ind. Rpt. No. 1964154, Modems *6 (Aug. 27, 1997) ("Equipment costs remain high, at about $1,000 per modem pair:for ADSL, a modem is necessary both in the user's home and at the central office."); J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global *21 (Mar. 13, 1997) ("One of the key barriers to ramping up quickly is the cost per subscriber to install ADSL. Currently, costs are around $1,500-3,000 per subscriber."); R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL *30 (June 20, 1997) ("The single biggest delaying factor in the DSL network rollout is the current, but dropping, per-subscriber cost of the equipment. Current DSL modems carry an average price tag of $2,000-3,000, and the price is dropping.").


105.  See J.J. Bellace, et al., Merrill Lynch Capital Markets, Ind. Rpt. No. 1869480, The ABC's of Wireline Equipment: Global *21 (Mar. 13, 1997) ("Industry sources estimate these costs would have to drop to around $500 per subscriber or below for ADSL to become profitable for deployment. The cost is likely to come down once service providers begin buying large amounts of modems and equipment."); R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL *30 (June 20, 1997) ("While [$2,000-3,000 for an ADSL modem is] fine for tests, it is generally considered that the 'sweet spot' where DSL service becomes commercially viable on a wide scale is $500/unit. While IDC is projecting this price inflection point to be crossed in 1999, we think this is overly conservative."). Most analysts predict that agree that ADSL will not be commercially viable until the price drops to $500.




107.  T. Hanrahan, The Bandwidth Oasis, Wall Street Journal Interactive Edition, December 15, 1997,


108.  See R.M. Maclellan, et al., Dillon, Read & Co., Ind. Rpt. No. 1924563, Introduction to xDSL (June 20, 1997); id. at Tables 8 (showing residential ADSL deployment to be 1.52 million, or 3.5 percent of the residential PC/modem base, by 2000); id. at Table 9 (showing small-business ADSL deployment to be 290,000, or 5.3 percent of the small-business PC/modem base, by 2000).


109.  See Bowermaster, Cable Modems Outpace ADSL.


110.  Jill Abeshouse Stern, Towering Above Us, New Jersey Law Journal, Dec. 16, 1996, at 32.


111.  New Paradigm Resources Group and Connecticut Research, 1997 Annual Report on Local Telecommunications Competition 552 (8th ed.1996). WinStar bills Wireless Fiber as "the functional equivalent of fiber in terms of reliability, data transmission quality, and bandwidth provided to the end user." WinStar and U S West Sign Interconnect Agreement for Colorado, Edge, Apr. 7, 1997.


112.  NextWave Telecom and Sprint PCS may enter using their PCS spectrum in a manner similar to AT&T. Advanced Radio Telecom, which holds 38 GHz licenses in 47 markets nationwide is rolling out wireless local loop services. Teleport Communications Group recently purchased BizTel Communications, which has 38 GHz licenses in 48 states. In February 1997, AT&T introduced a new fixed wireless system that will provide households with the equivalent of an ISDN line, with two voice lines and 128 kb/s data speeds. The technology is intended to provide high-speed, high-quality, secured bandwidth to provide faster data services and full-motion video conferencing. AT&T is currently trialing the system in Chicago, and plans to deploy it commercially in 1998.


113.  LMDS systems consist of a multicell configuration distribution system with return path capability within the assigned spectrum. Generally, each cell will contain a centrally located transmitter (hub), multiple receivers or transceivers, and point-to-point links interconnecting the cell with a central processing center and/or other cells. FCC, Wireless Telecommunications Bureau, LMDS Fact Sheet, Sept. 5, 1997,


114.  VIPC, Inc., LMDS: Answers to Frequently Asked Questions, Oct. 18, 1997, The FCC's decision to allow 1150 MHz of spectrum to be auctioned to a single operator in each BTA (with a much smaller 150 MHz available as well) allows for unparalleled services to be offered, now and well into the future. US Wavelink, What Is LMDS?, Sept. 18, 1997, what_is_lmds.html.


115.  Hewlett-Packard envisions enough bandwidth for "most homes in a neighborhood to watch separate digital movies, teleconference, and surf the Internet at high speed all at the same time." VIPC, Inc., LMDS: Answers to Frequently Asked Questions, Oct. 18, 1997, LMDS is capable of providing a variety of one-way and two-way broadband services. Because of its numerous applications, LMDS has the potential to become a major competitor to local exchange and cable television services. FCC, Wireless Telecommunications Bureau, LMDS Fact Sheet, Sept. 5, 1997,


116.  A business case was compiled which evaluated an LMDS deployment in the Santa Clara Valley of California. The primary uses assumed in this study were work at home and high speed Internet access. The scenario assumes 7 business cells and 22 residential cells, each assumed to be 4km x 4km. There are an estimated 13,150 homes per cell. The cost to obtain LMDS spectrum is estimated to be $16 per household covered. Ten percent of covered homes are assumed to have signed up by the third year for service costing $150 per month (which is expected to be paid by employers), 60 percent of major employers are also assumed to have signed up by the third year for service costing $3000 per month. Customer premise equipment is estimated at $650 per home (includes a roof mounted transceiver, a downconverter, and an Ethernet adapter). The cost of the LMDS hub is expected to be $370,000 and interconnection charges are expected to cost $10,000 per hub. VIPC, Inc., LMDS: Answers to Frequently Asked Questions, Oct. 18, 1997,


117.  CellularVision was awarded a pioneer's license in Brighton Beach area of New York for its role in developing LMDS. CellularVision provides over 40 analog channels of video programming to its 12,000 subscribers, and is currently expanding the number of operating cells in the New York area. VIPC, Inc., LMDS: Answers to Frequently Asked Questions, Oct. 18, 1997, LMDS enables customers to bypass over-burdened phone lines to access the Internet at 500 kb/s, at four times the speed of an ISDN line (128 kb/s) and 20 times the speed of a 28.8 kb/s modem. Jason Daponte, CellularVision Offers Wireless High Service, Total Telecom, June 23, 1997, CellularVision of New York has begun marketing CVDN 500 in Manhattan and Brooklyn, where it offers a 49-channel subscription television service. Ibid.


118.  Total Telecom, U.S. Operators on Blocks for Wireless LMDS, Aug., 11, 1997,


119.  Gerry Blackwell, Wireless Access Enters Real-World Trials, Internet World, May 1997, at 15.


120.  MMDS operates in the 2 GHz band of radio spectrum and has a range of 35 miles. Although operators are currently licensed only for one-way, the FCC recently granted a license to MMDS operator CAI Wireless systems to offer two-way services in Boston. Various operators have already begun limited deployment of MMDS for high speed Internet access service under this arrangement.


121.  Wireless Data Communications: An Overview, The Choices, The Considerations, Motorola Inc., Wireless Data Group, available at


122.  A new sense of what modems could be, Online U.S. News, (April 6, 1998), and from


123.  DBS providers began delivering on-line services and the Internet to their customers beginning in 1996. Steve Higgins, Direct Broadcast TV May Go Further Than Many Predicted, Investor's Business Daily, Nov. 16, 1995, at A8.


124.  Subscribership is growing 100 percent annually. D.H. Leibowitz, et al, Donaldson, Lufkin & Jenrette Securities, Ind. Rpt. 2541409, Media and Entertainment (Mar. 24, 1997).


125.  Teledesic promises to build the "Internet in the sky". Teledesic's current plan is to deploy 288 LEO satellites, which is significantly less ambitious than it original plan to deploy 840. The company plans to launch its first satellite sometime in 2001. When it actually begins service (expected in 2002), Teledesic states that "from day one" it will be able to offer "fiber-like . . . broadband telecommunications access for businesses, schools and individuals everywhere on the planet." Teledesic's network will operate in the Ka-band (28.6 - 29.1 GHz uplink and 18.8 - 19.3 GHz downlink), and will offer most users a 64 Mb/s downstream connection and a 2 Mb/s upstream connection. Users with "broadband terminals" will enjoy a two-way 64 Mb/s connection. The terminals can interface with IP, ISDN, ATM, and other network protocols.


126.  In June 1997, Motorola announced the formation of Celestri, a $13 billion broadband LEO satellite network that is expected to begin providing service in 2002. Unlike Motorola's other LEO network, Iridium, Celestri is intended to serve as a broadband Internet backbone.


127.  Behind Teledesic and Celestri, the most ambitious broadband LEO project is Skybridge, which is being backed by Alcatel and Loral. The $3.5 billion Skybridge proposal calls for 64 LEOs to be launched in time for operations to begin in 2001. Skybridge would provide upstream connections of between 16 kb/s and 2 Mb/s, and downstream connections of 16 kb/s to 60 Mb/s. The remaining broadband satellite proposals call for geostationary satellites instead of LEOs. Loral is backing a $1 billion project called Cyberstar. It relies on a yet to-be-determined number of geostationary satellites. It would provide data and video services at between 400 kb/s and 30 Mb/s. Lockheed is backing Astrolink, which relies on 9 geostationary satellites to provide up to 9.6 Mb/s connections. The $4 billion Astrolink project is scheduled to begin operations in late 2000. Finally, GM-Hughes is backing the $3.5 billion Spaceway project, which would (initially) use 8 geostationary satellites to provide 6 Mb/s connections.


128.  Iridium was conceptualized in 1987 by Motorola engineers as a LEO-based, wireless personal communications network "designed to permit any type of transmission – voice, data, fax, or paging – to reach its destination anywhere on earth." Iridium, The Iridium System, systm/systm.html. In 1992, the FCC granted Iridium an experimental license; an operational license was granted in 1994. Motorola is Iridium's primary investor and contractor. Other major investors include Lockheed Martin, Raytheon, and Sprint. Iridium plans to launch 66 satellites total. Half of Iridium's constellation was in orbit by the end of September 1997. Service is expected to commence in 1998. The Iridium system interacts with end-users over the L-band, and the satellites, gateways, and earth stations interact over the Ka-band.


129.  D.H. Leibowitz, et al., Donaldson, Lufkin & Jenrette, Ind. Rpt No. 2502083, Wireless Communications Industry: Global, at 53 (June 5, 1996); Denise Pappalardo, Satellites Ready for Data Service Launch, Network World, Mar. 24, 1997 at 12. American Mobile Satellite has also launched a satellite service in the United States to offer fixed telephony and data services to areas unserved by the public network. A report by the Federal Aviation Administration predicts that up to four big LEOs and three little LEOs will be deployed between 1996 and 2005. GlobalStar has publicly projected subscriber counts of 30 million mobile voice and data users by 2010, See also the recent news on Iridium, Iridium delays launch of global system, in Yahoo!Finance,


130.  See Peter Huber, Law and Disorder in Cyberspace (New York: Oxford University Press, 1997), especially Chapter 5.


131.  Ibid, pp. 61-2.


132.  For an excellent discussion, see Thomas J. Duesterberg and Peter K. Pitsch, The Role of Competition and Regulation in Today's Cable TV Market (Washington, DC: The Hudson Institute, 1998), pp. 5-10.


133.  See Duesterberg and Pitsch, pp. 6-7, supra.


134.  See Barbara Esbin, Internet Over Cable: Defining the Future in Terms of the Past, OPP Working Paper No. 30 (August 1998).


135.  47 U.S.C. '214(a).


136.  47 U.S.C. '201.


137.  47 U.S.C. '541.


138.  47 U.S.C. '543.


139.  47 U.S.C. '22(6).   


140.  See Duesterberg and Pitsch. See also Peter K. Pitsch and David Murray, "Are Telecom Mergers Anticompetitive: Lessons of the Cable-Telco Mergers of 1993," Future Insight 3.1 (Washington, DC: The Progress & Freedom Foundation, June 1996).


141.  Indeed, the NPRM associated with this NOI is a perfect example. Reacting to petitions filed in early 1998, the Commission issued its NPRM in August, hoping (in a best case world) to issue a final rule in February 1999. Even if the Commission is able to meet this goal, litigation will almost surely preclude a "final" outcome at the Federal level before mid-1999. At that point, as the Commission has emphasized, further action by state PUCs would be required to implement the separate subsidiary approach promised in the NPRM. In many cases, states thus far have been reluctant to approve the actions that would be required under the NPRM's approach, and even those that do so may impose further requirements. In short, the NPRM – no matter how deregulatory its intent – sets in motion a chain of regulatory and legal processes that seems unlikely to produce a stable regulatory framework for digital broadband services offered by Incumbent LECs before the end of the millenium – if ever.


142.  See First Report and Order and Further Notice of Proposed Rulemaking at 127, Implementation of the Non-Accounting Safeguards of Sections 271 and 272 of the Communications Act of 1934, CC Dkt. No. 96-149 (rel. Dec. 24, 1996). The Commission did emphasize, however, that "[i]f a BOC's provision of an Internet or Internet access service (or for that matter, any information service) incorporates a bundled, in-region, interLATA transmission component provided by the BOC over its own facilities or through resale, that service may only be provided through a section 272 affiliate, after the BOC has received in-region interLATA authority under section 271." Id.


143.  Action, Common Carrier Bureau Will Not Enforce Current Rules on Application of Subscriber Line Charges to ISDN Service, 10 FCC Rcd 13473 (1995), reversing, In the Matter of NYNEX Telephone Companies, Revisions to Tariff F.C.C. No. 1, Transmittal No. 116, Memorandum Opinion and Order, 7 FCC Rcd 7938 (1992), affd. on recon., 10 FCC Rcd 2247 (1995).


144.  First Report & Order at & 116, Access Charge Reform, FCC 97-158, CC Dkt. No. 96-262 (F.C.C. rel. May 16, 1997).


145.  In the Matter of the Application of Bell Atlantic-Delaware, Inc., to Replace the Existing Tariff for Residential BRI Service, Dkt. No. 96-009T, Order No. 4246, 1996 Del. PSC Lexis 224, & 14 (Del. PSC Jul. 2, 1994) (noting with approval that hearing examiner had ordered a charge that was identical to that of the unlimited local usage package offered subscribers of residential dial tone.


146.  Bell Atlantic - Delaware, Inc. v. PSC, C.A. No. 96A-07-003, 1997 Del.Super. Lexis 113 (Super. Ct. for Kent County, Apr. 4, 1997).


147.  The FCC set the X-factor at 6.5 percent, even though historical productivity gains (the measure the Commission admittedly considers most reliable) have never showed productivity gains even approaching 6.5 percent. See Price Cap Order, para. 137 (highest five year average 6.1 percent); see also id., para. 139 ("there is no extended time period over which the measured [productivity gains] remained substantially above 6 percent.") According to the FCC's own studies, the most recent five-year average was a mere 5.2 percent, as was the overall average for the past ten years. Price Cap Order @ para.137. By the Commissions's own admission, the 6.5 percent offset is outside the range of reasonableness. The A lower bound of reasonableness, the Commission concluded, was 5.2 percent. The "upper bound" of reasonableness was around 6.1 percent based on the Commission's data, or "6.3 percent if the FCC included AT&T's admittedly inflated estimates in the analysis.


148.  Notice of Proposed Rulemaking, Amendments of Part 69 of the Commission's Rules relating to Enhanced Services Providers, 2 FCC Rcd 4305 (1987).


149.  47 U.S.C. '251. The Commission has recently proposed to extend its unbundling rules to permit competitive information service providers to obtain unbundled network elements. See Computer III Further Remand Proceedings, Further Notice of Proposed Rulemaking, CC Dkt. Nos. 95-20, 98-10 at  94-96 (released Jan. 30, 1998). 


150.  47 U.S.C. 251(d).


151.  Even then-Chairman of the FCC Reed Hundt noted that the unbundling rules do not "create economic incentives for the telephone companies that are proprietors of parts of the Internet, particularly the local loop . . . to upgrade those particular businesses." See FCC Drafting Rule Proposal to Address ILECs' Innovation Concerns, Communications Today, Sept. 17, 1997.


152.  See Illinois Public Telecommunications Ass'n v. FCC, 117 F.3d 555, 570 (1997) (since price caps went into effect, "investors rather than ratepayers have borne the risk of loss" on telephone company investments).


153.  47 USC ' 251(c)(4).


154.  See, e.g., 47 U.S.C. '531 (requiring cable operators to carry channels for public, educational or government use); 47 U.S.C. '532 (requiring carriage of channels for commercial use); 47 U.S.C. '535 (requiring carriage of local commercial television signals); 47 U.S.C. '535 (requiring carriage of noncommercial educational television).


155.  See 47 U.S.C. ' 571(a)(4).


156.  47 U.S.C. ' 534


157.  47 U.S.C. ' 532.


158.  47 U.S.C. ' 531.


159.  47 U.S.C. ' 535.


160.  See 47 U.S.C. ' 271(b)(1).


161.  The FCC has held that, for purposes of ' 271(b)(1), "interLATA services" includes "interLATA information services." Implementation of the Non-Accounting Safeguards of Sections 271 and 272 of the Communications Act of 1934, as Amended, First Report & Order & Further Notice of Proposed Rulemaking, FCC 96-149, & 55 (rel. Dec. 24, 1996).


162.  The most that Bell companies are permitted to do is provide Aa service that permits a customer that is located in one LATA to retrieve stored information from, or file information for storage in, information storage facilities of such company that are located in another LATA, 47 U.S.C. '271(g)(4).


163.  See Bell Atlantic Tel. Cos. Offer of Comparably Efficient Interconnection to Providers of Internet Access Services, 11 FCC R 6919 (1996).


164.  See 47 U.S.C. ' 272; First Report and Order and Further Notice of Proposed Rulemaking, Implementation of the Non-Accounting Safeguards of Sections 271 and 272, CC Docket No. 96-149, FCC 96-489 (rel. Dec. 24, 1996).


165.  See 47 CFR ' 64.1903.


166.  See Report and Order, Amendment of the Commission's Rules to Establish Competitive Service Safeguards for Local Exchange Carrier Provision of Commercial Mobile Radio Services, WT Dkt. No. 96-162 (released Oct. 3, 1997); 47 C.F.R. ' 20.20.


167.  The challenge now facing the Commission with respect to regulating data services offered by Title VI providers is discussed in length in the paper by Barbara Esbin, Internet Over Cable: Defining the Future in Terms of the Past, OPP Working Paper No. 30 (August 1998).


168.  NPRM paras. 154-155, 163.


169.  NRPM, Separate Statement of Commissioner Michael K. Powell.


170.  It is interesting to note that the invention of the mechanical telephone switch was motivated largely by economic, not technological reasons: Almon B. Strowger was a Kansas City, Missouri undertaker who found that all "new business" was being directed to his competitor. As it turns out, his competitor's wife was the telephone operator for the city and promptly routed all calls concerning new deaths to her husband. Strowager invented the telephone switch in 1889 to remedy this situation in by allowing subscribers to dial each other directly without operator intervention. The Strowger switch, also known as the two-motion, or step-by-step switch, was patented in 1881 and began to be deployed by the Bell System in 1918. As of 1978, 53% of the Bell System exchanges in service (serving over 23 million subscribers) used Strowger switching.


171.  See Daniel Minoli, Telecommunications Technology Handbook (Artech House, Boston, 1991).


172.  For a detailed discussion of networking standards and protocols see W. Stallings, Networking Standards (Addison-Wesley Publishing, Reading, Massachusetts, 1993).


173.  D.C. Comer, Internetworking with TCP/IP (Prentice Hall, Englewood Cliffs, NJ, 1995).


174.  M. de Prycker, Asynchronous Transfer Mode (Ellis Horwood, New York, 1993).


175.  P.P. White, "ATM switching and IP routing integration: the next stage in Internet evolution?," IEEE Communications Magazine, vol. 36, no. 4, pp. 79-83 (April 1998).


176.  There are numerous types of delay and echo which can occur. Packetization delay, the time required to fill a packet with voice samples from a subscriber, can be significant, and when combined with coupling between the speaker apparatus and microphone at the subscriber location can result in annoying echo. Echo cancellation technology can be used to mitigate these effects. Transmission delay can also become an issue in some networks, and if the delay is appreciable (e.g. 500 ms or longer) the one-way delay in itself (absent echo) is perceivable and annoying. For a further discussion of delay in the local loop see C.A. Eldering and A. Martin Minguez, "System specification and requirements for Fiber in The Loop systems," in SPIE vol. 1786, Fiber Networks for Voice, Video, and Multimedia Services, pp. 156-165 (1992). One-way delay and user perception of such delays is discussed in N. Kitawaki and K. Itoh, "Pure delay effects on speech quality in telecommunications, "IEEE J. on Selected Areas in Communications, vol. 9, no. 4, pp. 586-593 (May 1991).


177.  Based upon the OSI sever layer reference model, the packet based transport services described here can be considered to include the physical, data link, network, and transport layers. The upper three layers in the OSI model; session, presentation, and application, can be considered, for the purposes of this discussion to be part of the end-user application or service.


178.  The migration away from private services, such as those offered by America On Line and Compuserve, towards public services such as Internet web site access and Internet e-mail, clearly demonstrate the incentives global network externalities offer.


179.  Two-way cable systems are the best example of systems which can be incrementally upgraded to provide broadband services to a limited number of subscribers. A two-way cable plant based on Hybrid Fiber Coaxial (HFC) technology can support "cable modems" for Internet access, once a small amount of additional equipment is placed at the head-end, and modems distributed to the subscribers. However, if the penetration rates become significant (e.g. over 20%) substantial investment in the plant (to reduce the node size) and at the head-end (for multiplexing and encryption equipment) is required. Once the penetration rate rises above a certain level, it becomes more cost effective to have a telephone type "switched" infrastructure in place. For a discussion of the costs of HFC vs. switched digital infrastructures see N. Omoigui, M.A. Sirbu, C. Eldering and N. Himayat, "Comparing integrated broadband architectures from an economic and public policy perspective, " in The Internet and Telecommunications Policy (G. Brock and G. Rosston, eds., Lawrence Erlbaum Associates, Mahwah, NJ, 1995).


180.  47 U.S.C. 251


181.  For unbundled loops where competitors (typically entrants and incumbents) are simultaneously DSL technologies


182.  For a discussion of pricing regulation and the move towards price caps see R.W. Crandall and L. Waverman, Talk is Cheap (Brookings Institution, Washington D.C., 1995).


183.  CC Docket No. 96-112, Allocation of Costs Associated with Local Exchange Carrier Provision of Video Programming Services, was released on May 10, 1996 and raised cost allocation issues with respect to Open Video Systems, but no rulemaking has taken place to date.


184.  Id. paras. 58-60.


185.  As an example, with respect to allocation rules for apportioning the common fixed costs of service between different classes of customers, most economists agree that "common costs cannot be uniquely and nonarbitrarily allocated among customers and that the average costs that result from such an apportionment procedure, based on historical costs, are likely to result in incorrect prices (Id. p. 103).