Media Access Project: Internet User Coalition CommentsMedia Access Project: Internet User Coalition Comments
Return to Internet Access Charges Page
|
____________________________________________ Usage of the Public Switched |
) ) ) CC Docket No. 96-263 ) ) |
The Internet User Coalition ("IUC") and its member organizations,1 respectfully submit the following comments in response to the Commission's Notice of Inquiry ("NOI"), FCC No. 96-488 (released December 24 1966). The IUC appears on behalf of citizens who depend on the Internet for civic and social discourse, and seek to realization of its potential as a limitless source of information and entertainment, a powerful tool for education and research, and a platform for previously unattainable economic efficiencies.
The IUC has been formed to advocate federal policies which expand American citizens' Internet access. The Internet is a new and hitherto unimaginably widespread communications medium, the development of which can improve the quality of both speech and commerce.
The IUC's premise is that, in assessing how the nation's telecommunications infrastructure should accommodate the Internet, the FCC must adapt its policy and legal analysis beyond traditional telephony principles. In particular, it must ensure that its actions also advance the First Amendment goal of promoting the free flow of information and creation of a well-informed electorate. Application of this paradigm should compel the Commission to reject proposals to impose new charges on citizens seeking Internet access via the nation's switched telephone networks.
New "access charges" will impede the development of more efficient, high-bandwidth, open architecture data networks for the digital age. Those telephone companies claiming that such fees are needed to protect the functionality of the nation's telephone network use flawed arguments and unsupportable factual premises:
For more than sixty years, the FCC has separately regulated telecommunications carriers providing mass media or "common carrier" services. The long-awaited convergence of these functions is now well under way, in large part because of the sudden and unexpectedly explosive growth of the Internet.
The Internet resembles no other medium. It seamlessly facilitates both speech and commerce, so that the expansion of each of them assists the expansion of the other. In fashioning its policies, the FCC must recognize and accommodate this extraordinary attribute of the Internet. For the first time, its supervision of the nation's switched telephony infrastructure must also address First Amendment concerns it has not previously confronted only in its regulation of mass media. As it recognizes in mass media jurisprudence, the Commission should treat Internet users not just as "ratepayers," but also as citizens using these services for political speech, access to information, lifelong learning, communications, and commerce.
Because the issue of Internet access charges arises in the context of the Commission's revision of the nation's telephony infrastructure, its resolution necessarily implicates the existing framework of traditional common carrier regulation. These policies have focused less on the uses made with the network than on the network itself. They have emphasized the realization of reasonable carrier profits and efficacy of the public switched telephone network ("PSTN"), while assuring "just and reasonable" rates to consumers.
While the need to integrate mass media elements into its decision is less immediately obvious, it is no less important. The Internet and other information services have layered new, higher-order technologies on top of the existing local exchange carrier loop, private data backbones, and transmission protocols. These new technologies have begun to enrich the marketplace of ideas in the process. Technological convergence will recast what had heretofore been common carrier and private transmission services into new forms of media for political, civic, artistic, and commercial speech.
New forms of mass communication require new First Amendment applications, but the underlying goals remain the same. The Supreme Court has repeatedly ratified the First Amendment ideal that government should ensure the "widest possible dissemination of information from diverse and antagonistic sources." Associated Press v. United States, 326 U.S. 1, 20 (1945). It has held that "the people as a whole retain their interest in free speech...and their collective right to have the medium function consistently with the ends and purposes of the First Amendment." Red Lion Broadcasting v. FCC, 395 U.S. 367, 389 (1969).
The task, then, is to apply 18th-century constitutional principles to 21st century technologies. Professor Cass Sunstein has described the dilemma the Commission now faces in this way:
Sometimes constitutional doctrine seems to have lost sight of the point of central constitutional commitments. Sometimes the commitment to free speech seems like an abstraction insufficiently...connected with democratic goals, or indeed with any clearly describable set of governing aspirations.
Cass Sunstein, Words, Conduct, Caste, 60 U. Chi. L. Rev. 795, 797 (1993).
Professor Sunstein's response is that government must infuse its actions with policies which create an environment for democratic discourse. Building upon the writings of Alexander Meiklejohn, he emphasizes that "[o]ur constitutional system is one of deliberative democracy," and that government's role is to stimulate and nurture democracy. Government cannot restrict speech, but it can, and should, promote "attention to public issues,...[and]...diversity of view...." Sunstein, The First Amendment in Cyberspace, 105 Yale L.J. 1757, 1762 (1995). He concludes that
A well-functioning democracy requires a degree of citizen participation, which requires a degree of information; and large disparities in political (as opposed to economic) equality are damaging to democratic aspirations.
Id., at 1762-63 [footnotes omitted].
The IUC asks the Commission to follow this model. It should regard the creation of an environment for free expression and civic discourse, as well as electronic commerce, as a central object of its decisionmaking, not an afterthought.
The most important characteristic of the Internet may be that it brings freedom to the citizens using it - freedom from technological, regulatory, and geographic boundaries. The Commission's goal in framing Internet policies should be to ensure that every American can receive affordable access to the Internet, and that competition for provision of Internet service flourishes.
The Commission's NOI confronts some of the important issues of the decade. It will use comments filed here to review, inter alia: (1) how its rules can most effectively create incentives for the deployment of facilities that enable efficient transport of data traffic; and (2) what effects the current system of regulation2 may have on the PSTN, ILEC cost recovery, and the development of the information services marketplace. NOI at ¶¶313, 315. As it evaluates the record in this proceeding, the Commission must not lose sight of the fundamental principle that its policies should maximize citizens' freedom to use the Internet.
The Internet User Coalition ("IUC") is comprised of organizations and individuals that depend on the Internet for civic and social discourse, as a potentially limitless source of information and entertainment, as a powerful tool for education and research, and as an efficient platform for economic activities. These parties believe that the open, decentralized, and interactive nature of this medium makes it a unique and invaluable way to broaden access to political and cultural speech. The ability to access and share a wide array of data, text, and multimedia files gives the Internet a utility for education, research, and collaboration-at-a-distance that cannot be rivaled. Moreover, the Internet offers a potentially efficient means of marketing and distribution which will reduce frictional, transactional, and marketing costs, inuring to the benefit of economic participants on a global scale. To help realize this potential, and to expand citizen access, members of the coalition have endorsed seven underlying goals:
1. Digital Bandwidth for Internet Citizenship. The most significant bottleneck preventing individuals and small organizations from taking full advantage of the Internet today is inadequate bandwidth from the end-user to the ISP.3 The Internet will reach its full potential only when digital bandwidth at affordable prices is generally available. Therefore, the Commission's policies should encourage the development of a variety of high-bandwidth, digital access services for the home and small organizations.
2. Open, Decentralized Architecture. The Internet is becoming an important means to promote the free flow of information precisely because of its decentralized architecture and open standards. This allows all users to access the information network regardless of what type of computer, modem, or software they use. The Internet's decentralized nature provides users with the utmost freedom of speech, which permits them a wide variety of expressive activity in a wide variety of ways. Therefore, citizens must have multiple, competitive access options which are free from gatekeepers.
3. Universal Accessibility to an Interconnected, Global Network of Networks. The Internet will only reach its full potential when every citizen who wants to connect has the ability to do so. The same reasons to encourage that every American be connected to the public switched telephone network apply with equal force to interactive data networks. The Commission should adopt policies with an eye toward encouraging broad access, and preventing discrimination in deployment of interactive technologies.
4. Affordable Prices. Full participation on the Internet depends on affordable pricing. Indeed, pricing is closely related to, and a critical underlying element of, the goal of universal accessibility. Flat rate pricing plans, which result in fixed charges to the end-user, have been most effective at meeting user needs and encouraging greater usage. Changing the pricing of the end-user's "last mile" connection is unjustified; it has not been shown to be fairly allocable to actual net costs incurred, and may inhibit the deployment of data-friendly, higher-bandwidth services. Regulators will have a critical role to play in ensuring that basic telecommunications services enable affordable access, especially during the transition from monopoly to a competitive local telecommunications marketplace.
5. Consumer Choice. Competition will best provide a wide range of service options and prices. With competitive markets, including both "last mile" local transport services and information services, users will be able to select those services best tailored to their needs and budgets.
6. Non-discriminatory, Use-neutral Pricing. One of the unique features of the Internet is the degree of freedom it offers its users to apply the data they receive to whatever purpose they wish. At the present time, this includes publishing, sending, and receiving multimedia documents which can reach a worldwide audience with text, audio and even video messages. These uses are rapidly evolving, and new uses are being developed constantly, but the transmission medium remains the same whatever the use. It is this flexibility that allows the Internet to continue to offer its users innovative and efficient ways of accomplishing basic communications tasks which used to be more difficult and expensive. Therefore, users ought not to be subject to regulation or to price differentiation based on the ultimate use of the data they receive or transmit.
7. Community Involvement. All users of the Internet benefit when there is maximum participation from community institutions such as schools, libraries, and other local organizations. These organizations play a unique role in disseminating interactive media to future generations and to individuals who may not otherwise be able to access them. Therefore, Internet policies must be implemented in a way which promotes community involvement.
The Commission has asked several questions about how its rules can "most effectively create incentives for the deployment of services and facilities to allow more efficient transport of data traffic...." NOI at ¶313. Noting that ILECs have raised concerns about congestion in the PSTN, it asks whether there are technical solutions to these concerns, such as hardware routing data traffic around switches, new high bandwidth technologies such as ADSL, or wireless solutions. Id. Finally, it observes that the end-to-end dedicated channels created by the circuit switched PSTN are "unnecessary and even inefficient" for carrying packet-switched data traffic. Id.
While a myriad of complex and technical issues surround these questions, they are of paramount importance to Internet users. The transition to high bandwidth, data-friendly transport technologies is perhaps the most important challenge to the future accessibility and functional development of the Internet. Many of the technologies which could provide this greater bandwidth will, by their very nature, alleviate congestion on the PSTN.
The IUC believes that the best strategy to encourage this transition is the hands-off approach, together with aggresive efforts to promote competitive access to the local loop. It is dangerous to pick a particular technology or industry as a favorite in this race. Moreover, it is unnecessary; marketplace forces are already in line to ensure a higher bandwidth future. What the Commission must do is ensure that these competitive forces have the ability to operate. ILEC proposals to impose access charges on ISPs, however, would diminish these forces by giving ILECs incentives to continue to use the PSTN for ISP traffic.
The eventual migration of Internet traffic from the PSTN to more efficient data networks will benefit Internet users and PSTN users alike. From the standpoint of the Internet user, it is crucial that emerging technologies allow for: adequate bandwidth to allow easy, fast retrieval and transmission of information; competitive provision of local data transport; fair competition among ISPs; and transport provision free of bottlenecks.
It is also crucial that new technologies preserve the Internet's open, decentralized architecture. The open nature of the Internet's standards and protocols, which permits user access regardless of what equipment is used, is an essential factor in its development.4
Such technologies are in fact being developed. Comments being filed in this docket on behalf of the Internet Access Coalition and America Online will include a more detailed discussion of the technologies that are in development or currently available for carrying data traffic. These will include technologies that make more efficient use of existing wireline facilities and those that provide for facilities-based competition. The IUC, therefore, will not discuss each of these in detail. It will similarly leave it for other parties to discuss when these new transport platforms could be deployed, which of them would best suit ISP needs, and what regulatory relief would best provide an incentive for the deployment of these technologies.
Some ILECs have argued that it is necessary to impose access charges on Internet use because it is the only way to give ISPs the proper incentives to migrate from their current use of the PSTN to more data-friendly technology. According to this argument, one "consequence of today's pricing signal is to retard the adaptation of more appropriate technologies." Bell Atlantic Study at 6. See also, Letter from James Young, Vice President and General Counsel, Bell Atlantic, to Senator John McCain (February 28, 1997)("current policies provide little incentive for Internet providers to embrace...new technologies").
In fact, the very opposite may be true. The ILECs' argument misapplies the economic principle that it is desirable to place the costs of an activity on the party which is best able to control those costs. Actually, the best way to encourage innovation in the management of data traffic is to place the costs on ILECs themselves. They are, after all, the parties best able to control costs in the PSTN. A surefire way to discourage innovation is to let ILECs pass costs along to parties who are in no position to manage deployment of network resources. If ILECs can collect compensation for voice network access to ISPs, they will have a strong incentive to build more voice network facilities. There will be little incentive for them to invest in newer, data friendly networks.
The notion that new charges are necessary to motivate ISPs to improve their facilities is based on an inaccurate portrayal of the Internet industry, because it ignores the preferences of Internet users. It assumes that, much the same as the local telephony market, the information services market is not competitive. This misconception leads the ILECs to conclude that ISPs will fail to act unless there are regulatory incentives. But this is not the case; the ISP market is presently highly competitive. There is every indication that users who want more data-friendly networks and/or less congestion will demand it, and will vote with their feet. Indeed, CompuServe and Erol's quickly moved to capitalize on reports of congestion faced by America Online customers by advertising that they offer fewer busy signals. Therefore, an ISP already has a powerful, direct economic incentive to adopt more efficient service.
Moreover, the rapid development of technology, and recent growth in the number of Internet users, ensures that ISPs will not become entrenched with their existing systems. The ILECs' physical plant is designed to be in service for many years, and its modular nature allows additional equipment to be added to work alongside existing equipment. It would not make sense, therefore, for a typical ILEC to discard its investment in existing architecture. In contrast, computing technologies evolve constantly and at lightening speed, so ISPs will not face the same inertia. Moreover, Internet communications operate because they use a common protocol, and do not require all users to be operating the same equipment. An ISP can receive data easily on a 14.4 kbps modem, an ISDN modem, or an ADSL network, side-by-side without interference. Bell Atlantic's own study documents how ISPs have already embraced technological improvements. It finds that ISPs in its area have installed the more efficient primary rate interface ISDN ("PRI ISDN") on about 50% of their lines. Bell Atlantic Study at 6. Other experts have noted that they are eager to adopt newer transport methods. Jeff Caruso, "ISPs Drive ADSL Ahead," CommunicationsWeek, March 17, 1997.
When - and only when - digital, data-friendly networks are available, users will be willing to pay a reasonable, cost based price. The ongoing, enormous user demand for improved computer modems demonstrates this. In just a few years, the industry standard has gone from 2400 baud to 28.8 kbps, and now to 56 kbps. At the same time, the price for these high-speed modems has greatly declined to reflect costs. There is every reason to believe that similar consumer demand exists for data-friendly networks.
Users should not be made to pay until these data-friendly networks are available from ILECs, at cost based prices. As just noted, there is no guarantee that revenues generated from data transmission will ever be devoted to development of networks, or anything besides ILEC general funds.
The Commission asks several questions concerning the "effects of the current system on network usage, incumbent ILEC cost recovery, and the development of the information services market- place." NOI at ¶315. It asks commenters to provide data on the characteristics of information service usage and its effects on the network, including the "incumbent ILEC's costs directly related to ESP's use of the PSTN,...revenues attributable to ESP traffic (including second phone line revenue), and...a comparison of what PSTN services ESPs desire, as opposed to what they currently have access to." Id.
These questions are important, but their focus is far too narrow. In asking how to allocate costs between the industries, the NOI fails to explore adequately the needs of the citizens who use the Internet. The IUC urges the Commission to pay greater attention to the needs of Internet users - both institutional users and citizens. Given that the new fees proposed by the ILECs will be borne by these users, the Commission cannot overlook them.
The NOI contains many questions about the effects of the information services industry: about congestion that ISP traffic may cause on the PSTN, ILEC costs and revenues, and the ISPs' current and hoped-for use of the PSTN. NOI at ¶315. Nowhere, however, does it promise to consider what the effects might be on the parties most directly affected by the outcome of these issues -- the Internet's users.
As several members of the IUC have already observed, MAP, et al. NOPR Comments at 1-3, information services are more than another product; they are an evolving medium of speech that has assumed an important and growing role in the social, economic, and political life of the citizens using it. Imposing new fees will affect far more than the information services market. It is no exaggeration to state that it would impact the very nature of the modern democratic process - the freedom and ability of citizens to speak.
In making policy judgments on imposition of fees on Internet users, the Commission must treat the Internet as an arena of citizen speech and political discourse. The Internet is unique in that it offers all speakers a level playing field - it is the "most participatory form of speech that this country - and indeed the world - has yet seen." ACLU v Reno, 929 F.Supp. 824, 881 (E.D. Pa. 1996), probable jurisdiction noted, ___ U.S. ___ (December 9, 1996). The Internet is also an ever-expanding and boundless source of information and entertainment. "Citizens from all nations are finding additional outlets for personal and political expression, utilizing interactive fora...to voice their opinion and to listen to the views of others." Information Infrastructure Task Force, "A Framework for Global Electronic Commerce," (Draft #9, December 11, 1996) at 1. Finally, at the very same time it has the potential to enhance speech and personal freedom, it offers greatly improved efficiency of commerce, transcending distance and borders. "[T]he Internet has blossomed into an appliance of everyday life, a medium accessible form almost every point on the planet, brimming with an inestimable range of data and information. In essence, the Internet has become the vehicle of a new, global digital economy which has enveloped the physical world, altering traditional concepts of economic, political, and social relations." Id.
The current system of regulation, under which ISPs do not pay access charges, has been highly effective in promoting speech, retrieval of information, and commerce over the Internet. The affordable, flat rate pricing and open architecture it promotes has fostered a competitive, creative, and dynamic market for ISP services.
Applying access charges to ISPs could destroy this. ISPs facing new, per-minute access charges will have two alternatives: they can either absorb the new costs, or pass them through to users in the form of higher prices. Absorption of these costs will be difficult, since high levels of competition have already driven prices down to the point where profit margins are very thin. See, e.g., Wayne Rash, Jr., "Expensive Access Lessons Loom for Internet Users," CommunicationsWeek, January 13, 1997. The result is likely to be a widespread industry fallout, which in turn would lead to less competition, little innovation, and insufficient consumer choice. The IUC believes that, therefore, ISPs are far more likely to pass these new costs along to users.
The parties best able to absorb these costs are the ISPs owned and operated by the ILECs themselves. Backed by its parent's immense capital resources, an ILEC's ISP business unit will have an unfair advantage over unaffiliated or entrepreneurial competitors. If it wished, an ILEC's ISP unit could keep its rates low by not passing along the access fee to its subscribers, withstan ing any resulting losses for a longer period than its unaffiliated competitors. For the ILEC, the access charge payment is in the nature of an internal transfer, so that while the profits of the its ISP unit would be reduced if it did not pass along the increase, the net result would be a wash. Doing this, the ILEC's ISP would presumably draw market share from non-ILEC ISPs. The reduction in volume and profits could force many of these unaffiliated, entrepreneurial ISPs out of business. In the alternative, an ISP unit of an ILEC could raise prices anyway, enjoying larger profits per customer than competing ISPs. It could still afford to raise prices less than the full amount of the access charge, so as to still unfairly increase its market share.
Non-ILEC ISPs are more likely to raise prices so as to pass their new costs through to citizens; this could also be looked upon as a private tax on Internet use. As with any tax placed on an activity, the incidence of the activity will decrease. Indeed, high per-minute charges are the number one reason for lower Internet use in foreign countries, according to a recently-reported study by the Organization for Economic Cooperation and Development. Douglas Lavin, "High Access Costs For Net Depress Usage, Study Says," Wall Street Journal, March 14, 1997, at B3C. Moreover, new users may be discouraged from trying the Internet, which will have a disproportionately great impact on a medium which is still very much in development. Finally, this tax will diminish the uniquely egalitarian nature of online speech - lower income users, who may be most sensitive to fluctuations in price, could be foreclosed entirely. The economic cost of this tax, moreover, will be borne not only by the individuals whose speech would be silenced, but by all Americans, who would be deprived of the benefits of a widely used speech and commercial forum.
The Commission seeks comment on the effects of new applications for Internet-delivered data, such as "Internet telephony, [and] real-time streaming audio and video services..." NOI at ¶316 (footnote added). It asks how these new services should affect its analysis. Id.
In short, they should not. The Commission should decline to follow requests to regulate the Internet based on uses made of the data communicated over the network.5 No matter what software application they are running, Internet users transmit and receive the same thing over the network every time and all the time: packets of data. The Commission could not determine which citizens were using which applications at any given moment, no more than it could determine the subject of particular voice conversations being carried out on the PSTN.
Moreover, to devise a system of charges based on present day uses would be problematic. As noted above, the uses of those data packets are constantly and rapidly evolving. It would be nearly impossible to predict with certainty which specific uses might ultimately become stable and prevalent, especially within the time frame of agency rulemaking proceedings. Attempting to determine the uses made with every packet may well be a violation of user privacy, perhaps conflicting with the Electronic Information Privacy Act. 18 USC §§2510, et seq. And it would be unwise, because it would reduce the flexible, open nature of speech and activity which has driven the development of this medium. Indeed, the complexity added by a regulatory process of approval and classification of new uses may well stifle innovative forces. This could destroy the freedom from regulatory burdens which has enabled introduction of groundbreaking developments in such fields as democratic speech, multimedia presentations, educational applications, and telemedicine.
In the Telecommunications Act of 1996, Congress directed the Commission to adopt broad reforms to ensure that educational institutions and libraries had access to advanced telecommun cations services. 1996 Act, §254(h). To that end, the Federal-State Joint Board on Universal Service has recommended a 20-90% discount for all telecommunications services, Internet access, and internal connections. Recommendations of the Federal-State Joint Board on Universal Service, at ¶440. Its recommendations of general discounts on all telecommunications services, and discounted Internet access, were designed to further Congress' goal of "schools and libraries [having] access to the wealth of information available on the Internet, and, therefore,...to advanced telecommunications and information services...." Id. at ¶465.
Imposing access charges on ISPs would work at cross-purposes with these initiatives. Just like all other Internet users, schools and libraries would face increased costs to connect to their ISP, and would have to reduce their amount of use drastically. Thus, the Commission would be taking with one hand what it had given with another. The result would be a diminution of the enormous potential of the Internet as a tool for learning. It could altogether deprive those students who do not have home computers with Internet connections of any opportunity to use this medium.
In addition to potentially stifling the growth of speech and commerce which the Internet has experienced to date, imposition of access charges at this time would be unfair as a matter of regulatory policy. It would reverse a wise policy judgment, which has promoted the growth of ISPs, alarm monitoring, and many other services, and it would do so at a time when the reasons for the previous policy have gotten even stronger. Yet, given the uncertainty surrounding whether Internet traffic causes a net cost to ILECs, pages 22-34 below, reversing this policy would provide few, if any, benefits. Moreover, it would allow a degree of market segmentation which is so unprecedented in common carrier regulation that it would depart from the very core of common carrier principles.
In comments to the NOPR, and other public statements, many parties have argued that it is time to discontinue the exemption for "enhanced service providers," because it was a temporary policy designed for an infant industry which has grown up. America's Carriers Telecommunication Association, for example, states that the enhanced services industry "is no more or less an `infant' industry and is no more or less mired in a state of change and uncertainty than is the local/long distance [calling] industry." ACTA NOPR Comments at 26. Pacific Telesis asserts that "the Enhanced Services industry is now well developed." PacTel NOPR Comments at 76. See also, Dorman Speech at 6.
This is an inaccurate portrayal. The Commission's policy rationale for not imposing access charges on ISPs6 has always been based on the industry's stability, not its size, and it is as applicable today as it has ever been. The policy was originally created in the Commission's 1983 MTS and WATS order. Memorandum Opinion and Order, MTS and WATS Market Structure, 97 FCC 2d 682 (1983). There, the Commission realized that it had to minimize "customer impact" and "market displacement." Id. at 715. It found that if it were "to impose full carrier usage charges on enhanced service providers...[they] would experience huge increases in their costs of operation which could affect their viability." Id.
In 1988, the Commission revisited its decision and found that imposing access charges would still be "inappropriate," because the enhanced services industry was in a "uniquely complex period of transition." Order, Amendments of Part 69 of the Commission's Rules Relating to Enhanced Service Providers, 3 FCCRcd 2631, 2632 (1988). It found two transitional forces to be specifically compelling: the hoped-for provision of unbundled elements of network architecture, so as to allow enhanced service providers to make efficient, … la carte purchases of needed elements; and the impending entry by Bell Operating Companies as competitors in the enhanced services market. Id. at 2632-33.
The logic that the Commission followed in these previous orders applies with even greater force today. First and foremost, the unbundled elements of the local network will hopefully soon be provided via competitive pricing, thanks to the 1996 Act and the Commission's careful implementation thereof. ISPs already face uncertainty as to the ultimate pricing of the unbundled service elements and the technical effects to their operations. It does not make sense for the Commission to subject ISPs to paying access charges before there is real, effective competition. There is little equity in forcing these companies, and all their users, to shoulder the costs of non-competitively-provided network elements. Perhaps upon such time as competition is widespread for provision of these services, it will finally make sense to revisit the issue; now it does not.
Moreover, the state of flux in the market that concerned the Commission in 1983 and '88 has not changed - it has only increased. Since 1988, the ILECs have entered the market for enhanced services, but they have only recently begun to offer Internet service. Indeed, as noted above, pages 15-16, they stand to improve their competitive positions in these markets as a direct result of a repeal of the exemption. But this is not the only agent of radical change. The market has undergone well-publicized dilemmas involving new pricing structures, growth in the number of users, adoption of new types of uses and standards, content regulation, encryption, intellectual property, trade and taxation. Imposition of access fees, and the resulting decline in use, would only add to the disruption of an already volatile industry. Now, as in 1988, it would be inappropriate for the Commission to add to the confusion, especially in light of the doubts over whether ISP traffic really causes net costs to ILECs. See pages 22-34 below. The Commission should wait until the ISP market has stabilized.
As common carriers, ILECs must take all traffic on their communications platform, regardless of content, i.e. what is said on this platform. 47 USC §202. They can segment their market based on desired levels of service, such as distinguishing between residential or business users, or customers desiring additional services, such as call waiting. But ILECs are not presently permitted to segment the market based on what a particular user says or how long it takes him to say it.
That is exactly what they seek to do here. Access charges, if imposed, would treat ISPs differently from other end users7 which receive calls over the local loop, simply because there are allegedly heavier traffic patterns. In effect, this is a "popularity charge." Following this logic, higher charges would apply to families with teenagers, businesses with booming sales orders lines or customer inquiry lines, or any end user with a PBX system that is designed to make efficient use of resources. See page 28, below. ILECs should not be able to increase charges to an end user simply because that user receives more traffic than average.
Laudably, the Commission has indicated that it will be critical of the claims made by all parties to this proceeding. In the NOI, it has demanded greater certainty and specificity from the parties claiming that ISP traffic poses costs and dangers to the PSTN, or that new charges would threaten the health of the burgeoning information services industry. It has requested "better empirical data" to support such assertions. NOI at ¶311.
Faced with this request, the ILECs have simply failed to prove their case. Instead, they have sought to justify the imposition of new access charges by offering only anecdotal accounts of congestion, many that do not even apply to the PSTN. They have made vague and inflated estimates of the costs of Internet traffic. Finally, they ignore almost all revenue derived from Internet traffic. Especially in light of the potential for an enormous diminution in Internet speech and usage, the Commission should not act based on such suspect evidence.
A number of ILECs' have offered doomsday scenarios that ISP traffic has already begun to cause congestion in the PSTN, and that this problem will only increase in the future. As evidence, they offer little more than anecdotes. In many cases, they seize upon stories that actually have no relevance to PSTN traffic. For example, a Bell Atlantic press release cautions that ISP traffic may cause "disruption of vital public safety services such as 911 emergency call service." Report of Bell Atlantic on Internet Traffic, November 19, 1996 at 1. The president of Pacific Telesis, pointing to a single incidence of increased traffic in a single switch in the middle of Silicon Valley, claimed that there would be a "flood" of data traffic which would cause "drowning switches." "Telecom Deregulation and Internet," remarks of Dave Dorman at the Association for Corporate Growth, Los Angeles, January 15, 1997 (downloaded from http://www.pactel.com/about/mgnt_perspectives/dorm11597.html) at 6-7 ("Dorman Speech").
These stories are no substitute for accurate, empirical evidence of congestion. Indeed, ILECs use these tales to claim that ISP traffic will cause congestion in all parts of the network, even those geographically remote from the ISP's points of presence. This is an exaggeration of the problem; it is like citing instances of rush hour traffic at 20th and M streets to prove that traffic will be congested in the suburbs. Instead, Internet traffic is typically diffuse over most parts of the network, but concentrated in only one location.
Many of the most widely-publicized incidents of congestion have been internal to the Internet itself, and thus have no impact on the PSTN. Some ILECs have pointed to service outages on Netcom, AOL, and other ISPs, as proof that the PSTN itself is being clogged. "Internet Congestion: Crisis or Come On?," speech by Michael Fitzpatrick, PacTel President and CEO, Wescon/96, October 23, 1996. Congestion within the Internet is simply irrelevant to the PSTN congestion problem. The Internet is a wholly distinct, largely privately- operated network, which intersects the PSTN only at certain specific ISP points of presence. Although these stories show that the ISPs themselves face congestion problems, they occur on the private, data networks only.
It is always possible to concoct scenarios under which ISP traffic could create congestion on other network locations, but they are highly unlikely ever to occur. For example, since Internet user traffic is so diffuse in locations remote from the ISP, it might well be possible for traffic to spike so high as to block all paths between two central offices ("COs"). This is actually quite improbable, however, because interoffice trunks are laid out to ensure that there are redundant paths to complete a given call. ETI Study at 8. Moreover, if a sufficiently large number of users were served by a single CO switch, and all of them tried to contact their ISP at the same time, this traffic could block all paths through the single serving switch. But calls not going through this switch would be unaffected.
Another possible location for greater congestion is the business trunk lines between an ISP and its serving CO. But because these are individual end user lines, any congestion at this location in the PSTN would not impact other COs, or even other traffic through the ISP's serving CO.8 Of course, such traffic might impede the ability of the ISP's subscribers to place calls to that ISP. But marketplace forces will be a strong incentive to reduce this type of congestion: an ISP must either purchase more lines and modems to alleviate the congestion, or it will risk losing customers who are dissatisfied with getting busy signals.
The final likely target for congestion is the switch at the serving CO that terminates calls to the ISP. This might be the only location currently beyond the ability of the ISP to minimize. But it is far from clear that ISP traffic has caused problems at these at these serving CO switches. See discussion below at pages 28-29. Moreover, technological solutions are reducing the probability of congestion. For example, ETI found that ISP traffic using trunk port connections, so as to bypass the line concentration module at the CO, will not cause blocking. ETI Study at 30. Bell Atlantic notes that half of the ISPs in its territory use PRI ISDN, a method of trunk port connection that bypasses the line concentration module ("LCM"). Bell Atlantic Study at 2. Therefore, half the ISP traffic in Bell Atlantic's area could not possibly contribute to any congestion problem. Yet the ILEC studies inflate the problem by aggregating ISP traffic that bypasses the LCM with traffic going through the LCM.
The ILEC studies have failed to show how ISP traffic could possibly contribute to their costs, because the majority of that traffic is off-peak and therefore makes efficient use of spare capacity. It is generally accepted that "[t]he cost of operating the PSTN and many of its components is sensitive to the peak demand placed on each network resource and to the relationship between that peak demand and the aggregate capacity of the individual network components." ETI Study at 11.
ISP traffic is largely off-peak, however, as attested to by even the ILECs' own studies. Bell Atlantic Study at 2-3 (ISPs' busy hour 10:00 p.m.); Pacific Telesis Study at 2 (same). There should be little cost incurred by such off-peak use, or even significant growth of off-peak traffic. ETI Study at 11.
The ILEC studies, however, have avoided a direct comparison of peak hour demand on system components at 5:00 p.m. with the ISP hunt groups' hour of highest demand at 10:00 p.m., opting only to show that the duration per call is longer for ISPs. Moreover, no ILEC studies have shown that ISP traffic has caused the busy hour to shift. Even though ILECs claim that the PSTN is engineered for shorter calls, the number of these short calls placed at 5:00 p.m. would certainly be greater than at 10:00 p.m. Thus, the total call volume during the peak hour may still be far greater, and may still use much more switch capacity, than during the ISP hunt group busy hour.
An increase in off-peak use of the PSTN does not lead to the conclusion that ISP traffic increases the ILECs costs. This is equivalent to saying that more nighttime traffic on an interstate highway will cause more congestion during rush hour. The ILECs have not explained how off-peak traffic from Internet users would somehow produce costs that do not occur for other off-peak traffic. They have not shown that off-peak traffic causes greater wear on network components. Nor do they attempt to show that traffic at 10:00 at night interferes with the traffic from the majority of users during the business day.
In responding to the Commission's questions in the NOPR about Internet traffic, a number of commenters stated that ISPs are functionally identical to interexchange telephone carriers. See, e.g., Pacific Telesis NOPR Comments at 78; Southwestern Bell Telephone Company NOPR Comments at 19; ACTA NOPR Comments at 30-31.
This is a flawed analogy for several reasons, however. First and foremost, as noted above, the Internet is more than just a medium for interexchange communications. It has assumed a role as a speakers' platform, a potentially limitless source of information, and an important mode of commerce. As such, the Commission should not regulate Internet traffic rates as it would interexchange tariffs and access charges, but should also consider the effects of the rate structure on the speech and information access interests of Internet users.
ISPs have been classified as end users, buying many lines for the purpose of processing user traffic and transferring it to a private data network. See n. 2, above. Few would argue that other end users, such as large businesses or government agencies, should have to pay access charges, even if their PBXs received a large, steady volume of calls. Customer service lines, catalog sales lines, and local call in contests may also receive a large concentration of traffic, but they do not pay usage sensitive fees. Many end users pass traffic on to data networks, including bank-by-phone lines, credit verification networks, and ticket purchasing agencies.
Indeed, busy hour line usage may be as great for these other end users as for ISPs. An efficient, cost-minimizing telecommunications customer, such as a large business or government agency, would try to engineer its trunk line usage so that it purchased only as many lines as necessary to achieve a desired grade of service. Shared utilization of these trunks among the individual users during busy hours would lead to high concentrations of traffic, perhaps higher than seen on ISP trunks. For example, the ETI Study calculates that a PBX with 900 individual station lines would require 97 trunk lines, which would have a traffic volume of 28 CCS during the busy hour. See ETI Study at 18. Thus, in this case, the business might routinely generate a concentration of traffic even greater than the levels of ISP traffic that ILECs claim is congesting the network. The effect of access charges, therefore, would be to tax the ISPs' efficient use of trunk lines.
The ILEC studies are also noteworthy for the degree to which they rely on anecdotal evidence to extrapolate to costs and congestion on the PSTN as a whole. Put another way, the studies' sample selection is not representative of their ILECs' networks as a whole. Their unrepresentative sampling of a few extreme, isolated incidents of congestion grossly overestimates what would be found in a more statistically valid review.
For example, as the ETI Study has observed, Pacific Telesis operates 790 CO switches in the state of California. It nonetheless selected only 29 of those to study, and discusses only one of these as having service problems. That one switch is located in Silicon Valley, where Internet traffic is likely to be far above average. ETI Study at 20. Similarly, Bell Atlantic examined only 9 COs, or under 5000 out of the 20.1 million circuits in its territory. Even so, it could find supposed congestion at only 5 of these 9. Bell Atlantic Study at 2; "2nd Quarter 1996 Results" at 1 (downloaded from "http://www.bellatl.com/invest/ fin_info/q296/p01.htm").
This means that the Bell Atlantic study found congestion problems at only approximately 0.014% of the total access lines in its territory. Using similar analysis for the Pacific Telesis study, PacTel found congestion problems at only 0.00038% of its access lines.9 Compare Pacific Telesis Study at 1-2 with "Pacific Telesis Reports Strong Quarterly and Annual Growth," January 21, 1997, at 2 (downloaded from http://www.pactel.com/cgi-bin/getrel?1457).10
Another manner in which the ILECs advocating the imposition of new fees fall short of establishing their case is that they have greatly overestimated the costs Internet traffic causes to their networks.
Pacific Telesis has stated that it was forced to spend $14 million in 1996 to upgrade its network to carry ISP traffic - $2.6 million in incremental capital expense, and $11 million "to meet the forecasted ESP demand for ISDN primary rate."11 PacTel Study at 2. As noted above at page 12, however, Internet user traffic routed over these PRI ISDN connections will not contribute to congestion over the PSTN, and ISP payment for PRI ISDN results in a net profit for the ILECs. So the only costs that are relevant for this proceeding would be the incremental capital expenses.
Although Bell Atlantic can point to specific instances totaling only $2.5 million in costs, it claims that the total cost from serving ISPs in 1996 was $30 million. Bell Atlantic Study at 6. To reach this number, however, it makes two unjustified and self-serving assumptions: that every line going into a 512 line capacity line unit serves an Internet user and has a traffic load of 30 CCS; and that a tenfold increase in traffic will result in a tenfold increase in costs.12 It then multiplies this inflated cost figure over every ISP line it presumes to be in its region. Id. at 5-6. It never explains any of these premises, nor does it show that ISP traffic will exceed peak hour traffic at the COs serving the ISPs (and that therefore network reengineering will be required).
Similarly, U.S. West merely asserts without proof that an ISP's usage cost per line is 8 times greater than a business line's cost. U.S. West Study at 2. It does not discuss how it arrives at this figure.
One critical fact never specified in any of the ILEC studies is what proportion of their costs are attributable to routine network maintenance. All PSTN traffic, voice, data, or otherwise, causes some incremental maintenance cost. ILECs are constantly spending money to maintain, redeploy, or replace physical plant. Without knowing the ILEC's baseline level of upkeep for the COs studied, it is impossible to determine specifically whether ISP traffic has caused an increase in costs.
While ILECs have not provided meaningful data for direct comparison, it is still possible to compare the sums the ILECs claim to contribute to Internet user traffic with their reported annual costs. Bell Atlantic has reported annual gross capital expenditures of $2.488 billion for 1995, and average annual gross capital expenditures of $2.300 billion over the last 5 years. Network Services Capital Expenditures (downloaded from http://www.bell-atl.com/invest/ businv r/netserv/nce.htm). Assuming similar expenditures for 1996, the costs Bell Atlantic detailed in its study comprised 0.11% of its annual total.
Another reason for the Commission to be skeptical of ILEC estimates of net costs from Internet user traffic is that ILEC studies significantly understate the revenues this traffic generates.
First, the ILEC studies fail to state that they already collect use-based revenue from Internet traffic. They claim that ISPs avoid paying usage fees because they terminate large amounts of traffic but originate almost none. See, e.g., Dorman Speech at 6; Bell Atlantic Study at 1. These ILECs fail to state, however, that the same calls terminating at the ISPs originate from and are paid for by the ISP user. In no way, therefore, is this unpaid traffic. See, also, ETI Study at 23-24, citing Interconnection Order I in CC Docket 96-98 (state-regulated use fees typically exceed the underlying incremental cost of local network usage). In effect, new access charges would allow ILECs to double-recover for ISP traffic.
Moreover, the ILEC studies make no mention of revenues from sales of residential second lines. According to the Commission's own statistics, the second line penetration rate in 1995 was almost 15% of all households. FCC Industry Analysis Division, "Percentage Additional Residential Lines for Households with Telephone Service (End of Year Data)," December 6, 1996.
It is quite clear elsewhere, however, that second lines produce significant profits. The ILECs' own financial statements and press releases proclaim that residential second lines are an engine driving great revenue growth, and are becoming an increasingly important part of their business. For example, in the third quarter of 1996, U.S. West declared that its residential second line sales had increased 31.7% over the previous 12 months, a number 8 times greater than residential sales growth over the same period. "U.S. West Communications Records Another Quarter of Strong Growth In Core Operations," October 23, 1996, at 1 (downloaded from "http:// www.uswest.com/aboutusw/newsreleases/comm/102396.html"). At the end of 1995, over 8% of its residential customers had purchased second lines. "U.S. West Communications, Inc. Access Lines" (downloaded from U.S. West website at http://www.uswest.com/ aboutusw/investorinfo/factbooks/95comfactbook/oandt/access.html).
Similarly, in the second quarter of 1996, Pacific Telesis reported that second residential lines caused over half its annual residential line growth. "Pacific Telesis Reports Record-Setting Increases in New Customer Lines in Second Quarter," July 18, 1996, at 1 (downloaded from http://www.pactel.com/cgi-bin/getrel?1233) In October, 1996, it reported that the number of second lines had increased by 105% over the previous year. "Pacific Teleses Continues Earnings Growth in Third Quarter," October 17, 1996, at 1 (downloaded from http://www.pactel.com/cgi- bin/getrel?1309).13
Bell Atlantic predicts that one-third of the residential market will have second lines by the year 2000. "Network Services, Strategic Overview" at 3 (downloaded from "http://www.bell- atl.com/invest/businvpr/netserv/overview.htm").
In addition to their usual profits per line in service, ILECs' costs-per-line are even less, and profit margins are even greater, for additional residential lines than for first lines. Second lines make use of capacity that is already in place, because "most ILECs have for many years been providing a minimum of two pairs per residential premises, sometimes as many as four...." ETI Study at 24, citing Deposition of William L. Vowell, March 11, 1996, California PUC.
In total, observers place the nationwide revenue in 1995 from second residential lines at over 1.4 billion, including annual recurring revenues of over 1.2 billion. David Braun, "Baby Bells Score $1.4B From Net, TechInvestor, March 13, 1997 (downloaded from "http://www. techweb.com:80/investor/newsroom/tinews/mar/0313bells.html.body?"); ETI Study at 27. This is a windfall far exceeding even the Bellcore estimate of national costs attributable to ISP traffic. Id. at 26.
Finally, the ILECs have underestimated the revenues they receive directly from ISPs. They assert that the only revenues received are the simple monthly cost of a business trunk line. Bell Atlantic Study at 1; Dorman Speech at 6. This ignores the fees received from additional services like hunt groups, and connection and installation fees. ETI Study at 25.
The steering committee of the Network Reliability and Interoperability Council ("NRIC") has been focusing its efforts on an evaluation of the extent of the problem to the PSTN that might be posed by Internet traffic. NOI at ¶317.14
The steering committee's findings are yet another indication that the ILECs have failed to prove the existence of any problem. As reported to a recent meeting of NRIC, the steering committee has examined "FCC-reportable outage data" and found "there is nothing that would suggest an FCC/network reliability issue." Corrected Minutes of the October 31, 1996 Meeting of the Network Reliability and Interoperability Council, at ¶37. NRIC found that it "is not likely that the Internet as it is presently being used could cause a reportable outage," and it "saw no indication that Internet traffic presented any immediate reliability threat." Id. at ¶¶36-37.
This inquiry marks an important turning point in FCC policy. The issue it raises, whether and how telephone companies can charge for providing local transport and switching services for particular kinds of customers, superficially resembles telephony questions the Commission has routinely resolved for more than sixty years. But this time, the question is infused with fund mental First Amendment issues. The wrong answer will diminish growth of a tranformative tool for advancing both democratic discourse and economic growth.
As the IUC has shown, the imposition of new charges for Internet access would slow the growth of Internet usage and create disincentives for construction of data-friendly networks. ILECs are adequately compensated for their services under the current scheme. Internet use is not a source of congestion on the PSTN.
Access charges are a solution in search of a problem. The FCC should reject them.
|
Daniel J. Weitzner Leslie A. Harris March 24, 1997 |
Respectfully submitted, Joseph S. Paykel Andrew Jay Schwartzman Gigi B. Sohn MEDIA ACCESS PROJECT Counsel for the Internet
User |
Back to new media main page2. The Commission's current system of regulation classifies Internet service providers ("ISPs") as "enhanced services." It considers such enhanced service providers to be end users, buying many lines for the purposes of processing user traffic and transferring it to a private data network. It does not impose the per-minute access charges for terminating traffic on incumbent local exchange carrier ("ILEC") networks. NOPR at ¶284-85.
3. For example, in its "Bandwidth Forum" on January 23, 1997, the Commission received voluminous testimony and data on this subject. Although they may have differed on solutions, virtually all witnesses agreed that inadequate bandwidth was a serious problem.
4. Daniel J. Weitzner, one of the undersigned counsel to this document, testified at length about the importance of the open architecture of the Internet's delivery system at the Commission's Bandwidth Forum.
The strength of the Internet derives from both affordable access and a uniquely open architecture. Traditional communications media such as radio and television have been affordable and readily available around the country, but have failed to enable full democratic [discourse] because of architectural limitations. For example, online discussions of political issues enable users to exchange views, and even pose questions to political figures, in a way that broadcast television can never support. The Internet's architecture allows for a diversity of views and exchange of information which are simply impossible in any other communications medium.
The Internet manifests five critical attributes in its basic architecture that give it such potential to enhance democratic discourse: Decentralized, gatekeeper-free access; Bi-directional, interactive capability; Multiple, competitive access points; Open standards; and Affordable service.
In considering policies to promote broader access to advanced communications infrastructure, we hope that the Commission will take these characteristics as baseline policy goals. The democratic potential of the Internet will only be realized with broad access to the Net for both individuals and community organizations. The Commission can help bring that potential of the Internet to the broad cross-section of the population by encouraging the development of new access options that promote these essential attributes of the Internet architecture.
Daniel J. Weitzner, Deputy Director, Center for Democracy and Technology, "Expanding Access To the National Information Infrastructure For Individuals and Community Organizations: Open Architecture and Affordable, Digital Bandwidth," presentation at Federal Communications Commission Bandwidth Forum, January 23, 1997.
5. For example, the Commission has received a petition from the America's Carriers Telecom- munications Association observing that some Internet users have begun to use the information networks for real-time two-way conversations, and seeking regulatory protection. Petition for Declaratory Ruling, Special Relief, and Institution of a Rulemaking, RM-8775 (filed March 4, 1996)("ACTA Petition"). The Commission plans to address the issues raised in that petition in a separate proceeding. NOI at n. 438.
6. As noted above, at n. 2, the Commission includes ISPs within the broader category of enhanced service providers.
7. See n. 2, above.
8. As the ETI study demonstrates, this is especially true in the majority of ILEC switches and network components where common channel signalling system 7 ("SS7") is present. With SS7, if it is determined that the ISP's lines are in use, the calling party will receive a busy signal generated by the originating CO. ETI Study at 7-8 n. 13. Therefore, no interoffice network resources will be used by these customers, and there is little chance that traffic caused by inadequate line purchase at the ISP would "`back up' into the public network." Id.
9. This assumes the same number of lines are served by each CO. Even removing this assumption would not change the end result -- the congestion alleged by these studies would still affect a vanishingly small fraction of one percent of these ILECs' lines in service.
10. Total number of access lines for Pacific Telesis calculated as 18.169 million by dividing total of 48,330 employees over reported workforce efficiency ratio of 26.6 employees per 10,000 access lines. Id.
11. Aside from the speculative nature of the second element of Pacific Telesis' costs, it did not specify whether ISPs paid any initiation of service fees to defray the cost of these new ISDN trunks.
12. Indeed, the Commission has received evidence that an increase in peak-hour capacity does not cause a 1:1 proportional increase in costs. ETI Study at 11. This is because many network components would not need replacement because they are "non-blocking," i.e. deployed in such a way that they can handle greater loads than the peak capacity of other network elements. Id.
13. It is especially revealing that Pacific Telesis' fourth quarter, 1996, report made no mention of second line sales to Internet users. It released this report after it had commenced its bid to impose access charges on ISPs. Instead, for the first time, it attributed the growth of residential lines only to "[a]ggressive marketing efforts and the strengthened California economy." "Pacific Telesis Reports Strong Quarterly and Annual Growth," January 21, 1997, at 2 (downloaded from "http:// www.pactel.com/cgi-bin/getrel?1457").
14. The Commission has noted that NRIC, itself comprised of representatives from leading parties including many ILECs, trade associations, and computer equipment manufacturers, has appointed this special steering committee to investigate Internet developments affecting the PSTN and to report whether specific actions would be appropriate. The Commission declared that it did not intend for the NOI to supersede NRIC's findings, and that these recommendations will compliment the record the Commission develops. Id.