In reporting on yesterday's announcement by the FCC that it intends to formalize its net neutrality guidelines, the Washington Post quotes Google Chief Executive Eric Schmidt as saying "It is possible for the government to screw the Internet up, big time." And who can argue with that? In fact, that is the reason that David Farber, a former FCC Chief Technologist, has come out against government imposition of net neutrality regulation of Internet service providers. Farber expressed it colloquially at a recent ITIF event, when he said that, after being around Washington, DC for many years, as between government and industry, he'd rather take his chances with industry. Farber understands that mistakes made in the competitive marketplace can be quickly identified and remedied; regulatory mistakes can last a lifetime.
So, how can net neutrality rule proponents and opponents agree on the same fundamental premise about government screwing up the Internet, big time? Because they are talking about two different things, and this has tremendous implications for the debate. For Google, it and other Internet-based applications and service providers are "the Internet," but the infrastructure providers acting as Internet service providers are not. They, in the words of FCC Chairman Genachowski, merely provide the "on-ramps to the Internet;" they are not "the Internet." Farber understands, quite rightly, that one should not make such distinctions: in the Internet ecosphere, as in a two-sided market, both ends are "the Internet."
In fact, years ago (in the late '90s) I attended a Washington DC policy luncheon hosted by the Legg Mason Precursor Group. The topic du jour was cable "open access," the precursor to today's "net neutrality." I asked Internet doyenne Ester Dyson where the Internet began and ended. Her response, as best I recall, was: "It begins and ends with me and you." In other words, every computer (and the user behind it) is part of the Internet network of networks by virtue of its utilization of Internet addresses (the Domain Name System) and Internet communications and transport protocols (TCP/IP). There is no "big I Internet" (as we used to say), there are millions or billions of locations linked together through a common addressing system and protocols, used by you and me. That, at least, is how I understand the term "Internet ecosystem."
The FCC's net neutrality rulemaking is aimed at adopting "fair rules of the road" for ISPs. Here "net" is understood as "network," and the highway analogy allows the FCC to regulate the Internet's "on-ramps" without, in its view, regulating the Internet itself. Here, "Internet neutrality" is segregated from "network neutrality." This harkens back quite directly to the common carrier framework applicable to the narrowband Internet, where regulated common carrier telephone lines were the means by which telephone subscribers accessed largely third-party, non-facilities-based ISPs. The model was initially applicable to telephone company DSL, a broadband service; the underlying transport was treated as a common carrier transport service and the ISP portion of the service as an unregulated "enhanced" or "information" service. The broadband cable modem service as provisioned from day one contained no separate offering of underlying transport. The FCC recognized this in its 2001 Cable Modem Declaratory Ruling, classifying the cable modem service as a (then unregulated) "information service." It declined to classify cable modem service as a "telecommunications service" by refusing to tease apart the constituent elements of the Internet service for the purpose of recognizing a separate transport service within the bundle of attributes offered to the public by the cable operators. The agency found the content and conduit elements of the service to be "inextricably intertwined." By doing so, it effectively declined to first force the cable operators to act as transport common carriers (or one might say, simple highway "on-ramps") so that it could then classify the entire cable modem service as a common carrier "telecommunications service." This classification decision was upheld by the U.S. Supreme Court in its Brand X decision in 2005. Subsequently, with the aim of encouraging additional broadband infrastructure investment and deployment, the FCC extended this regulatory classification to wireline, wireless and broadband-over-powerline Internet services.
The current FCC appears poised to extricate or de-intertwine the Internet content elements of broadband ISP offerings from the transport (on-ramp) elements, and effectively impose common carrier -- or public utility -- requirements on the providers of such services. At least, this is what the rhetoric and text of the proposed rules suggests. Of course, even common carriers are permitted to engage in just and reasonable discrimination under Title II of the Communications Act. Section 201(b) declares unlawful "unjust or unreasonable" rates, terms, practices, and conditions of service; section 202(a) declares it unlawful for any common carrier to make any "unjust or unreasonable discrimination" in rates, terms, practices and conditions of service. Ironically, or sadly, the FCC's proposed rules for information service providers like the broadband ISPs are even more stringent: they prohibit all discrimination in the carriage of information, use of service and attachment of devices of the user's choosing, subject only to "reasonable network management," to be defined by the FCC. Maybe it is all a feint to get ISPs to decide to "voluntarily" become Title II common carriers in their provision of Internet services to avail themselves of greater statutory flexibility and protection for their services. After all, in America, anyone can choose to act as a common carrier (although the reverse process is nowhere near as simple).
Will we all be better off if the FCC moves forward and effectively imposes common carrier mandates on broadband ISPs? If I thought so, I would be in favor of such action. But I don't. We have achieved what few other countries have: facilities-based broadband competition for the vast majority of the U.S. population, in part through a "light touch" regulatory framework that gave the providers flexibility to invest and provision Internet services according to their business judgment and consumer demands. As I have written in the past, the number of actual violations of the FCC's Internet policy principles are "infinitesimal compared to the number of broadband Internet access service subscribers." And, as my colleagues Adam Thierer and Berin Szoka have noted, the potential for regulatory creep up the Internet stack from infrastructure to applications is sometimes glacial, but almost inevitable.
Proponents of prescriptive regulation such as Google fear that the infrastructure providers will act on anticompetitive incentives to favor or disfavor Internet traffic that competes with a service they provide (e.g., voice or video). But we already have antitrust authorities that are more than capable of policing such anticompetitive activity. The case for imposing industry-specific ex ante "rules of the road" for broadband ISPs has not yet been made, and we have a year or more of FCC fact-finding to tell us if it can be. I don't think it can, and that is why I am not in favor of bringing back the rack and re-imposing a stringent common carrier regime in a thriving marketplace. The risks of inaction in this arena seem small to me, whereas the risks of imposing preemptive common carrier-style regulation in the dynamic Internet ecosystem seem huge. I have quoted these words of wisdom before from former FCC Commissioner (and former Acting Chairman) James Quello, and return to them again here: "If it ain't broke, don't break it."