Beyond the Layered Model J Scott Marcus Senior

Beyond the Layered Model J. Scott Marcus Senior Advisor for Internet Technology, FCC The opinions expressed are solely those of the author, and do not necessarily reflect the views of the FCC or of the European Commission. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus Slide 1

Beyond Layers § Several theorists have proposed a layered approach to telecoms regulation, including: § § § Werbach Whitt Sicker/Mindel Nakahata Solum § Diagnosis of problem generally correct. § Underlying objectives are reasonable. Ø What about the recommended cure? February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 2

Beyond Layers § As a thought model, the layered approach provides interesting insights. § The Layered Model provides one useful caveat: Beware obligations that cross layers. Ø Otherwise, the Layered approach offers surprisingly little guidance to the regulator. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 3

The Diagnosis § The Communications Act as amended comprises regulatory silos: § Wired telecommunications (Title II) § Wireless (Title III) § Cable (Title VI) § The convergence of computing and communications renders those silos irrelevant. § Broadband § Voice over IP § The mismatch of law to reality causes problems. § Irrational conclusions. § Regulatory arbitrage. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 4

Layered Approach February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 5

Ostensible Goals of the Layered Model § Closer mapping to reality as all networks converge to an Internet-based model. § Straightforward regulatory categorization of services. § Regulation that responds naturally to technological evolution and convergence. § Avoidance of regulatory silos. § Straightforward determination of regulatory consequences. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 6

The Layers of the Internet Layers interact with peer layers Application Transport Network Data Link Physical Layers derive services from successively lower layers February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 7

Closer mapping to reality as all networks converge to an Internet-based model? § Are all services based on the Internet today? § Can we be sure that all services will be based on the Internet in the future? § Can we say with certainty that network products and services will never evolve beyond the Internet? Ø Risk of regulatory hubris. Ø So much for regulation that responds naturally to technological evolution and convergence. . . February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 8

Hybrid Services § The most challenging services represent a mix of Internet and traditional elements: § AT&T Internet-based calling card services § Vonage § Skype dial-out and dial-in § How useful is the distinction between the Logical and the Physical transmission layers for such services? § Challenges both correspondence to reality and ease of categorization. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 9

Straightforward regulatory categorization of services? § Layers were intended to be an aid to specification, not necessarily a detailed guide to implementation. § Products and services routinely and systematically violate layering. § Layers are often used recursively (e. g. tunneling). § If layers were tied to economic and regulatory consequences, we might expect systematic manipulation and arbitrage. Ø Determination of layer is notably easier than determination of telecommunications service versus information service. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 10

Avoidance of regulatory silos? § Old system had old silos. § Wired telecoms § Wireless § Cable Ø Implies asymmetric treatment for broadband? § New system has new silos. § Logical (Internet) § Physical (Non-Internet) Ø Implies asymmetric treatment for IP telephony? Ø Risks needlessly or inappropriately violating technological neutrality. Ø Vertical silos -> horizontal silos February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 11

Straightforward determination of regulatory consequences? § Computer Inquiries declined to regulate enhanced services: § Competitive entry would be straightforward, provided that any underlying bottleneck facilities were available on nondiscriminatory terms. § Competition could prevent competitive harms. § Promotion of innovation services for consumers. § Internet services have been unregulated, not because they constitute logical transmission, but rather as a consequence of the prospect of competition. § Economic tests can work in a converged world; semantic definitional tests often fail. Ø Triggering regulation based on an incidental consequence rather than an underlying root cause is a step in the wrong direction. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 12

Straightforward determination of regulatory consequences? § Advocates regulation of last mile facilities because they represent competitive bottlenecks. § A lightning rod for objections. § Must be balanced against other objectives: § Encouragement of investment. § Support for nascent technologies. § Possible to reach this conclusion without recourse to the Layered Model. Ø The Layered Model offers little or no specific guidance otherwise as to how to regulate within a Layer. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 13

Market Power in the Last Mile § Most papers on the Layered Model argue for regulation of last mile facilities because they represent bottleneck facilities. § Is this really true in all geographic areas? § Is this really true for all time? § What mechanisms does the Layered Model provide for avoiding regulation in geographies where market power is not an issue, or when it no longer is an issue? § If market power were to emerge at a higher layer, should we necessarily decline to take action? February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 14

Market Power in the Last Mile § Stevens Report and Layered model seek to: § Regulate underlying telecommunication services. § Leave Internet ISP services unregulated. § “Assess market power separately for each layer. ” (MCI comments on IP Enabled Services NPRM) § Why not instead assess market power separately for each market? § § Layers may not correspond to markets. Markets may not correspond to layers. Classification of services into layers may be intractable. Identification of markets, by contrast, could draw on existing Do. J/FTC guidelines and longstanding jurisprudence. § Economic metrics rather than semantic definitions. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 15

Market Power in the Last Mile § Assume arguendo that it were deemed desirable to use regulation to address bottlenecks that might exist in conjunction with last mile facilities. § A solution based on European principles leads much more directly to the same conclusion: § Explicitly identify the market using economic tests. § Determine who, if anyone, has significant market power. § Apply remedies (a minimally adequate set to address the likely competitive harms) solely to those parties. § Remove corresponding remedies where significant market power is not present, or when it is no longer present. § Provides far greater granularity and flexibility. February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 16

Vo. IP and the Layered Model § The FCC’s public forum on Vo. IP (12/2003) identified five key issues: § § § Intercarrier compensation Funding for universal service Lawful intercept Access to emergency services Access for those with disabilities § Three scenarios: § Red: Free World Dial (FWD) § Blue: AT&T Calling Card § Purple: Vonage Ø WHAT does the Layered Model tell us to do? February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 17

An Alternative Layering • Does this really add value? • How would we regulate content differently from applications? • Is this perhaps already obvious? February 13, 2005 Silicon Flatirons, Boulder CO - J. Scott Marcus 18