The Uses and Misuses of the Layers Model

The Skeptical Regulator*
Modest Observations on Regulatory Policy
Release 2.3 n August 2004

The panel discussions at this year's Aspen Summit are loosely organized around the "layered model" of the Internet. The 'generic' layers model - comprised of the physical, logical, applications and content layers of the Internet - is an engineering metaphor that more closely captures the reality of communications today. It is based upon the layered protocol stack, one of the two defining characteristics of Internet design (the end-to-end principle being the other).

While a number of academians and other industry commentators have proposed using a layers approach, normally after a transition period, in a new federal regulatory regime, MCI's Rick Whitt was the first to utilize the model for a discrete set of policy recommendations. The MCI model would extend an unbundling mandate to ILEC DSL services for at least thirty months, and until an absence of market power is shown or three other commercially viable broadband providers exist in a given market. [1] Where market power is shown, "other" broadband providers could be subject to the requirement as well, unless they comply with a set of "Internet access rights." [2] As a default measure, it would require "light-touch" regulation such as resale on all broadband platform providers, regardless of technology. [3] And although it does not bar vertical integration, it presumes that the "failure to regulate last-mile broadband facilities will allow those providers to extend their market power into the higher layers. . . this form of vertical integration would cause undue harm to the Internet." [4]

More recently, the New Millennium Research Council (NMRC) released a report which contains a series of articles opposing the MCI proposal. [5] In attacking the transfer of the existing telecommunications regulatory regime to broadband, the incentives (and disincentives) it would create, and its suspicions of vertical integration, the NMRC report levels a number of valid criticisms at the MCI proposal.

In one sense, and to its credit, by moving first to incorporate the work of certain scholars in recent years MCI has "seized" the layers model. But this also may be the reason why other aspects of the NMRC report are overstated. Rather than confining its critique to the MCI proposal, the report mounts a general assault on the layers model and concludes that it is "fatally flawed as a framework for new regulation or legislation." [6] Several articles within the report appear to blur the distinction between the specific MCI proposal and the layers model in general. [7] Other articles imply that "layered model advocates" are of the same regulatory bent, [8] or have a "hidden agenda" of extending legacy telephone regulations to the broadband world. [9] This is not a fair reading of some of the pertinent literature on the subject.

Take a recent article by Professor Phil Weiser, for example. [10] Professor Weiser argues that the FCC should use its Title I ancillary jurisdiction and shift its focus to a layered model that "regulates functionally similar services in the same way regardless of the underlying platform." He then goes on to explain that the agency could employ a reactive, antitrust-like model of regulation for the broadband market by announcing a standard of nondiscriminatory access. This would allow parties to develop their own business arrangements, with only ex post "scrutiny of discriminatory conduct alleged to lack a redeeming efficiency justification."

Professor Weiser therefore argues that a model with minimal regulation is appropriate. While he stakes out a middle ground between a model that would mandate open access and one that relies entirely on market forces, his argument is a far cry from a layered model that seeks to, say, severely limit vertical integration or impose heavy-handed unbundling and price regulation on physical platform providers.

At bottom, the fundamental point about the layers model is that regulating, or deregulating , the physical platform is no longer related to regulating classes of service. Compared to the current silo approach to regulation, it is a more intelligent way to look at the broadband world and contemplate all sorts of policy changes. And, while it could provide a foundational premise for a new regulatory framework going-forward, it is, at its core, sedulously neutral on questions of regulation. Indeed, by underscoring the nature of Voice over Internet Protocol (VoIP) as a complementary, and competitive, application that rides on broadband platforms, the layered model underscores the case for limited, if any, regulation of VoIP.

As we consider "The Future of the Internet," it is seemingly premature, as the consensus on the need for a new Communications Act continues to grow, to throw the layers model in the analytical trash bin. Granted, it is not a perfect model. It does not obviate public choice concerns. It does not square technological dynamism with static antitrust analysis. And, if it is used as a premise for a new regulatory regime, that regime must properly define interfaces between layers and allow for networks to evolve (which may prove to be, as several authors in the NMRC report and commentators such as George Gilder suggest, sufficient barriers to the model's utility). [11] However, if one rejects the layers model, then another framework needs to be formulated, as the preservation of a regulatory framework based on the legacy services offered over a particular platform - be they voice, video, data, or what have you - is increasingly irrational and should be abandoned.

- The Skeptical Regulator


* The Progress & Freedom Foundation debuted The Skeptical Regulator in July 2003. The periodic essays are authored by president Raymond Gifford and other fellows at the Foundation. Send comments to

  1. Richard S. Whitt, Codifying the Network Layers Model at 21-22 (Mar. 2004).
  2. Id. at 22.
  3. Id . at 19.
  4. Richard S. Whitt, A Horizontal Leap Forward , MCI Public Policy Paper at 53 (Mar. 2004).
  5. New Millennium Research Council, Free Ride: Deficiencies of the MCI 'Layers' Policy Model and the Need for Principles that Encourage Competition in the New IP World (Jul. 2004)("NMRC Report").
  6. NMRC Report at viii.
  7. See, e.g., NMRC Report at 4-7, 8, 11-15, 16, 20-21.
  8. Id. at 5.
  9. Id. at 11-15.
  10. See Philip J. Weiser, Toward a Next Generation Regulatory Strategy , 45 Loyola U. Chi. L.J. 41 (2003), available here.
  11. See Testimony of George Gilder before the Senate Commerce Committee, Telecommunications Policy: A Look Ahead (Apr. 28, 2004).



The Progress & Freedom Foundation