Testimony given by Senior Fellow Geoge Gilder at a Full Committee Hearing:
Telecommunications Policy: A Look Ahead
Wednesday, April 28 2004 - 9:30 AM - SR - 253
U.S. Senate Committee on Commerce, Science & Transportation
Mr. Chairman and Senator Hollings, thank you for the opportunity to appear before your committee today. Your selected topic is crucial to the well-being of the U.S. and global economies, and I appreciate your deep interest in the subject.
Overthrowing matter and media with the new worldwide web of glass and light and air should be a happy and defining event in the history of man. Global information networks offer unprecedented potential opportunities for economic growth, cultural revival, and individual freedom and empowerment. Yet the United States has in large part blocked the path of the technologies and companies needed to consummate this vast new infrastructure of chips, fiber optics, antennae, digital storage, and software.
Although American companies invented almost all the technologies crucial to the Internet, we have fallen behind many other nations in the deployment of these technologies.
The U.S. now ranks eleventh internationally in residential "broadband" access. Using the FCC’s silly 200-kilobit-per-second definition, some now say that 25 percent of American homes have broadband. But by the standards of Asia—where most citizens enjoy access speeds 10 times faster than our fastest links—U.S. residences have no broadband at all. U.S. businesses have far less broadband than South Korean residences. South Korea, for instance, has 40 times the per capita bandwidth of the U.S. Japan is close behind Korea, and countries from China to Italy are removing obstacles to the deployment of vDSL, fiber-to-the-home, and broadband wireless networks.
Asian broadband also proves there was no Internet "bubble." Today, Korea runs over the net between a three and five times larger share of its economy than we do. Riding the bus to work, Koreans watch television news and exchange video mail over their mobile phones. They enjoy full-motion video education and entertainment in their homes. Many of the dot-coms that failed in America due to the lack of robust broadband links are thriving in Korea. Consider that by this time next year Verizon Wireless’s 38 million customers will enjoy faster Internet access via their mobile phones than through their Verizon DSL connections to their homes. Only the most severe disincentives to invest could have yielded such a result, which defies the laws of physics. The American Internet "bubble" was actually a crisis of policy.
The Telecom Act of 1996 was meant to "deregulate" America’s telecom infrastructure and technologies, the most dynamic sectors in the entire world economy. But after the usual lobbying and horse-trading, the Act turned into a million-word re-regulation of the industry. Regulatory actions by the FCC and the 51 state utility commissions greatly exacerbated the bad parts of the Act and distorted many of the good parts. As I predicted the day after it was enacted, the result was a carnival of lawyers, micro-mis-management by bureaucrats, price controls, the socialization of infrastructure, the screeching halt of innovation and investment in the "last-mile" local loop—and the Great Telecom and Technology Crash of 2000-2003.
In the last year or so, the FCC has partially reversed some of its most egregious errors. Some are still being adjudicated in the courts. But U.S. telecom remains a highly regulated, highly taxed sector of our economy. The mistakes of the last 10 years have greatly harmed the U.S. economy, and continued gridlock and inaction threaten to shift American leadership in technology to Asia, which has embraced the Internet with open arms.
Today, just as the telecom and technology sectors exit a three-year depression, we are in danger of repeating the very worst mistakes of the 1996 Telecom Act, but this time on an even grander scale. In today’s testimony I will address and refute one particular proposal that is being offered as the basis for the new telecom legislation. In doing so I hope also to offer an alternative vision.
The new "big idea" in telecom regulation comes from a host of learned and experienced telecom thinkers: the likes of former FCC authority Kevin Werbach, Stanford law professor and technology author Lawrence Lessig, industry analyst Roxanne Googin, and IPioneer Vint Cerf, to name just a few.
The idea is mandated "open access" to the logical layers of the network, and it is embodied in a new legislative proposal by MCI, "A Horizontal Leap Forward: Formulating a New Public Policy Framework Based on the Network Layers Model." A horizontal layers approach would supposedly be a radical shift from the "vertical silos" approach now used, where telephony, cable, and wireless, for example, are regulated based on historical industry definitions, not generic functional categories. The common denominator of Internet Protocol (IP)—supposedly the basis for all future communications networks—is said to necessitate the new layered regulatory approach.
Barely recovering from the FCC’s TELRIC and UNE-P "open access" mandates that chopped up and assigned ownership rights to the physical infrastructure—the hardware—of the Net, we now face the prospect of rigid reassignment of content, applications, services, and protocols, too.
Whatever it is called, it represents more micromanagement of a dynamic industry in the midst of major technological transitions.
The new proposal feeds on fear—fears that cable TV companies or the Bells might seek to leverage their broadband networks by wrapping content into their conduits, or that Microsoft might keep "tying" new applications into Windows, or that Google might monopolize information on the Net (yes, there is already an organized effort to turn Google into a public utility). MCI’s layering proposal defines rigid boundaries between content (voice, text, video), applications (e-mail, browsers, VoIP), protocols (TCP/IP, HTTP, FTP), and infrastructure (wires, switches, spectrum, PCs, handsets). In a paper entitled "Codifying the Network Layers Model," MCI proposes to "quarantine" major providers of one of the layers within that layer, and to prohibit them from vertically integrating into another layer unless they offer wholesale open access to all competitors. Lessig, MCI, and company worry that the "end-to-end" nature of the Internet—where any terminal attached to the net can be reached from any other terminal--will be threatened if these new layering rules are not adopted.
Layering proponents, however, make a fundamental error. They ignore ever changing trade-offs between integration and modularization that are among the most profound and strategic decisions any company in any industry makes. They disavow Harvard Business professor Clayton Christensen’s theorems that dictate when modularization, or "layering," is advisable, and when integration is far more likely to yield success. For example, the separation of content and conduit—the notion that bandwidth providers should focus on delivering robust, high-speed connections while allowing hundreds of millions of professionals and amateurs to supply the content—is often a sound strategy. We have supported it from the beginning. But leading edge undershoot products (ones that are not yet good enough for the demands of the marketplace) like video-conferencing often require integration.
Metaphors from the Telecosm help explain the fluid nature of these layers that MCI wants to preserve in concrete. Consider Corvis, our favorite optical equipment company and national fiber optic bandwidth provider. It blows apart the MCI approach on several fronts. First is CEO David Huber’s architecture of an all-optical network, devoid of electronic regenerators and protocol readers, which unites content and conduit by using colors of light both to bear the message and to determine the path of the circuit. It radically collapses the top layers of the OSI (Open Systems Interconnection) stack used in the Sonet voice and data networks of the past, not so much redefining the interfaces as transcending them. A "switchless" web of always-on fixed lambdas (wavelengths of light) can function as both the physical and logical layers of the Net because the intelligence is embedded in the path. There will be some controlling devices at the edge of the network, and IP will still be widely used, but the heyday of IP packet switched networks may well be over. Typically government enshrines the past in the name of progress. In uniting Corvis, a cutting edge equipment provider, with Broadwing, an infrastructure builder and service provider, Huber is also betting that IP networks are not inherently modular, where equipment from a thousand providers can easily be cobbled together to deliver high-bandwidth, low-latency services, but that networks are still in fact in an era of undershoot where an integrated provider can deliver a superior product at a much lower cost.
Our favorite digital chip company, EZchip, also explodes the idea that the layers of the Net can always be defined and "quarantined." Where until now data flowing through the seven layers and numerous sub-layers were parsed and modified by a gaggle of hundreds of chips connected by thousands of wires and glue-logic galore, EZ puts all seven layers of the OSI stack onto one-chip, performing all the essential functions of an Internet router on a single sliver of silicon. The "layers" are once again transcended when EZ’s software tools allow programmers to tell the chip what to do without even referring to the rigid layers, channelizations, protocols, and interfaces used in the previous software environment. Is this fair? Should EZchip be allowed to invade someone else’s turf, perhaps that of Cypress’s high-end content addressable memories (CAMs) or Broadcom’s Silicon Spice communications processors or the sacred code of the OSI idol? Or to blow apart someone’s whole field, like EZ could one day do to the many providers of communications ASICs (applications specific integrated circuits), or to Internet router king Cisco itself? It might be said that the “layering” proposals now in circulation are yet another (if more clever) attempt by competitors to target the Bell telephone and cable TV companies. Indeed, MCI’s own paper implies the cable companies (bundling network, ISP, and content) and the Bells (bundling network, ISP, and voice) are already stomping all over the layers, creating a muddy (and hopefully one day illegal!) mishmash of vertical integration. What a coincidence that the activities of its rivals violate MCI’s framework and cry out for cleansing and re-ordering (read structural separation, consent decrees, price controls, divestiture) by new teams of FCC horizontalawyers and IPolice.
But if the proposals are meant as anything more than political lobbygagging of rivals, if the proponents really mean their model legislation as a principled, generic set of rules, then we must consider the logical consequences of such new laws. If applied dispassionately, how would such general rules affect the rest of the Internet, communications, and technology industries?
Should Google be able to leverage search into Gmail, or to supply content using its proprietary algorithms and its physical network of 100,000 servers? Shouldn’t any rival search provider be able to feed off of Google’s advanced infrastructure? After all, wouldn’t it be impossible to recreate Google’s massive web of global intelligence? Doesn’t Google’s superior infrastructure exhibit "market power"? Might Google actually evolve into a general provider of web-based information management services, rivaling the PC-based Microsoft, or should Google be "quarantined" as a search provider? Or maybe we should structurally separate Google into three companies: an infrastructure provider (its 100,000 networked servers plus algorithmic IP), a content/advertising company, and an information services company (Gmail plus future knowledge management applications). Surely FCC bureaucrats can make these easy distinctions and explain the resulting penalties to weary entrepreneurs who have just spent 10 years of their life building a new service that people really like.
Should Sony be able to demand that its PlayStation gamers get access to Microsoft’s Xbox Live online video game network? Should Amazon be able to aggregate and make searchable the text of hundreds of thousands of books? Should Sprint PCS or Verizon Wireless be allowed to develop specialized content delivery platforms or applications that take advantage of their superior wireless data networks? Sprint was the first to build its own photo-sharing platform, and it is apparently the most user-friendly wireless photo-sharing system. Can we let such infrastructure-leveraging stand? What if Equinix (the data center company that almost defines of the integration of the physical, protocol, application, and content layers of the Net) succeeds in becoming the overwhelming meeting place (peering point) for the world’s network, e-commerce, and content providers? Network economics suggest the concentration of all the largest Internet players in Equinix facilities is possible, or even likely. If Equinix achieves such "market power," are we to assume that other "virtual data centers," like the CLECs before them, could force Equinix to "open up" its hosting facilities so that the new virtual competitors can offer services over infrastructure they did not build? Why should anyone build risky and expensive new infrastructure if it can be readily used by competitors. What about Microsoft integrating easy-to-use voice-over-IP software into its next operating system? Should Microsoft rival Real Networks be barred from aggregating music and video for download with its RealPlayer multimedia suite? All of these are, to one degree or another, inter-layer integrated products and services.
Proponents of "layering," or "Net neutrality," or a free Internet "commons," assume there is one network, that it is sufficient and timeless, that no new networks are possible or needed. They want innovation on the edge, in the form of software apps and Wi-Fi attachments. Innovation in the core is either assumed or ignored. The logical conclusion, however, is that since the "best network"—the free commons—cannot make any money, there will be no network. And just how much innovation at the edge will there be if there is no innovation—no bandwidth—in the core? MCI’s "horizontal leap" asks authorities to pursue vigilantly those who would exploit "network choke points" or take advantage of "network effects." In industries where "entities seek to obtain market power" (i.e., seek to make money in a business enterprise), policymakers need o ensure four things: "open architecture, open access, universal access, and flexible access." When imposed by regulators or courts in a national capital, these four euphemisms boil down to one hard reality: socialization and micromanagement of the "architectures" and "access" networks built by others.
The ability to tie and merge and break apart and outsource products, services, and technologies are the very stuff of business. As is the ability to pursue an unguaranteed return on one’s risky investment. As is the decision how to price these products and services. Some services will be bundled. Some will be free, loss leaders to leverage the purchase at another point of sale. But the entire system cannot be free. Everybody else’s product or service, except one’s own, cannot be a commodity, barred from bundling or profit.
The companies that enable this broadband world will be able to charge for it during the years that they provide the optimal service. Their initial margins will be high. When communications becomes a commodity, as it eventually will, the margins will drop. This is not a catastrophe. No one has a right to high margins for a commodity service. But the Telecosm is still an arena of innovators, such as Corvis, EZChip, Qualcomm, Verizon Wireless, Essex, AFCI, Agilent, and hundreds of others, who will enjoy large monopoly rents until their inventions are standardized and commoditized and the leading edge moves elsewhere. The telecom industry is nowhere near some mythical paradox of perfection or cul de sac bargain basement of commoditization. It is still engaged in a thrilling adventure of putting together worldwide webs of glass and light that reach from your doorstep or teleputer to every other person and machine on the planet. It is long distance and it is local, it is packetted and circuited, it is multithreaded and aggregated, it is broadband and narrowcast, all at once. These crystal palaces of light and air will be hard to do and the world will reward the pioneers who manage to build them. The real threat to monopolize and paralyze the Internet is not the communications industry and its suppliers, but the premature modularizers and commoditizers, the proponents of the dream of some final government solution for the uncertainties of all life and commerce.