From Wires to Waves

As wireless telephony goes digital, it gets very cheap very fast.
George Gilder
Forbes ASAP
June 1, 1995
Print ArticleU.S. Sen. Ted Stevens of Alaska wants to know: With deregulation of telecommunications, who will bring connections to Unalakleet, to Aleknagik and to Sleetmute? Who will bring 500 channels up the Yukon with the salmon to the people in Beaver? What will happen to the Yupik, the Inupiat and the Inuit? Will we leave them stranded in the snow while the world zooms off to new riches on an information superhighway?

A senior Republican on the Senate Commerce Subcommittee on Communications, Stevens is a key figure in the telecom deregulation debate on Capitol Hill. As he contemplates the issues of restructuring communications law, he has reason to be suspicious of the grand claims of an information age. He knows that universal service--the magic of available dial tone in your own home--has hardly reached rural Alaska at all. As George Calhoun points out in his sort invented by Alexander Graham Bell in 1881 and now extended to some 95% of American households) are simply not feasible, either technically or economically, in many remote regions.

In Beaver, for example, there is one telephone in a hut linked to a nine-foot satellite dish. Permafrost and cold economic reality make it impossible to extend dial tone to the several hundred households of this town, even though its average household income, mostly from salmon fishing, is some $120,000.

Ted Stevens is right to be concerned. Portentiously sharing his concern are other powerful Republicans from rural states, including Larry Pressler of South Dakota, the chairman of the subcommittee. Extended now from phone service to broadband digital superhighways, their concerns could pose a deadly obstacle to true deregulation of communications and thus to continued American leadership in these central technologies of the age. At stake is some $2 trillion of potential value to the U.S. economy (see Forbes ASAP, April 10). The problems of universal service in Alaska disguise the more profound paradox of telephone service in most of the world.

The fact is that the universality of telephones is crucial to their usefulness; yet universal service using current technology is totally uneconomical and impractical. Snow and ice are the least of it. The basic problem is the architecture of the system, with a separate pair of lines, on average two miles long, devoted exclusively to each user. It simply does not pay to lay, entrench, string, protect, test and maintain miles of copper wire pairs, each dedicated to one household that uses them on average some 15 or 20 minutes a day.

Connections in cities are one thing. Urban access systems comprise a bramble of millions of wire loops, each linking a home or business telephone to a nearby central office switch. Under a half mile in length, these lines still represent some 80% of the cost of the system. But because the lines are short and often bundled together, city telephony benefits from economies of scale and convenience. In rural areas, however, the copper lines cost between 10 and 30 times as much per customer as they do in cities.

Moreover, Calhoun reports that in general, phone companies cannot supply ISDN (integrated services digital network) and other digital services over twisted-pair wire more than 18,000 feet (some 3.5 miles) from the central office. Perhaps a third of all the nation's phones are more than 3.5 miles from a central office.

What saves us is socialism. Closing the huge differential between the costs of serving rural and urban customers is a Byzantine web of cross-subsidies, whereby inner-city and business callers in urban areas subsidize the worthy citizens of Kirby, Vt.; Vail, Colo.; Mendocino, Calif.; Round Rock, Texas, and Tyringham, Mass., among other bucolic locales, to the tune of billions of dollars. Overall, subsidies from business and urban customers to rural and other expensive residential users total some $20 billion a year. In case the cross-subsidies do not suffice to guarantee universality, Congress has established a $700 million "Universal Service Fund." For all that, some 5% of homes still lack telephone service (compared with 2% unreached by TV, which faces no universal service requirement).

Lending huge physical authority to this Sisyphean socialist scheme are some 65,049,600 tons of copper wire rooted deep in the rights of way, depreciation schedules, balance sheets, mental processes and corporate cultures of the regional Bell operating companies and other so-called local-exchange carriers. The minimum replacement cost of these lines deployed over the last 50 years or more--and still being installed through the mid-1990s at a rate of at least five million lines a year--is some $300 billion. By comparison, Calhoun estimates, the telcos could replace every telephone switch for one-tenth that amount while radically upgrading the system.

In this cage of twisted copper wires writhe not only the executives of the telephone companies, but also the addled armies of telecommunications regulators, from the Federal Communications Commission and other Washington bodies to 50 state public utilities commissions and the towering hives of lawyers in the communications bar. The coils of copper also subtly penetrate the thought processes of MIT Media Lab gurus, libertarian lobbyists from the Electronic Frontier Foundation and myriad political analysts who see this massive metal millstone as a fell weapon of monopoly power. The copper colossus even intimidates scores of staunch Republicans who have arrived in Washington determined to extirpate every government excess, but who bow before the totem of universal service in their districts.

Like any socialist system, the copper colossus will die hard. But die it must.

Some 20 years ago, AT&T's long-distance lines comprised a similarly imperious cage of copper wires, installed over the previous 50 years and similarly impossible for rivals to duplicate. Then too, analysts termed telephony a natural monopoly because the system could handle additional calls for essentially zero incremental cost and because network externalities ensured that the larger the number of customers, the more valuable the system. These assumptions had led to government endorsement of the Bell monopoly as a common carrier committed to universal service.

Regulators, politicians and litigators always imagine that they can control the future of telecom, awarding monopoly privileges in exchange for various high-minded goals, such as universal or enriched services. But their actual role, as Peter Huber and his associates show in their new text, Federal Broadband Law, is mostly to promote monopoly at the expense of such values as universality, which ultimately depend not on law but on innovation. As a form of tax, regulations reduce the supply of the taxed output. It is technological and entrepreneurial progress, impelled by low tax rates and deregulation, that brings once- rare products into the reach of the poor, always the world's largest untapped market.

In this case, the decline and fall of the long-distance monopoly was not chiefly an effect of politics or litigation but of technology. Effectively dissolving the copper cage of long distance were the millimeter waves of microwave radio. Over the years, it turned out you could set up microwave towers anywhere and duplicate long-distance services at radically lower cost without installing any new wires at all. But this realization came woefully slowly to the regulators.

In the "above-890-megahertz" decision of 1959, made possible by new Klystron tubes and other devices that opened up higher frequencies to communications, the FCC permitted creation of private microwave networks. On the surface, it was a narrow decision affecting a few large corporations. But as AT&T planners noted at the time, it represented a clear break from the previous principles of common carriage, cross-subsidy and nationwide price averaging in the telephone network.

Sure enough, over the next two decades a cascade of further decisions climaxed with the authorization of MCI to emerge as a direct competitor to AT&T. Within less than a decade, MCI added to its panoply of aerial microwaves the yet more advanced technology of single-mode glass fibers. Issuing some $3 billion of junk bonds over a four-year period, MCI built the first nationwide network of advanced fiber optics. GTE made comparable investments in Sprint, and AT&T rushed to excel its new rivals. Combining microwave with fiber, long-distance telephony became a technologically aggressive and openly competitive arena; AT&T's monopoly was a thing of the past.

Today, the remaining monopolies in local phone service face a threat from radio technology still more devastating than the microwave threat to AT&T in long distance. As with microwaves, the government--in the name of preventing monopoly--dallied for decades before acting to allow elimination of the monopolies it had earlier established. After the invention of cellular at Bell Labs in 1947, some 34 years passed before the FCC finally began granting licenses for cellular telephony. By the 1980s, the FCC and Judge Harold Greene, managing the Modified Final Judgment breaking up AT&T, permitted limited competition in wireless telephony. However, the FCC allocated half the metropolitan licenses to existing RBOCs, which had no interest in using wireless to attack the local loop monopoly. The other licenses it assigned by lottery to gamblers and financiers with no ability to create an alternative local loop. The process of buying out the spectrum speculators required leading wireless carriers to hobble themselves with huge amounts of junk-bond debt. Although McCaw Cellular Communications created a robust national system, its financial structure prevented aggressive price competition with wireline service.

As a result, the idea persists that wireless telephony is an expensive supplement to the existing copper colossus rather than a deadly rival of it. The installed base of twisted-pair wire still appears to many to be a barrier to entry for new competitors in the local loop, rather than a barrier to RBOC entry into modern communications markets. The conventional wisdom sees the electromagnetic spectrum as a scarce resource. Few believe that it will soon emerge as a cheaper and better alternative to the local loop, in the same way that microwave emerged as a cheaper and better substitute for copper long-distance wires.

Making Waves
At the foundation of the information economy, from computers to telephony, is the microcosm of semiconductor electronics. It reaches out in a fractal filigree of wires and switches that repeat their network patterns at every level from the half-adder in a calculator chip or the SLIC in a telephone handset to the coaxial trees and branches of a cable TV system or the mazes of switched and routed lines in the global Internet. In computers, engineers lay out the wires and switches across the tiny silicon substrates of microchips. In telecommunications, engineers lay out the wires and switches across the mostly silicon substrates of continents and seabeds. But it is essentially the same technology, governed by quantum science and electrical circuit theory.

Semiconductor engineers may still spend more of their time with circuit theory, contemplating the operations of resistors, inductors and capacitors on currents and voltages in the device. But quantum theory is most fundamental, because it allows humans for the first time to manipulate matter from the inside--to control the conduction bands and energy-band gaps of the internal atomic structure of silicon and other elements, and to make electrons, holes and photons leap and lase at the behest of the designer. It is quantum theory that allows chip engineers to control with exquisite precision, gauged in tenths of microns and trillionths of seconds, the movements of electrons at the heart of electronics.

At the heart of quantum theory, however, is a perplexing duality. Most of contemporary physics seems to deal with particles--electrons, quarks, leptons, neutrons, protons. In 1994, for example, scientists at Fermilabs in Chicago announced "discovery" of the "top quark," which they described as the "last building block of matter." Yet these entities manifest themselves only in the midst of explosions in which their wave signatures can be identified. So-called quantum particle theory is unintelligible without quantum wave theory.

The elements of quantum physics intrinsically combine the characteristics of particles--definite specks of mass--with the characteristics of waves-- an infinite radiance of fields and forces. Entirely unlike particles, waves merge, mingle and mesh in vectors and tensors propagating boundlessly through space.

It is this paradoxical combination of the definite with the infinite that gives the microcosm its promise as a medium, not only for computation in one place, but for communications everywhere. Spectrum unfolds in a global ethersphere of interpenetrating waves that reach in a self- similar fractal pattern from the plasmas of semi-conductor lasers through the ethers of the planet.

Today, the telecosm of modern communications brings decisively to the fore the wave side of the quantum duality. Wires may seem more solid and reliable than air. But the distinction is largely spurious. In proportion to the size of its nucleus, an atom in a copper wire is as empty as the solar system is in proportion to the size of the sun. The atmosphere and wires are alternative media, and to the electron or photon are only arbitrarily distinguishable. Whether insulated by air or by plastic, both offer resistance, capacitance, inductance, noise and interference. In thinking about communications, the concept of solidity is mostly a distraction. The essence of new devices emerges more and more as manifestations of waves.

Whether in the air or in a wire, the electrons or photons do not travel; they wiggle their charges, causing oscillations that pass through the medium at close to the speed of light. As in waves of water, the wave moves, but the molecules of water stay in the same place. Thus belied is the analogy of particles or even bullets favored by physics teachers who give primacy to the mass rather than to the wave. Since the age of carrier pigeons and catapults, communications systems have transmitted masses only in the postal services.

Today, even in entirely stationary electronic systems, the wave action is increasingly dominant. The microchip itself--a Pentium processor, say--now runs at 120 megahertz, a rate in cycles per second that puts it in the middle of the FM radio band. New computers must pass the FCC requirements for radio emissions. Texas Instruments now advertises its 486 SXL-66 microprocessors as selling for under 50 cents per megahertz. Increasingly in the world of computers, people speak of bandwidth and cycles, reserving the discussion of mass chiefly for the batteries. The world of the telecosm is subtly shifting from electronics, with its implicit primacy of electrons, to what might be termed spectronics, seeing the particle as an expression of the wave rather than the other way around--moving from Bohr's atom and Heisenberg's electronic uncertainty to Maxwell's rainbow and Schrodinger's wave equation.

In a global marketplace increasingly unified by telecommunications at the speed of light, the vision of waves as fundamental affords not only a better image of physics, but also a better purchase on economic reality than a spurious search for solid states, physical resources, national economies and commodity products.

Conceived as some irreducible essence, the particle of mass, whether in the form of a top quark or Higgs boson, wire conduit or central switch, pushes our thinking about the world toward a vision of ultimately discrete and confinable entities, with electrons moving through the p-n junctions of microchips like so many steel ingots crossing a national border. Conceived, by contrast, as a continuous span of waves and frequencies, tossing and cresting, reflecting, diffusing, superposing and interfering, the telecosmic vision accords with the ever-rising global commerce in information services--ubiquitous, simultaneous, convergent, emergent.

To grasp the next phases of the information economy, one begins not with the atom or any other discrete entity, but with the wave. In 1865, in a visionary coup that the late Richard Feynman said would leave the American Civil War of the same decade as a mere "parochial footnote" by comparison, Scotch physicist James Clerk Maxwell discovered the electromagnetic spectrum. This spread of frequencies usable for communications is both the practical resource and the most profound metaphor for the global information economy.

Is it a domain of limits, to be husbanded by governments and appropriately allocated by auctions at a price of billions of dollars for a tiny span of wiggle rates? Is it beachfront property to be coveted as a finite and unrenewable resource? Is it a constricted domain to be exploited under the iron laws of diminishing returns? Is it a zero-sum game to erupt in Star Wars and street fights as satellite magnates and personal communications entrepreneurs crowd into a feudal fray of frequencies? At the heart of the gathering abundance of the information economy, would it sustain a new economics of scarcity?

So one might imagine from today's conventional wisdom. Contemplating these limits, diminishing returns and zero-sum economics at Richard Shaffer's Mobile Forum in March was industry guru Carl Robert Aron. He sees the world of wireless entering a "new ice age," like the recent ordeal of the tire industry in the face of radials. He predicts that customers, capital and revenues will become increasingly scarce and many species of company will become extinct. Offering a similarly grim vision, BellSouth Vice President of Corporate Development Mark Feidler declares that the price elasticity of demand for telephony is negative--you lower the price and revenues will sink. On the same panel, AT&T-McCaw executive Rod Nelson asserted that he could see no threat from personal communications services, because McCaw was already offering "a low-priced, high-quality service." Even Martin Cooper of ArrayComm saw spectrum as a limited resource sure to grow more valuable over time.

What would Maxwell say? As he discovered it, the spectrum is infinite, ubiquitous, instantaneous and cornucopian. Infinite wave action, not the movement of masses, is the foundation of all physics. It ushers in an age of boundless bandwidth beyond the dreams of most communications prophets. As industry guru Ira Brodsky concludes in his authoritative new book, Wireless: The Revolution in Personal Communications, "We are quickly moving from the era of spectrum shortage to the age of spectrum glut." This expanding wavescape is the most fertile frontier of the information economy. In its actions are the essential character of the coming economics of abundance and increasing returns.

In contemporary networks, as Nicholas Negroponte stresses in his best-selling book, Being Digital, all bits are fungible. In spectronics, all spectrum is fungible. In particular, the distinction between wireline and wireless service dissolves. A wire is just a means of spectrum reuse. Down adjacent wires, appropriately twisted or insulated, you can transmit the same frequencies without fear of interference or noise.

Using new digital radio technologies, such as code division multiple access or smart and directional antenna systems, you can similarly beam the same frequencies through the atmosphere, insulated by air. The chief difference is that the wire system costs far more to install and inhibits mobility.

The only wire technology commanding a decisive edge over wireless for critical applications is fiber optics. The intrinsic bandwidth of a fiber thread is nearly 1,000 times larger than the bandwidth of all the "air" currently used for terrestrial radio communications. In both media, capacity is largely governed by the need to avoid the water molecules that absorb many frequencies of electromagnetic
waves--in air, from humidity or precipitation; in fiber, from the unremovable residue of water in the structure of the glass.

Compared with perhaps 30 gigahertz of currently accessible frequencies in the air, every fiber thread can potentially bear 25,000 gigahertz. This huge bandwidth derives from the possibility of using infrared light frequencies for long-distance communications rather than radio or microwave frequencies. When you are dealing in terahertz (infrared light encompasses some 50 trillion hertz worth of frequencies between 7.5 X 10[11] and 3.5 X 10[14]), there is a lot of room for sending messages.

One fiber thread the width of a human hair can potentially use about 25 trillion of those hertz for communications (the rest tend to be fraught with moisture). This span is enough to carry all the phone calls in America on the peak moment of Mother's Day, or to bear three million six-megahertz high-definition television channels--all down one fiber thread the width of a human hair. As Paul Green sums it up, fiber commands 10 orders of magnitude greater bandwidth than copper telephone lines and 10 orders of magnitude lower bit-error rates. Optical engineers have packed as many as a million such threads in one bundle with a cross-section a centimeter square. Such feats plausibly support the assertion that, as a practical matter, spectrum is infinite.

The capacity of fiber is so large that the best way to think of it is as a radio system in glass--a fibersphere that can potentially accommodate as many as 10,000 separate wavelength bitstreams. Under a system called wavelength division multiplexing, users will tune in to a chosen frequency band in the same way they currently time in to a chosen radio or television channel, whether in the air or in a coaxial cable. Indeed, engineers can take the same infrared frequencies used in fiber and move them to the air for shorter distance applications such as local-area networks, point-to-point connections between buildings, links between handheld computers and desktop hosts, and even television remote controls. As tunable laser transmitters and photodiodes, along with other optoelectronic gear, become more sensitive and efficient, airborne infrared will become more robust and useful. Experiments by the Israelis with ultraviolet frequencies suggest that even these superhigh frequencies above visible light might someday be used for communications through the atmosphere (offering tens of thousands of TV channels, for example).

Now the FCC has auctioned off 120 megahertz of frequencies for personal communications services. The most prominent winning bidders were consortia led by Sprint, TCI, Comcast and Cox (a long-distance carrier and three cable companies going under the name Wireless Co.); by AT&T; and by AirTouch, Bell Atlantic, NYNEX and U S West as PCS PrimeCo. Most analysis has focused on what is called the wireless market and has assumed the major competitor to PCS to be the current cellular companies. Aron's ice-age ruminations stemmed from contemplation of this radical increase in competition for a limited number of cellular customers who currently cost some $540 each to sign up (counting handset subsidies) and whose per-capita revenues are declining at a pace of some 8% per year. Remember BellSouth's Feidler's vision of a negative elasticity of phone markets, meaning that lower prices bring lower revenues?

From a spectronic perspective, all this analysis is deeply misleading. Whether channeled down wires or through the air, spectrum is spectrum. Digital wireless is a cheaper and better way of delivering service. The market for PCS is not the cellular customer, but the one billion wireline customers in rich countries and the several billions of potential phone and teleputer customers around the globe. In pursuing these customers, the price elasticities will be dramatically positive, with various price points reachable with new wireless technologies releasing torrents of new demand and new revenues. What Aron calls an ice age will in fact prove to be a gigantic global warming, unleashing huge new growth in telephony, using spectrum in all its various forms (except perhaps the twisted-pair copper wires that currently dominate the installed base of the industry).

The winning bidders from AT&T and Sprint did not put up their $3.7 billion in order to join a zero-sum straggle for new cellular customers. These bidders are dominated by long- distance businesses that can use PCS to reduce their some $30 billion in access charges to the local exchange carriers by creating an alternative local loop. Similarly, MCI, though avoiding the auction, created a subsidiary called MCI Metro that may seek to manage service for spectrum winners in 17 cities, again harvesting the benefits of obviated access charges. Then all these companies can use their PCS technologies to pursue customers around the world without any thought of wire.

A chart created by industry analyst Herschel Shosteck illustrates the opportunity. The Shosteck chart is a bell curve relating the incomes of the world's households to telephone penetration rates. He shows that telephony has so far penetrated only to countries representing the top tail of the curve, where national wealth suffices to reduce the cost of telephony to a threshold of between 4% and 5% of incomes. As incomes rise around the globe, more and more people cross the telecom threshold. A chart of GDP in real dollars per capita versus telephone penetration shows that a 40% rise in incomes could bring a 1,600% increase in potential customers.

Compounding the surge in incomes, however, will be the plummeting cost of wireless telephony. Shosteck estimates that between 1985 and 1994 the price per customer dropped 80%, from $5,000 to $1,000. Combining these two trends, he calculates that there will be between 400 million and 800 million new wireless subscribers by the end of the year 2000. These numbers represent an awesome upsurge from the world's current level of some 60 million cellular customers. Any further acceleration in income growth or decline in telephone prices will increase these numbers. A 50% further drop in telephone prices combined with a 50% rise in incomes would quickly thrust the vast bulk of the world's population above the Shosteck threshold. Far from the negative elasticities that U.S. phone executives see in their saturated wireline voice business, the world-wide communications market will be a financial trampoline.

Just Chips And Antennas
In an ordinary industry, a 50% drop in price seems a major obstacle. But telephony is becoming a branch of the computer industry, which doubles its cost effectiveness every 18 months. The wireless convergence of digital electronics and spectronics will allow the industry to escape its copper cage and achieve at least a tenfold drop in the real price of telephony in the next seven years.

Sen. Stevens should meet Martin Cooper, a former research chief at Motorola and now CEO of ArrayComm. Located in San Jose, ArrayComm is devoted to drastically reducing the cost of telephone access over the next two years while entirely obviating the problems of twisted-pair wiring that afflict Alaska.

The current pitch of ice-age cellular providers is "pay more and get less. . .and don't even think about universal service." Although they claim penetration rates in industrial countries of nearly 10%, most cellular users make most of their calls on wireline systems. The real market share of cellular is in fact under 1% in the industrial world. The cellular companies' formula for success is to exploit the public hunger for mobility by charging more money for worse service--extracting premium prices for calls with acoustics and reliability far inferior to wireline telephony. Followed by both sides of the cellular duopoly--by Bells, McCaws and other suppliers--this pay-more- for-less-and-worse formula has concealed from much of the industry the basic technological fact that wireless will soon be acoustically better than wireline and drastically cheaper as well. As the CD example shows, after all, digital sound systems are superior to analog. And without wires, phones finesse the largest capital and labor costs of conventional telephony.

In economic terms, the intrinsic cost advantage of wireless is concealed by the colossal installed base of copper. Already mostly paid for and largely written off, the 154 million twisted-pair access lines will allow the Bells to compete in price for some time with wireless rivals that have lower real costs.

Nonetheless, technical reality will prevail in the end. Spectronics offers technologies in four dimensions for dividing and conquering spectrum: Frequency division, time division, code division and space division. All address in various ways the issue of frequency rouse--how many times in a system particular frequencies can be roused without causing interference in other calls using the same frequencies. Of the four techniques, so far only frequency division has been widely exploited. As these other methods come on line, the cost of telephony will go over the same kind of digital cliff long familiar in computers.

Surveying all these proposed schemes and their promised upgrades (see sidebar next page), it is safe to project between a 60% and 90% drop in the cost of wireless telephony over the next five years, depending mostly on the progress of CDMA. Qualcomm's CDMA could reduce costs tenfold, compared with the threefold gains from current global services mobile (GSM) technology, which contemplates an upgrade path chiefly through downgrading the voice quality with a half-rate vocoder.

All these gains in wireless efficiency from dividing by time, code or frequency are compounded by dividing spectrum by space. Mathematically, every 50% reduction in the cell radius yields a 400% increase in the number of customers who can be served in a given area with a given technology. Huge theoretical gains accrue from cell-splitting--reducing the physical extent of cells and multiplying their numbery--converting current macrocells as large as 35 miles in diameter into microcells a mile or so in width, and into picocells measured in hundreds of yards in buildings, shopping centers or congested urban streets.

All these gains, however, could be nullified by the expense and difficulty of implanting base stations all over cities and neighborhoods. The key to the gains of space division, therefore, is creation of base stations drastically cheaper, smaller, more discreet and more functional than the current cell sites, costing between $500,000 and $1 million, occupying 1,000 square feet and containing between 55 and 416 radios, depending on the frequency reuse factor. The most notable breakthrough in base stations is the Steinbrecher MiniCell, to be demonstrated in July and launched at the end of the year.

Putting a base station into a briefcase, Steinbrecher uses a single broadband digital radio to perform the functions of between 55 and 416 analog transceivers. The key breakthrough is a proprietary mixer that can flawlessly down-convert all the waveforms in the entire cellular spectrum into a stream of baseband digital bits without losing any information or introducing spurious signals. Containing all the electromagnetic contents of the cell, this digital bitstream is broken into channels by a 0.4- micron technology application-specific integrated circuit and is interpreted by digital signal processors. Governed by the learning curves of semiconductors, the MiniCell promises to reduce the cost of a cell site by an initial factor of 10 and by an eventual factor governed chiefly by Moore's Law.