radio revolution: the coming age of unlicensed wirelesswerbach.com/docs/radiorevolution.pdf ·...

The Coming Age of Unlicensed Wireless

By Kevin Werbach

Radio Revolution

N E W A M E R I C A F O U N D A T I O N

P U B L I C K N O W L E D G E

NEW AMERICAF O U N D A T I O N

1875 Connecticut Ave. , NWSuite 650

Washington, DC 20009Phone: 202-518-0020 • Fax: 202-986-2539

PUBLICKNOWLEDGE.ORG

1630 Connecticut Ave. , NW7th Floor

Washington, DC 20009Phone: 202-986-2700 • Fax: 202-986-3696

W W W. N E WA M E R I C A . N E T

Radio RevolutionThe Coming Age of Unlicensed Wireless

By Kevin Werbach


Washington, DC

Author

Kevin Werbach is the founder of the Supernova Group, a technology analysis and consulting firm. He advises organizations on the strategic and legal implications of emerging trends in communications,digital media and software. As Counsel for New Technology Policy at the Federal CommunicationsCommission from 1994 to 1998, he helped develop the United States Government's e-commerce policy, shaped the FCC's approach to Internet issues, and authored Digital Tornado, a seminal analysis of the Internet's impact on telecommunications policy. He has also served as editor of the influential publication Release 1.0. His writing has appeared in Harvard Business Review, Fortune, Wired, HarvardLaw Review, Slate, Red Herring, and Business 2.0, among other publications, and he appears frequently as a commentator in print and broadcast media.

Radio Revolution is the second publication Kevin Werbach has authored for the New America Foundation.His Working Paper, "Open Spectrum: The New Wireless Paradigm" was published in October of 2002and is available at www.spectrumpolicy.org.

The New America Foundation is a non-partisan, non-profit, public policy institute in Washington, D.C. Relying on a venture capitalapproach, the Foundation invests in outstanding individuals and ideasthat transcend the conventional political debate. Through its Fellowship

Program and Strategic Initiatives, New America sponsors a wide range of research, published writing, conferences, and events. New America’s Spectrum Policy Program advocates a more fair, efficient, anddemocratic allocation of the public airwaves. Many additional publications on this topic, including New America’s Citizen’s Guide to the Airwaves, can be found at www.spectrumpolicy.org.

Public Knowledge is a public-interest advocacy organization dedicatedto fortifying and defending a vibrant information commons. This

Washington, D.C.-based group works with a wide spectrum of stakeholders – libraries, educators, scientists, artists, musicians, journalists, consumers, software programmers, civic groups, and enlight-ened businesses – to promote the core conviction that some fundamental democratic principles and cultural values – openness, access, and the capacity to create and compete – must be given new embodiment in the digital age.

Contributors

Nigel Holmes, who is principal of Explanation Graphics, www.nigelholmes.com, created four originalillustrations for this report; two of his illustrations from the Citizen’s Guide to the Airwaves are reprintedhere as well. In addition, Donald Norwood Design created the layout and design of the report.

Matt Barranca, a Program Associate at the New America Foundation, wrote the WISP profile side-bars. The Acoustic Analogy sidebar was adapted from New America’s forthcoming “The CartoonGuide to Government Spectrum Policy: What if the Government Regulated the Acoustic Spectrumthe Way it Regulates the Electromagnetic Spectrum?” by J.H. Snider. Hannah Fischer led the copy-editing and production efforts and was assisted by New America’s Michael Calabrese, J.H. Snider,Matt Barranca, and Max Vilimpoc. Spectrum experts Dewayne Hendricks, Anthony Townsend,Patrick Leary, and Mark McHenry provided valuable feedback for some of the technical content of this report.

Acknowledgments

We thank the Ford Foundation, the Open Society Institute, the Arca Foundation, and the JoyceFoundation for their support for New America Foundation’s Spectrum Policy Program. Withouttheir support, this report would not have been possible.

This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlikeLicense. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/1.0/ orsend a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.


Part I: Introduction ........................................................................................1Wireless Fundamentals ...................................................................................2Believe in Magic ..............................................................................................2

Part II: Wireless Fundamentals.............................................................5Basic Concepts .................................................................................................5The Role of Government ................................................................................11

Part III: Paradigm Shift: From Static to Dynamic.................13The Traditional Approach .............................................................................13When the Devices Get Smart ........................................................................14Survey of Dynamic Wireless Techniques......................................................16Implications of Dynamic Approaches ..........................................................19WiFi as a Case Study.....................................................................................22

Part IV: The Unlicensed World ............................................................25The Spectrum of Spectrum-Use Regimes .....................................................25Current Unlicensed Products ........................................................................27Success Stories ..............................................................................................30

Part V: Future Scenarios ..........................................................................37Expanding the Space of Possibilities...........................................................37The Last Wireless Mile ..................................................................................39Interoperable Public Safety Communications.............................................40Adaptive Mobile Phones ...............................................................................40Personal Broadcast Networks .......................................................................41

Part VI: Policy Recommendations.....................................................43

Part VII: Conclusion ....................................................................................47

Bibliography .............................................................................................................49

Endnotes.....................................................................................................................51

Contents

Introduction

These are not just incrementaladvances. The fundamental assumptionsgoverning radio communication sinceits inception no longer hold. The staticwireless paradigm is giving way todynamic approaches based on cooperat-ing systems of intelligent devices. It istime for policy-makers to consider howregulation should change in response.

The radio revolution is the singlegreatest communications policy issue ofthe coming decade, and perhaps thecoming century. The economics ofentire industries could be transformed.Every significant public policy challenge

could be implicated: competition; inno-vation; investment; diversity of program-ming; job creation; equality of access;coverage for rural and underserved areas;and promotion of education, health care,local communities, public safety, andnational security. Yet the benefits of theparadigm shift are not guaranteed.Exploiting the radio revolution willrequire creativity and risk-taking by boththe private and public sectors. At everystep, there will be choices between pre-serving the status quo and unleashingthe forces of change. The right answerswill seem obvious only in hindsight.

The only way to appreciate theopportunity before us is to comprehendthe fundamentals of radio communica-tion, and the profound ways they arechanging. For all its significance to dailylife and economic activity, wireless tech-nology is poorly understood. This paperseeks to explain the established “static”wireless paradigm, the emerging“dynamic” alternative, and the implica-tions of the coming revolution.

1

We stand at the threshold of a wireless paradigm shift. Newtechnologies promise to replace

scarcity with abundance, dumb termi-nals with smart radios able to adapt totheir surroundings, and government-defined licenses with flexible sharing ofthe airwaves. Early examples suggestthat such novel approaches can provideaffordable broadband connections to awide range of users.

Wireless FundamentalsWireless. The very word belies its significance.Wireless communication is defined by what it isnot, like the horseless carriage or the fat-freemuffin.1 Yet the real value of a satellite televisionbroadcast, a WiFi connection to a laptop, or amobile phone call from your car to your motherisn’t the absence of dangling wires. Mobility,portability, ubiquity, and affordability are allenhanced when signals pass through the air rather

than through strands ofcopper or optical fiber.Talking on a mobilephone is different, and inmany ways better, thanusing a landline connec-tion. If it weren’t, morethan one billion peoplewouldn’t have signed upfor mobile phone service,despite the alternative of acentury-old wired phoneindustry.

Wireless communica-tion is the foundation ofindustries generating hun-dreds of billions of dollarsin revenue and sellinghundreds of millions of

devices every year. It is crucial to how we commu-nicate, work, learn, entertain ourselves, accesshealth care, and protect our nation. It is also heavily regulated everywhere in the world. Governments today face critical decisions con-cerning the future of wireless communication. Isthere a “spectrum shortage,” and if so, how can itbe alleviated? Should more spectrum be set asidefor “unlicensed” uses? Should spectrum licenseesbe given property rights to resell or otherwisecontrol their spectrum more thoroughly? Do weneed different rules to deal with interference?Should new “open spectrum” technologies beallowed to “underlay” or “interweave” with exist-ing licensed services? Can government, military,and public safety spectrum be managed moreeffectively? These questions will shape the com-munications environment of the 21st century.

“The existing legal

and policy framework for

spectrum management

has not kept pace with

the dramatic changes

in technology and

spectrum use.”

— WHITE HOUSE MEMO TO

FEDERAL AGENCIES, JUNE 2003

Unfortunately, wireless communication remainsdeeply misunderstood and under-appreciated. Basicconcepts like spectrum and interference sufferfrom widespread misconceptions. Technologicaldevelopments of recent decades have not pene-trated the public consciousness, even as the fruitsof these developments become part of daily life forhundreds of millions of people. The great para-digm shift from static to dynamic wireless commu-nication has barely registered in business andpolicy circles. Just as economists know that infor-mation technology must have a role in productivitygrowth but have trouble finding it in their statis-tics, the wireless industry is experiencing a trans-formation that even many of its own experts do notfully appreciate.

Believe in MagicIn the words of legendary science fiction authorArthur C. Clarke, “any sufficiently advanced tech-nology is indistinguishable from magic.”2 Wirelesscommunication is a form of magic. Words and pic-tures fly over invisible pathways with near instanta-neous speed. We control devices at a distance, withno apparent means of connection. Scores of sig-nals, carrying many different types of messages,traverse the air simultaneously. A time travelerfrom the Middle Ages would surely see divineintervention – or witchcraft – all around.

Yet for us, wireless communication is a familiarform of magic. It drives the radios we have had inour homes since our grandparents’ day, the mobilephones that many of us use to communicate, thetelevisions we watch an average of seven hourseach day, the remote controls that start those TVs,and even the throwaway boxes that open ourgarage doors. This familiarity breeds contentment.We think we understand how wireless communica-tion works. We don’t.

Our intuitions about wireless, by and large, aremistaken. They are based on outdated technologiesand inaccurate analogies. If we hope to move forward in exploitation of the airwaves, we musttake a step back. We must understand wireless com-munication for what it really is. And then we mustre-evaluate our assumptions about what it could be.This paper presents a set of analogies to help

2

Introduction

explain the basic physics of radio, and the radicalshift that emerging technologies represent. Thestrangeness of wireless communication vanisheswhen we see that it is no different than acousticcommunication, otherwise known as speech.

Paradigm shifts are both difficult and essentialfor progress.3 Copernicus and Galileo showed thatthe Earth revolves around the sun, contrary to theperceived wisdom of the day. Eventually their viewprevailed, launching an age of extraordinary dis-covery. In the last century, quantum mechanicsoverthrew the long-established Newtonian world-view. A hundred years of subsequent physics exper-iments confirm that our universe contains no suchthing as solid matter or definite cause and effect.These ideas are so weird that most of us simplyrefuse to accept them. We live in the familiar classical environment of our commonsense aware-ness. At the same time, we blithely accept tech-nologies such as the integrated circuit and the laser,which could not exist without the scientific fruits of the alien quantum world.

The new dynamic paradigm reveals that wirelesscommunication is more magical than we assume.More than one service can occupy the “same”spectrum, in the same place, at the same time. Thefrequencies that now carry one signal could some-day carry thousands...or billions. There could be asmany video broadcasters as today there are mobilephone subscribers. Government could cease thefrustrating and inefficient task of parceling outspectrum, and instead allow users to share the airwaves without licensing. Broadband Internetconnections could be far more ubiquitous andaffordable. Innovation could proceed by leaps andbounds rather than a hesitant, drawn-out shuffle.

Appreciating the potential of wireless technologyhas always been difficult. When Guglielmo Mar-coni invented the radio, he envisioned it being usedfor person-to-person communication, not one-to-many broadcasting. Alexander Graham Bellinvented the telephone while developing tools tohelp deaf people, and thought it would be used to

Radio Revolution

broadcast music concerts. If these scientific giantscould be so wrong about their own creations, mightwe not be wrong in our assumptions about wireless?

This is not mere idle speculation. Decisionsmade in the 1920s to zone spectrum by serviceand to assign exclusivelicenses to users havedefined the contours ofwireless communicationever since. A huge marketsits atop the existing regulatory framework,which in turn sits atopconceptual and technicalassumptions. Alter thoseassumptions, and we canalter the framework. Alterthe framework, and themarket could becomesomething far greater thanit is today. Maintain thestatus quo or worse, andthe opposite might result.

“Unlicensed” wirelesscommunications systemsare the manifestation ofthe dramatic change fromthe static to the dynamicparadigm. The word unlicensed, like the wordwireless, emphasizes what is missing rather thanthe true significance of the concept. What is soextraordinary about unlicensed devices is whatthey can do, and the incentives they create for innovation and growth. Already, wireless Internet service providers (WISPs) and non-profitcommunity networks are using unlicensed systemsto deliver broadband connectivity where it wasotherwise unavailable. Several are profiled in sidebars throughout this paper. In the future, unlicensed systems may support more significantnew communications scenarios, which are detailedin the last section.

3

A huge market sits atop

the existing regulatory

framework, which in turn

sits atop conceptual and

technical assumptions.

Alter those assumptions,

and we can alter the

framework. Alter the

framework, and the

market could become

something far greater

than it is today.

Wireless Fundamentals

Basic ConceptsThe Secret Life of Radio WavesA wireless communications systeminvolves one or more transmitters andone or more receivers. There is nothingin the middle. Transmitters radiate, andreceivers receive, within a certain rangeof frequencies known as the radio fre-quency spectrum. However, these areproperties of the equipment, not somedistinct medium the signals passthrough. By the same token, what gov-ernments regulate are the capabilities oftransmitters, and to a lesser extentreceivers, rather than the spectrum itself.

Radio waves are a form of electro-magnetic radiation, like lasers or light-ning bolts. “Radio frequency” signalsoscillate at frequencies between about 3kilohertz (kHz) and 100 gigahertz(GHz). Their propagation characteris-tics are well-understood by physicists. Infree space, radio waves can propagateindefinitely, with declining power overdistance, unless dissipated by obstacles

such as walls or the Earth’s atmosphere.Their susceptibility to such obstaclesdepends on the frequency, bandwidth,and power of the transmission.

The point of this physics lesson isthat most of the topics spectrum policyconcentrates on, such as “interference”and “spectrum,” are value judgmentsbased on our uses of wireless communi-cation. Radio waves do not bounce offone another, or cancel each other out.When two or more signals share thesame space at the same time, it can bedifficult for receivers to distinguishthem, just as the human ear has difficulty focusing on two simultaneoussound sources. In practical terms, theTV picture gets fuzzy or the mobilephone drops a call (see Figure 1).

We call this interference. The effect isidentical to what happens when you tryto listen to a radio broadcast and a CDat the same time. The sounds still reachyour ear, but you may have trouble sort-ing them out.

5

6


The Acoustic Analogy*

The sound waves used in ordinary speech are analo-gous to the radio waves used in wireless communica-tion. Both are radiation employed to send messagesbetween transmitters and receivers. The acoustic spectrum involves lower frequencies than the radiofrequency spectrum, but this has no effect on thephysics involved. Our ears are tuned to pick upacoustic waves, just as radio receivers are tuned toreceive radio waves. And our vocal chords produceacoustic waves, just as radio transmitters produceradio waves.

The major difference between acoustic and radiocommunication is that humans and other animalshave evolved exquisitely sophisticated tools for encod-ing and interpreting speech. Our vocal apparatus andears are magnificently precise yet highly adaptive.Standing behind them is the human brain, the mostpowerful computing device ever created. Our brainscan pick out sound waves from the surrounding background noise and quickly interpret them with phenomenal accuracy. We take all this for granted,because speech is so basic to our very existence.

Radios have historically been far less intelligentthan the human systems forvoice communication. In partic-ular, radio receivers are simpledevices that tune to a specificfrequency. Because radio deviceshaven’t been smart enough to dis-tinguish among overlapping sig-

nals, government regulates the electromagnetic spectrum to minimize interference in ways that wouldbe inconceivable for acoustic communications.

Dynamic wireless systems are closing the gapbetween acoustic and radio communication. Newerdevices employ sophisticated computer processing toencode and decode wireless signals. They also employcooperative techniques and adaptive mechanismsthat bring to mind the social behavior of humanbeings. The more wireless systems can discriminatethe way the human ear does, the less regulation isneeded to avoid confusion.

Imagine a crowd of people at a football stadium.4

Though thousands of them are talking at the sametime, many of them screaming at the top of theirlungs, there is no need for regulation to ensure effec-tive communication. There isn’t even a need for rulesto ensure that the public address system can be heardover the thousands of independent voices. A privateregime of property rights to speak in the stadium isjust as superfluous as a government licensing system.

Our mouths and ears are sufficiently adaptive toseparate signals from noise when both parties aretrying to communicate. This despite the fact that theacoustic spectrum is far narrower than the radio

spectrum, and our biologicalsenses are much less precisethan today’s digital communi-cations devices. When transmit-

ters and receivers are as smart ashumans, the best rules to prevent

interference are no rules at all.

No talking in this stadium!Violators will be thrown in jail!

* Adapted from J. H. Snider, “The Cartoon Guide to Federal Spectrum Policy: What if the Government Regulated the Acoustic Spectrum the Way it Regulates the Electromagnetic Spectrum,” New America Foundation, Forthcoming.

Interference among wireless systems sounds similarto what happens when a landline telephone callgenerates an “all circuits are busy” message. Inreality, the two situations are quite different. In thelandline case, the connection literally stops at anoverloaded phone company switch. The callreaches that point and goes no farther. In wireless,the signal keeps on going. Only the useful informa-tion is lost, not the actual radio waves.

This seemingly arcane distinction is critical. Forthe blocked phone switch, nothing the caller or thecalled party can do makes any difference. The elec-trical or optical signal terminates in the middle ofthe network. In the wireless case, the signal gets toits destination but cannot be understood. If thetransmitter or receiver were smarter, the same signal might be intelligible. Better technology atthe endpoints can reconstruct useless noise backinto useful information.

In other words, change the communicationsdevices or the regulatory environment, and youchange the capacity of the system. Therefore, anystatement about interference or spectrum scarcityassumes a particular set of technical and regulatoryconditions.

CapacityAs the previous section demonstrates, interferencematters because of its implications for capacity.Capacity is the essential metric for wireless com-

Radio Revolution

munications.5 Only so many radio stations,TV channels, phone conversations, orInternet connections can successfully operate at once. However, this number isnot fixed. Marconi, the inventor of radio,originally thought that only one signalcould be transmitted in a given geographicarea, because other radios would interferewith it. He later developed a technique foradding capacity based on the principle thattuning forks can be made to vibrate at thesame frequency across distances. Using thismodel, one radio signal can be associatedwith a carrier wave of a particular fre-quency, and additional radios on differentfrequencies can operate in the same area.

In effect, Marconi figured out how to use frequency to multiplex radio signals. Each stationgot its own unique frequency: hence the familiarradio call numbers like 102.7, 88.5, or 97.1. Becausefrequency division was the only viable means ofoperating multiple simultaneous transmitters whenradio developed as a commercial service, it becamethe basis for government radio policy. Regulatingradio meant regulating frequencies, by parceling out the usable spectrum to licensees and service categories. And so it remains today. We don’t usethe same numbers to identify TV channels ormobile phone networks, but these systems areassigned frequencies in a similar way.

Frequency division, however, is not the onlymeans of multiplexing radio signals.6 Another possibility is time. The government could haveallowed each broadcaster to transmit only during a certain hour of the day, for example. Frequencydivision was obviously a better solution, both oncapacity and practical grounds. In other cases,though, time division makes sense. Some mobilephone systems, for example, chop up their licensedfrequencies into split-second time slots, and interweave digital communications signals amongthem.7 In addition to time and frequency, spatialmultiplexing can be done based on the three-dimensional relative location of the transmitter and the angle at which a signal hits an antenna. But again, spectrum regulation talks primarilyabout frequencies.

7

FIGURE 1 – NOTIONS OF INTERFERENCE: When two or more signals

share the same space at the same time, some receivers have difficulty

distinguishing them and interference can result.

These multiplexing techniques, along withimprovements in tuners and signal processors, arethe reason the radio spectrum can now accommo-date services such as television, mobile telephony,satellite radio, and wireless Internet connectivitywhere once there was only radio. Instead of theoriginal two mobile phone competitors in eachmarket, for example, we now have six national carriers in the U.S. The “usable” spectrum of frequencies is five thousand times wider than it was in 1927, when the Radio Act was passed.

Have we now exhausted the capacity of the spec-trum? Yes and no. If we only focus on frequencies,virtually every useful band has been allocated to

one use or another. Any new service, such as digitaltelevision or third-generation (3G) mobile phonenetworks, requires that incumbents be “cleared”out of existing bands to make room. On the otherhand, if we actually sample the spectrum to test forsignals, we get a very different picture. Most of thespectrum is empty in most places most of the time.There may be a rectangle filled in on a frequencychart, but there is no detectable signal actuallyusing that frequency8 (see Figure 2).

Many advances in wireless technology make useof excess capacity that shows up in the real worldbut not on the official chart. Because the frequencychart represents official boundaries, however, these

8


(see note* at bottom of page)

100megahertz (MHz) 200 300 400 600 700 800 900

500MHz

1 GHz(1,000 MHz)

CBradio

SFM radio

S

BroadcastTV

BroadcastTV

BroadcastTV

Cordlessphones

Remote-controlledtoys

Car alarms

Garage dooropeners

Medicalimplants

FamilyRadioService(walkietalkies)

Wirelessmedicaltelemetry

Mobilephones Highway

toll tags*

Short-wave

WCordlessphones*

C

Radio:AM

BroadcastTV

*overlappinguse

Pagers

Mobilephones

3

FIGURE 2 – TWO VIEWS OF CAPACITY: The top image, from The Citizen’s Guide to the Airwaves, depicts allocations

on the low-frequency, broadcast bands. The bottom image represents the actual usage of the most active channels

of the broadcast bands, as recorded in New America Foundation’s spectrum usage measurements taken during peak

hours in the highly populated, Dupont Circle area of Washington, DC.

technologies can only be used in certain frequencybands. Between the view of spectrum as filled upand the reality of its emptiness lies an almost incalculable opportunity.

ArchitectureEven the gaggle of capacity-enhancing techniqueslisted earlier is incomplete. Capacity depends notjust on the way a device distinguishes one signalfrom another, but on the design or “architecture” ofthe overall communications system it is part of.Two systems with the same “amount” of spectrummay have very different capacity profiles. Whatmatters is the use that can be made of the spectrum,not just the abstract width of the frequency band.

What does architecture mean in this context? Asimple example would be to compare a radio broad-cast with a mobile telephone call.9 Radio is a broad-cast service, meaning that a tower sends out a signalat the maximum allowed power in all directions. Itblankets an area, so that every receiver within range(typically a metropolitan area) can tune in the signal.

Mobile phone networks, by contrast, use a cellulararchitecture. Each tower sends relatively low-power

Radio Revolution

signals to any handset within a few square miles. Forhandsets out of that range, there are other towers.Because of the low power, users talking to one towerdon’t notice the signals that other users are exchang-ing at the same time with a different tower. Thesame spectrum is being “reused” (see Figure 3).

9

Shannon’s Law

Bell Labs researcher Claude Shannon, in his seminalwork in the late 1940s, created the mathematicalconcept of information theory. Shannon developedequations to measure the ability of a communicationschannel to carry useful information. As usuallyexplained, “Shannon’s Law” defines a maximumcapacity, which is proportional to the frequency orbandwidth of the channel. As a result, bandwidth hasbecome almost synonymous with capacity. Becausethere is only so much bandwidth, and any service canonly have a small portion of it, this model impliesstrict limits on possible capacity.

This paraphrase of Shannon’s capacity theorem leavesout critical facts. The version in question describes thesimplest possible case – one transmitter and one

receiver. Add more of either, and the solution is nolonger so easy. More devices may cause “interference,”or may create opportunities to interweave signals or oth-erwise add intelligence to the network. Similarly, a widerband adds capacity, but it’s not the only way to do so.

The scientific field of multi-user information theorytakes Shannon’s work and extends it for these morecomplex (and more realistic) situations. It is a fruitfularea of research and experiment. In the half-centurysince Shannon’s initial work, we have learned manythings about what is possible with wireless communi-cations. But there are many things we still do notknow. For example, we don’t even know the maxi-mum potential capacity of a system with an arbitrarynumber of devices. This core uncertainty should makeus hesitate before uttering any statements aboutwhat is or is not possible in the wireless realm.

FIGURE 3 – COMPARING BROADCAST VS. CELLULAR

ARCHITECTURE: The broadcast architecture (left) services

more users over a larger area with one signal, while the

cellular architecture (right) lets more users receive and

send distinct transmissions.

Notice the distinctions. The broadcast modellets the transmitters and receivers be simple (andtherefore cheap), because there is only one trans-mitter sending data in one direction. The cellularmodel requires many more towers and moreexpensive devices, but in return it lets many moreconversations occur simultaneously, in both direc-tions. The broadcast system services more userswith one signal, but the cellular system lets moreusers receive separate and distinct transmissions.There are other significant differences, and othernetwork architectures with their own characteris-tics. Which architecture is chosen depends ontechnological capabilities and the service beingoffered. However, the way the legal regime dividesand regulates spectrum defines the range of possi-ble architectures to choose from.

Like capacity multiplexing techniques, wirelessarchitectures have evolved over time. Cellular

phone service was first conceived in the 1940s, butcouldn’t have been deployed economically at thattime even if there were no regulatory hurdles.Today, thanks to massive increases in computingpower and miniaturization of digital devices, manyarchitectures are possible that once were not evenimaginable. Asking about capacity or interferencewithout considering architecture and the tradeoffsit enforces is like asking how many people can livein a city. It depends.

LayersThe third fundamental concept for understandingwireless systems has traditionally not been part ofthe wireless world at all. Wireless networks havehistorically used integrated, special-purposedevices. A short-range data connection in a televi-sion remote control does one thing; a microwaverelay for sending long-distance telephone signalsacross the country does something else. This self-contained approach mirrors that of pre-digitalwired networks. Telephone systems had little incommon with cable TV systems, for example.

In contrast, data networks tend to operate usinga layered model.10 Rather than defining the entiresystem as an integrated whole, engineers split it upinto a communications “stack.” This allows forseparation of functions that can be optimized indi-vidually. For example, technology developed fortelephone networks to encode voice signals can beapplied to different “physical” layers, includingcable TV networks and wireless environments. Intoday’s digital world, every signal is just a stream oftheoretically interchangeable digital bits. Cable andphone companies can compete to offer voice,video, and broadband services, even though theirnetworks use different types of wires.

Layering has several benefits. It makes systemsmore flexible, allows innovation in one area to benefit systems elsewhere, opens up new possibili-ties for competition, and gives users “best of breed”solutions rather than bundles which may not meettheir needs.

Network engineers often use a seven-layer model published by the International StandardsOrganization.11 A simplified version for policy purposes includes four layers, from bottom to top:

10


FIGURE 4 – LAYERS: In the digital era, wireless signals con-

sist of a stream of interchangeable digital bits. Engineers use

a layered framework to provide order and optimize and man-

age transmissions.

Physical(WiFi, OFDM, UWB, etc.)

Logical(Addressing, meshed routing, handoffs, etc.)

Applications(Voice, video distribution, Web, etc.)

Content

physical, logical, applications, and content. Thewireless technologies described in this paper operate at the bottom two layers. The physical layerinvolves how information is actually sent across the communications channel. The logical layer concerns addressing and management functionsthat ensure the right information gets to the rightplace efficiently (see Figure 4).

The Role of GovernmentWhat is “possible” in communications always has two meanings: technical and legal. Technicalpossibility is a function of scientific discovery andcommercialization. Legal possibility is defined bythe regulatory system. Many things are possible in the technical realm but not in the legal.(Occasionally, the reverse is true!) No discussion of the business and social fundamentals of wirelesstechnology would be complete without taking government rules into account.

From its earliest days, the communications indus-try has been subject to pervasive government regula-

Radio Revolution

tion, in the U.S. and elsewhere. In wireless, that reg-ulation takes the form of spectrum policy. TheFederal Communications Commission (FCC) tellssome entities that they can use particular frequen-cies, usually for specific purposes and with detailedtechnical and economic requirements. It tells every-one else that they cannot communicate on those frequencies. And it enforces those rules, punishingviolators. The National Telecommunications andInformation Administration (NTIA) manages thefederal government’s use of spectrum, and serves asthe principal advisor to the President for spectrumpolicy formulation.

Extensive government regulation of spectrum is taken for granted. But let us ask for a moment,why should that be so?12 Communication over the airwaves is speech, just like communication throughthe wired phone network or through a microphoneat a political rally. Government regulation of speechis strictly limited under the Constitution. Yet wetolerate a government agency, the FCC, bestowingthe ability to speak upon individual companies,

11

FIGURE 5 – NTIA SPECTRUM ALLOCATION CHART: The National Telecommunications and Information

Administration (NTIA) uses a complex chart with more than 500 colored patches to illustrate spectrum-

use allocations.

telling them exactly how they can speak, and pun-ishing others who attempt to speak (see the NTIASpectrum Allocation Chart, Figure 5).

There are several rationales for government regulation of spectrum. The airwaves are considered a public asset, not to be left to thevagaries of the private market. Regulation pro-motes a diversity of access by different voices, and maximizes efficiency in use of the airwaves.Government involvement prevents the chaos ofruinous interference that might occur in a vacuum.And there are important public safety and nationalsecurity uses of wireless communication that government promotes and manages.

Behind all these rationales stands a singleassumption: scarcity. If spectrum were not scarce,and simultaneous uses of the same spectrum werenot mutually exclusive, there would be no reasonto treat it differently from other forms of speech.Whether spectrum is in fact as scarce as we assumeis a major theme of this paper.

Spectrum regulation developed early in the 20thcentury in response to two developments: a bur-geoning commercial radio broadcast industry, andfears of chaos if government did not step in. Thefailure of nearby ships to heed the distress signal of the Titanic was seen as evidence for more extensive government intervention. Back at home,nascent radio broadcasters were squabbling aboutinterference, arguing over who had the right totransmit on particular channels. Secretary ofCommerce Herbert Hoover pushed for federaloversight of spectrum allocation. He was rebuffedby the courts, who found he lacked statutoryauthority. So the Radio Act of 1927 and theCommunications Act of 1934 were passed, establishing the Federal CommunicationsCommission as the prime arbiter of the airwaves.13

For sixty years, spectrum policy meant decidingwhich uses – and which users – were entitled tofrequencies.14 Federal spectrum allocation operatesas a kind of industrial policy, choosing some services for favored treatment and often protectingproviders from competition. The biggest change inrecent years has been the shift to auctions and flex-ible licenses as the preferred initial assignmentmechanism. Responding to the critique first articulated by Nobel Prize-winning economistRonald Coase in 1959,15 the FCC and many othergovernments now use auctions as the primaryassignment tool, rather than comparative hearings(“beauty contests”) or lotteries.

An important element of the FCC’s licensingregime is what the licensees receive. In virtually allcases, licensees do not receive the right to controlthe spectrum absolutely. The CommunicationsAct does not view spectrum as a tangible commodity. Licensees receive a right to use the frequency to provide a particular service, and only that service.16

A small amount of spectrum is assigned not toany specific users, but for “unlicensed” operation.The government sets technical requirements, suchas power limits, for users of the band, and providescertification mechanisms for devices that operatewithin it. Users of unlicensed devices have no formal protection against interference from other users in the band, but they need no specialpermission to operate there. The FCC also allowsvery low power devices (less than one watt) tooperate in significant sections of the spectrumunder its Part 15 rules, on the grounds that theyare too quiet to interfere with any other service.

12


Paradigm Shift: From Static to Dynamic

Wireless systems today are not just better and faster than those of the past.They can be fundamentally different.One could explain a car as simply aspeedier and more durable horse, or acomputer as nothing but a very fast calculator. Those descriptions soundlaughable. We know that new technolo-gies have countless benefits and impactsthat their predecessors don’t. Horses arehorses, and cars are cars, even though insome circumstances one can substitutefor the other.

So too with wireless communications.Systems built using modern techniquessuch as spread spectrum, software-defined radio, and mesh networking canserve the same purposes as systems builtusing older approaches. At this rela-tively early stage in the radio revolu-tion, the two types of systems look verysimilar. If the newer dynamic systemscan develop and be deployed, however,they will eventually seem as differentfrom their predecessors as a car doesfrom a horse.

The Traditional ApproachTraditional wireless systems are static.They assume dumb receivers and dumbtransmitters, whose function is to blastout a signal at the maximum allowablepower level. The model is quitestraightforward: Imagine throwing alarge rock into the middle of a roundpond. Ripples will radiate out from thepoint of impact, eventually reaching theshore. No skill is required for a personstanding on the shore, or sitting in aboat on the water, to come into contactwith the ripples.

The good aspect of this model is thatit doesn’t require much sophistication inthe endpoints. Transmitters andreceivers that don’t think much forthemselves are relatively cheap to build.When the costs of radio hardware andcomputing power are high, such savingsmake a difference in what’s economicallypossible. The downside of the staticapproach is that the devices aren’t smartenough to get out of each other’s way ifthere are multiple signals in the same

13

When the Devices Get SmartIntelligent transmitters and/or receivers can engagein a different form of wireless communication thanthe traditional static systems. Rather than merelywaiting for an incoming signal, the receivers can contribute to the communications process. Ratherthan radiating constantly toward static targets, thetransmitters can craft what they send for maximumefficiency. Call this dynamic wireless communication.

Changes in the nature of wireless devices alsoaffect the way devices interact with each other, andwith their surroundings. In other words, theychange the interference environment. As discussedearlier, interference is a consequence of systemdesign, rather than an inherent property of the radiospectrum. Interference is also inherently a legal con-struct. No radio signal on planet Earth is perfectlypure. There is always some external radiation thatimpinges on transmissions. Regulations or otherlegal mechanisms distinguish between permissibleincidental noise and impermissible interference.19

The static mechanism to overcome ambient“interference” is to raise power output. The louderthe signal, the easier it is to find among othernoise. Of course, raising power increases the likeli-hood of impinging on other signals, especiallythose adjacent either geographically or frequency-wise. Regulators must therefore define power out-put and license geography limits carefully.Historically, large amounts of spectrum were keptas guard bands where no one could transmit, toallow a wide “buffer” between licensed signals.

Static wireless systems have traditionally dealtwith interference from other transmissions throughlegal means. No one else may transmit in licensedbands. The FCC’s rules provide penalties for“harmful interference,” which is defined in termsof the effects of the second signal on the licensedservice.20 When the FCC proposed to authorize alarge number of lower-power FM radio stations foruse by community groups, licensed radio and TVbroadcasters expressed alarm that their transmis-sions would be threatened. These opponents con-vinced the FCC to significantly scale back thefreedom it granted to low-power FM broadcasters.

Dynamic wireless systems look at interference dif-ferently. It’s the equivalent of listening closely rather

space and frequency band. Throw two rocks intothe pond, and it will be impossible to determinewhich originated the resulting ripples.

For wireless communication, the solution to this“interference” problem was to create exclusivitywithin frequency bands. In the early days, such aswhen radio and TV broadcasting were established,affordable receivers weren’t even smart enough totell what was in their band. The FCC establishedwide “guard bands” around licensed frequencieswhere no one could transmit. That’s why the threeoriginal U.S. broadcast TV networks are typicallyon either channels 2, 4, and 7 or channels 3, 6,and 10.17 The “white space” in between is dark toensure receivers in each channel don’t becomeconfused.18

Static wireless systems are static because of thecost and computational capability of hardware,more so even than scarcity of spectrum. As is wellknown, computers have become more capable overtime. In a famous formulation, Intel co-founderGordon Moore noticed that transistor density onmicroprocessors doubled every 18 months thanksto advances in manufacturing technology. Hisobservation became a prediction, and then a law,which has held true for 35 years.

The implication of Moore’s Law is that whatevera computer can do today, it can do twice as well in18 months, or twice as cheaply. If you bought a PCfor $1000 three years ago, today you’d be able tobuy a machine four times as fast for that same$1000. On the other hand, if you bought a colorTV three years ago for $300, you might be able toget a slightly bigger screen or sharper picturetoday, but you’d see the same programming.

A radio itself is not a computer. It is a device fortransmitting and/or receiving signals. However,computers can be used to control radios, or toprocess those signals. Just as devices from automo-biles to air conditioning systems benefit from having computer “brains,” computers can improvewireless communications devices. With today’scomputing power, in fact, they can totally trans-form them.

14


than asking the person you’re conversingwith to talk more loudly. Because of theirflexibility, dynamic systems often have theability to “maneuver” around potentialinterference, whether by splitting up signals into packets spread across a widerange of frequencies, hopping from placeto place in the spectrum, or sending com-munications through a physically distinctroute across a mesh network.

Think about a group at a cocktail party.Many people can hold conversations withone another simultaneously. They can doso not because they each shout over theothers, or because they agree on a set ofrules to define who can talk when andhow. It’s obvious to us that the reason somany people can talk at once is that thespeakers modulate their volume and the listeners use their brains to distinguish theirpartner from the ambient noise (see Figure 6). Ifyou’ve ever tried listening to a piece of music andconcentrating on different instruments to pull themout of the mix, you’ll understand this process imme-diately. Now transfer the setting from smart humanlisteners to smart digital radios.

Another analogy is the Internet. The Internetdoesn’t have master directories or switches control-ling the flow of information. Every router candecide independently where to send traffic. Thisworks effectively thanks to cheap capacity andcheap computers that power the routers and otherdevices such as caches and Web servers at theedges of the network.

As these analogies show, the switch from static todynamic approaches has two major consequences.First, the capacity of the system to transmituseful information increases. The same spectrumcan hold more communications. The intelligence ofdevices is substituting for brute-force capacitybetween them. Imagine what highways would belike if cars couldn’t be steered quickly to avoid colli-sions and slowdowns. There would have to be hugebuffers between each vehicle to prevent accidents…precisely what exists in the spectrum today.

Dynamic wireless techniques effectively multiplythe usable spectrum. Robert Matheson of the

Radio Revolution

National Telecommunications and InformationAdministration (NTIA) at the U.S. Department ofCommerce has proposed an “electrospace” modelto describe the ways that today’s wireless systemscan coexist. He proposes the following variables:physical location (latitude, longitude, and height);frequency; time; and direction of arrival (azimuthand elevation).21 These seven degrees of freedomfor sharing spectrum compare with the singledimension – frequency – under the static approach.And the electrospace framework is still too limited.It does not include techniques such as low-powerunderlay, cooperative mesh networking, and soft-ware-defined radio, which are discussed later.

Second, the architecture of wireless systemscan change. Instead of cheap, dumb terminals atthe endpoints, there are agile, intelligent devices.Networks as a whole become more decentralizedand more flexible. Two-way communicationreplaces one-way blanket broadcasting as the dominant mode of connectivity. What were oncesingle-purpose, hardwired systems dominated byproprietary radio components increasingly becomegeneral-purpose, adaptive platforms dominated bycommodity computing components. These subtlechanges have dramatic consequences. Just as commodity PCs vanquished powerful mainframes

15

FIGURE 6 – COCKTAIL PARTY: A good metaphor for the receiver capabilities

of smart devices is how people communicate at a cocktail party. Multiple

conversations can occur simultaneously despite ambient noise.

by bringing computing to the masses, dynamicwireless systems will replace central transmissiontowers with millions of interactive end-user devices.

Survey of Dynamic Wireless TechniquesDynamic wireless communication is a deliberatelybroad classification that includes many technicalmechanisms, with new ones being developed allthe time. Once the transmitters and receivers areseen as computers that can contribute to the effi-cacy of the communications system, all sorts ofpossibilities emerge. These possibilities do notrequire any particular spectrum, or any particularspectrum policy regime. However, as will be dis-cussed later, spectrum policy significantly influ-ences the kinds of techniques that are used, byestablishing the economic conditions and incen-tives for spectrum users.

Spread SpectrumAs discussed earlier, wireless systems designershave long understood how to “multiplex” multiplesignals by sending them along different frequen-

cies, or by splitting up spectrum into tiny slices oftime. These multiplexing techniques are entirelyconsistent with the industry structure that devel-oped under the dominance of the static wirelessapproach. Frequency-division multiplexingrequires the most limited possible intelligence indevices, with spectrum bands exclusively allocatedto specific users. Time-division multiplexing ismore complex, but traditionally requires all thedevices in a system to synchronize their “clocks” inorder to know which signal is in which time slice.Such synchronization effectively requires exclusivecontrol of a frequency.

There are newer multiplexing techniques withdifferent results. The first developed was “spreadspectrum.” Hollywood actress Hedy Lamarr andmusician George Antheil filed for a patent on aspread spectrum communications system in 1942,though real-world deployment occurred later. Aspread spectrum system inverts the static model oftransmitting with high power on a narrow channel.Using low power and spreading the signal across arange of frequencies, it’s possible to carry more

16


An Intelligent Device Bill of Rights

A key question in a world of dynamic wireless sys-tems is how to set the proper boundaries on howdevices can operate. More sophisticated equipmentcan reliably transmit signals in situations that wouldotherwise be subject to interference. “Cognitive”devices can sense the local spectral environment,adjust their transmissions to take advantage of tem-porarily empty spaces, and move their signals else-where as soon as another transmission is detected.Such devices are incompatible with traditional fre-quency-based licensing, which assumes a frequency isexclusively dedicated to a licensee.

One idea under consideration by the FCC’sTechnological Advisory Committee (TAC) is an“Intelligent Wireless Device Bill of Rights.”22 It wouldestablish a set of principles to allow effective sharing

of the spectrum. A draft of the Bill of Rights pro-posed in September of 2002 includes three articles:z Any intelligent wireless device may, on a non-

interference basis, use any frequency, frequencies,or bandwidth, at any time, to perform its function.

z All users of the spectrum shall have the right tooperate without harmful electromagnetic interfer-ence from other users.

z All licensing, auctioning, selling, or otherwise dis-position of the rights to frequencies and spectrumusage shall be subordinate to, and controlled by,Articles 1 and 2, above.

The Bill of Rights is at the early stages of discussionwithin the TAC, which itself has no formal authority.It suggests, though, how significantly the wirelessparadigm shift now underway could change thebasic framework of spectrum policy.

transmissions simultaneously. The basic notion isthat if the transmission is broken into pieces, eachof which is tagged with a code, a receiver thatknows the code can reconstruct the message. Thewider the spreading, the more space there is inbetween the coded packets to send other signals atthe same time (see Figure 7).

Taking spread spectrum to its logical conclusion,if the signal is spread wide enough, the powerdensity can be so low that the signal becomeseffectively invisible to other systems in the samebands. Radio frequencies are never totally emptyof noise. Radiation-emitting devices such as hairdryers and microwave ovens, as well as cosmicbackground radiation, create a “noise floor” thatall systems must contend with. Static systems doso by using high-enough power that it’s easy todistinguish the high-power signal from the low-level noise.

With enough smarts, though, a dynamic systemcan transmit and receive very faint signals withoutever raising above the noise-floor threshold. Onlywhen very large numbers of such devices operate inthe same location is interference even a realisticpossibility. This approach is known as ultra-wide-band (UWB). Many UWB systems employ veryshort “carrierless” pulses of electricity that givethem other unique and beneficial properties.

Radio Revolution

Because of fears about interference,commercial use of ultra-wideband forcommunication was illegal until early2002, when the FCC authorized itfor the first time.23

Space-Time CodingMany other multiplexing schemes are possible beyond spread spectrumand UWB. For example, companiessuch as Northpoint Technology havedeveloped systems that multiplexsatellite and terrestrial transmissionsin the same frequency band.24

Satellite signals arrive from above,while terrestrial signals are sent hori-zontally. A smart enough system candistinguish these two signals basedon their angle of arrival, and can

even do so without requiring modifications to theexisting satellite system.

Northpoint’s technology is an example of abroader class of techniques that take into accountthe physical location of transmitters and receivers.Static broadcasting uses a saturation approach.The receivers can be anywhere within the propa-gation footprint of the signal. The transmitter has no idea where they are beyond that, and thereceivers know nothing other than that the transmitter is in the same footprint. As theNorthpoint system shows, however, the locationof transmitters and receivers is a useful piece ofinformation. A signal arriving from thousands ofmiles overhead is different from a signal arrivinglaterally from a few yards away, even if both arewithin the same frequency band.

Space-time coding techniques use the physicaltopology of the network, or the surrounding envi-ronment, to add efficiency to wireless communica-tions systems. For example, the BLAST systemdeveloped at AT&T Bell Labs employs antennadiversity and “multiple input/multiple output”(MIMO) technology to increase capacity. Insteadof a single antenna at the receiver, BLAST employsarrays of multiple antennas at both the transmitterand receiver. Comparing the signal received at thedifferent antennas makes it easier to distinguish

17

A A A A

B B B B

C C C C

ABC ABC

FIGURE 7 – SPREAD SPECTRUM: In spread-spectrum communications, low-power

radio transmitters divide their signals into coded packets across a range of frequen-

cies and receivers reconstruct the message.

transmissions from noise, increasing effective capacity. Start-ups such as AirgoWireless are now using MIMO techniquesto enhance capacity and range of wirelessLAN chipsets.

Even factors that seemingly reducecapacity can be employed to increase it.The bane of many wireless systems is“multipath.” When radio waves encounterobstacles such as walls, some fractionbounce off the obstacle and the remainderpass through. The ones that bounce maystill reach the receiver. But they do sothrough a more circuitous, and thereforeslightly slower, path. The receiver seesthe same signal twice (or more), a splitsecond apart. This multipath effect canconfuse the receiver, degrading the signalquality.

With a properly defined system, how-ever, multipath becomes just another information-adding physical element. Knowing how a signal is bouncing around tells the receiver somethingabout the location and nature of the transmitter. Ifthe two temporally spaced signals are identified asthe same transmission, they can be combined in a buffer to enhance the output signal. Similar techniques can be used for other factors that traditionally cause “interference,” such as mobility.

Mesh NetworkingMesh networking is somewhat different than theprevious techniques. It is a family of cooperativenetwork architectures that can be applied throughsoftware to any radio technology. The basic defini-tion of a mesh is that receivers talk to each other aswell as to the transmitter.25 A good example of amesh network is the Internet. Every router has atable that allows it to send packets to many otherrouters, rather than through a central clearing-house. Thanks to this architecture, the Internetavoids congestion choke points and single points of failure. When one link is down or overloaded,traffic automatically shifts to other links.

Wireless systems have traditionally used either apure broadcast architecture (one central transmit-ter), or a hub-and-spoke approach as with cellular

telephone networks, where users connect throughlocal towers (see Figure 3). Many unlicensed technologies, including Bluetooth and WiFi, offersome mesh networking capabilities. Vendors suchas LocustWorld sell WiFi access points withsophisticated meshing software to discover and connect to other nodes automatically.

The benefit of a mesh approach is that there arelikely to be other end-users of the network closerto you than a tower or central broadcast facility.Shorter distances mean better signals, lower powerrequirements, and the ability to avoid obstaclessuch as trees.

Consider the task of providing “last mile” high-speed Internet connectivity to a neighborhood (seeFigure 8). The benefit of mesh networking is thatevery new house brought online adds something tothe network, improving performance and reliabil-ity. The difficulty is that a mesh network doesn’twork with one or two nodes. The system requires acritical mass of devices to operate effectively. Thatnumber depends on the service and deploymentenvironment. Several companies have tried to selllast-mile mesh networking gear, includingSkyPilot, Omnilux, and RoamAD. None has yetachieved a large-scale commercial deployment,though field trials are ongoing.

18


FIGURE 8 – MESH NETWORKING: In a mesh network, every device added to

the system augments the network. In the case of a "last-mile," neighborhood

network, end-users connect to the Internet by sharing connections with their

neighbors allowing for shorter distance, lower power transmissions.

Other providers are selling mesh networkinggear for public safety applications. One vendor,Tropos Networks, recently built a 17-cell meshnetwork for the police force of San Mateo,California so that officers could access crime data-bases from laptops in squad cars.26 MeshNetworksis in trials with the Orange County Fire andRescue service in Florida, offering vehicle and personnel tracking as well as mobile Internet connectivity through a distributed mesh network.

Software-Defined RadioAt the heart of any wireless communications systemis the radio. A radio transmits or receives wirelesssignals encoded into waves that oscillate at frequen-cies somewhere within the radio spectrum.Traditionally, those radios have been fixed in hard-ware. A radio talks to a fixed swath of spectrum, andunderstands a fixed modulation scheme for codingsignals. It’s like a dedicated telephone line betweentwo businesses. You simply couldn’t use it to call yourgrandmother, or another business across the street.

Software-defined radio (SDR) uses software tocontrol how the radio works.27 It’s like replacingthe dedicated phone line with a connection thatgoes through an electronic switch. Suddenly manyof the characteristics that were immutable becomeflexible. Capabilities not envisioned when thedevice was built can be added later.

SDR has several benefits. One device can sup-port multiple services transmitting on different frequencies with different encoding schemes. Amobile phone handset, for example, could receivesignals from more than one service provider, orfrom service providers in different countries,regardless of what technical standard they employ.This is particularly important for markets, such aspublic safety, where incompatibilities between systems, such as those used by police and firedepartments responding to the same emergency,are critical problems. The U.S. military has fundedsignificant research and development related toSDR for similar reasons. The Joint Tactical RadioSystem (JTRS) is now under development throughprime contractor Boeing and subcontractorsincluding Vanu, a start-up in Cambridge, MA thatis a leading SDR technology developer.

Radio Revolution

The flexibility of SDR reduces costs by eliminat-ing duplicate equipment, both for users and forservice providers. Imagine a cellular network, forexample, where each service provider only had toput up towers where others had not, rather thaneach of them having to build a redundant nationalfootprint.

The potential of SDR goes significantly beyondcost reduction. Because a software radio is soft-ware, it can run on general-purpose computerssuch as Windows and Linux devices, using mass-produced digital signal processors (DSPs) andother hardware. Such devices benefit from Moore’sLaw and the competitive dynamics that steadilypush costs down and capabilities up. As DSP chipsbecome more powerful for the same price, an SDRsystem can decode a larger swath of spectrum orperform other new functions. SDR systems canalso take the decoded radio signals and feed theminto other applications.

Agile or cognitive radios are a sub-category ofSDR currently in the development stage. Agileradios can “jump” from one frequency to anotherin a matter of milliseconds. Combined with pro-cessing capabilities that allow such devices tosample the spectrum around them, agile radioscan in effect manufacture new spectrum. Even achannel supposedly occupied by a licensed systemis empty much of the time in much of the definedphysical area. An agile radio could hop amonglocal, short-duration empty spaces in the spec-trum, moving whenever it sensed another trans-mission in the same band. Such devices couldeffectively become their own virtual networks,creating connections with other nodes whereverthey are.

Implications of Dynamic ApproachesThe switch from static to dynamic wireless communication has huge consequences. Technicalapproaches created to solve technical problemsturn out to have major policy, business, and evensocial consequences. These consequences, such asthe possibility of replacing spectrum licensingwith “commons,” have generated a great deal ofattention. They would not be possible without thetechnical advances we have described so far. In

19

many cases, the technical advances have grown upwithin the dominant wireless paradigm, as withQualcomm’s CDMA spread spectrum technologyfor licensed cellular networks. As wireless tech-nology moves forward, however, the possibilitiesfor radical change will become more difficult to ignore.

Author George Gilder identified the disruptivepotential of dynamic wireless technologies tenyears ago, in a prescient article in Forbes ASAP.28

At a time when policy-makers were infatuated with spectrum auctions, Gilder pointed out thatauctions were counterproductive if intelligentdevices could share spectrum and avoid interfer-ence on their own. The technological paradigmshift of dynamic wireless calls for paradigm shiftsin regulation and business as well.

CommonsThe first and biggest consequence of the radiorevolution is that licensing of spectrum frequen-cies is no longer required. Recall that the originalbasis for spectrum licensing by the governmentwas the fear of ruinous and pervasive interference.If devices can operate with sufficiently low powerand high intelligence to avoid one another, exclu-sive rights are no longer necessary to preventinterference. With a properly defined environ-ment, users effectively can’t prevent each otherfrom communicating. That opens the door for anew regime that allows anyone to transmit withingeneral technical guidelines.

This notion has become known as a “spectrumcommons.” The analogy here is to common landsin the Middle Ages in England, where anyonecould graze their sheep. Not all such environmentsare subject to the infamous “tragedy of the commons.” If there is enough open space, andgood enough legal or customary rules governingindividual action, a commons can thrive withoutrapid exhaustion. Until recently, these conceptswere rarely applied to spectrum, because capacityconstraints were thought to be so severe. As notedearlier, though, the revolution in wireless technol-ogy lifts the constraints that made spectrum scarce.

Beginning with the early 1990s challenge toFCC spectrum auctions from Gilder, network

engineer Paul Baran, and communications scholarEli Noam, and continuing in the late 1990s andbeyond with the work of Yochai Benkler, DavidReed, Lawrence Lessig, Tim Shepard, and myself,the argument for spectrum commons has graduallytaken shape.29 As will be discussed later, there areseveral varieties of spectrum commons, includingunlicensed bands and underlays. Evolving dynamicspectrum management techniques are rapidly cre-ating new possibilities for sharing rather thanlicensing spectrum.

If there is enough capacity to support a com-mons, exclusive rights are unnecessary. By analogy,the ocean is not infinite in size, but it is largeenough that ships can be trusted to navigatearound one another (see Figure 9). The ships, likedynamic wireless devices, can intelligently altertheir routes to avoid collisions. There is no need togive companies exclusive shipping lanes, and pro-hibit other ships from using those routes unlessthey pay a toll. Such exclusivity would significantly

20


FIGURE 9 – UNLICENSED VS. LICENSED BUSINESS

MODELS: Technology is transforming the capacity of the

spectrum from scarcity to abundance, akin to the vastness of

the oceans. But current airwave regulation is similar to giving

companies exclusive shipping lanes, and requiring non-

licensed users to pay a toll to access them.

reduce the level of shipping traffic, with no corre-sponding benefits. Technology is making the wire-less world look more and more like the ocean.

Well-functioning commons produce severalnormative benefits. Because access is no longercontrolled by a designated gatekeeper, everyonecan participate. Virtually anyone could distributeTV programs, promote an idea, or engage in agroup conversation using the same mechanismthat today supports a limited set of broadcastersand operators. This is freedom of speech on aspectacular scale.

Market StructureThe second implication of dynamic wireless sys-tems is that the business structure of marketschanges. Static systems necessitate exclusivity. Thedownside of exclusivity is that no one else can con-tribute to a network. The licensee must bear thetotal cost of building network infrastructure. Thelicensee typically recovers that cost by chargingfees to users for both devices and communicationsservices. Any vendor of user or network equipmentmust sell to or through the licensed operators.They are the only customers who can take technologies and legally put them into the market.

There is nothing inherently wrong with this“infrastructure” market model. It is traditionallythe market structure used for services that havehigh capital costs and benefit from economies ofscale. Having every person responsible for buildingthe roads passing their home wouldn’t make muchsense, nor would having every person responsiblefor bringing their own connection to a central tele-phone exchange. There are, however, seriousdownsides. Deployment is slow because it is costlyand requires proven models for recouping thatcost. Innovation is constrained because only a fewlicensees control access to the market. Uniformityand interoperability are enforced by the licensee,but services and equipment are costly because theyare provided in limited volumes and based on proprietary standards.

As wireless devices become more intelligent andcommons arrangements become viable, new mar-ket structures become possible. If the spectrum isno longer part of the service equation, the primary

Radio Revolution

element of the “service” offered to end-usersbecomes the devices those users purchase. Becausethose devices run on standards defined by industrybodies rather than mandated by spectrum licensees,there can be open, competitive markets to buildbetter and more cost-effective equipment.30 Wemove, therefore, from a market for centralizedinfrastructure and proprietary services, to a marketfor consumer devices, software, and ancillary serv-ices. Users pay a significant fraction of the totalnetwork build-out costs directly, by purchasinghardware, greatly reducing the expenses serviceproviders must undertake. For many services, thereare still core network costs—for access points,backhaul to wired Internetbackbones, authentication,roaming, and security—but these are limited compared to the all-encompassing networkbuild-out that licensedoperators must undertake.

Furthermore, dynamicwireless devices allow formarkets with greaterdiversity at several points.Many equipment vendors, several providers, andapplication or content providers can compete,because there is no mandatory control point andeach provider can leverage the infrastructure builtby others.

Incentives for RobustnessIntelligent or dynamic spectrum management tech-niques may be used in any regulatory environment.However, the nature of spectrum regulation heav-ily influences incentives for deployment of intelli-gent devices. The traditional, and still dominant,environment is exclusive licenses for frequencybands. Spectrum licensees have incentives tosqueeze as much capacity as possible out of theirspectrum. On the other hand, they have incentivesto make the devices users must purchase as inexpensive as possible. In a static system, themoney is all in the service; the devices are dumb.There is no need to make them robust againstinterference, because interference from other

21

“Change the technology,

and the economics and

the law of spectrum use

must change, too.”

– ELI NOAM, PROFESSOR OF ECONOMICS

AND FINANCE, COLUMBIA UNIVERSITY

systems is prohibited and policed by the FCC.Similarly, there is no great incentive to make thedevices flexible, because the service provider isfocused only on supporting its own service. Indeed,to keep the cost of switching to another serviceprovider high, licensed providers have an incentiveto discourage software-defined radios, which couldswitch frequencies (and hence service providers) atthe click of a mouse.

All that changes in a commons environment.Because wireless devices in a commons have no

legally guaranteed protec-tion against interference,they must guard against itusing technical means.Fortunately, that is whatdynamic wireless devicesare good at.

Static systems createincentives to make thereceivers as dumb as pos-sible. Dumber meanscheaper, after all. Becausethe central transmitterdoes the heavy lifting,there are no significantbenefits from intelligenceat the edge devices. Whenend-user devices becomedynamic, however, theycontribute to the integrity

and performance of the overall system. Makingthem smarter and more robust improves perform-ance. And without license restrictions keepingother devices from transmitting on the same frequencies, robustness based on intelligence isthe only path open.

“We need to think of

ways to bring [WiFi]

applications to the

developing world so as to

make use of unlicensed

radio spectrum to

deliver cheap and fast

Internet access.”

— UN SECRETARY GENERAL

KOFI ANNAN

WiFi as a Case StudyWiFi DefinedWiFi is the most prominent unlicensed wirelesstechnology available today. It is a family of spreadspectrum wireless local area networking standardsdesigned to allow users to send and receive data at11-to-54 Mbps within a few hundred feet ofanother WiFi device or access point. WiFi is agreat case study for the impact of dynamic wirelesstechnologies.

Wireless data networking services have beencommercially available since the 1980s, beginningwith Ardis, a joint venture of IBM and Motorola,and RAM Mobile Data. Ardis was purchased byAmerican Mobile Satellite and renamed Motient;after reorganizing through bankruptcy in early2002 it is still trying to right itself financially.RAM Mobile data was renamed Mobitex and isnow part of Cingular Wireless. It is the primarynetwork used by wireless email devices such as theRIM Blackberry and the Palm VII. Motient andMobitex are wide-area systems that target enter-prise and messaging markets. Another wide-areawireless network, the Metricom Ricochet system,offered services directly to end-users, providingwireless Internet access in several U.S. cities forlaptop users. Metricom filed for bankruptcy in2001; its assets were purchased by Aerie Networks,which is attempting to re-launch the service.

These early wireless data networks generallyused licensed spectrum, though Ricochet employed900 MHz unlicensed frequencies in some areas.They offered low-speed connections (19.2 kbps orless) with wide-area coverage in cities or nation-wide. Today, third-generation cellular networks arebeginning to offer packet data networking servicesas well, typically at higher speeds. None of theseofferings has yet become a mass-market success.

WiFi has been exactly the opposite story. TheInstitute for Electrical and Electronic Engineers(IEEE) ratified the 802.11b standard for wirelesslocal area networking (WLAN) in 1999. Vendorssuch as RadioLAN and Proxim had been offeringproprietary WLAN systems, for both office envi-ronments and home networking. 802.11b, relatedto the 802.3 Ethernet standard, was envisioned pri-marily as a wireless replacement for wired Ethernet

22


connections in corporate environments. In 1999,though, Apple Computer introduced a consumer802.11b device, the Airport, using chipsets fromLucent. The market exploded.

The WiFi market grew to $1 billion annually(primarily in hardware sales) by 2002. Most of that period has been a time of contraction in the technology and telecom sector making theachievement even more impressive. More thanhalf of U.S. companies now support WiFi net-works, and another 22 percent plan to do sowithin a year.31 And WiFi sales are projected tokeep growing. Cahners Instat sees the marketreaching $4.6 billion by 2005, and other researchfirms have issued similar projections.32 By 2008,says Allied Business Intelligence, 64 million WiFinodes will be shipped annually.

Secrets of WiFi’s SuccessWhat made WiFi such a success, especially com-pared to previous wireless data systems? After all,WiFi provides only short-range connections; on itsown, one access point can’t provide ubiquitous cov-erage in a neighborhood or city.

WiFi has thrived because it has benefited froman ecosystem that could only exist with the type oftechnology it uses. Because WiFi is a low-power,spread-spectrum technology, WiFi devices cancoexist without the requirement of spectrumlicensing to prevent interference. That means thereis no need for service providers, cell towers, con-trolled hardware markets, or expensive spectrumlicenses. Anyone can buy a WiFi device and estab-lish a network.

Because WiFi is an open standard and an equip-ment-centered rather than service-centered market(again, both of which flow from the nature of thetechnology), costs are subject to computer industrydownward pressure. A WiFi access point that costhundreds of dollars when introduced is availablefor less than $100 today. Chipsets are down in the$10 range, allowing laptop, personal digital assis-tant (PDA), and mobile phone vendors to incorpo-rate them with little or no price increase for theoverall system. According to Intel CTO PatrickGelsinger, a WiFi network costs half as much peruser per month to operate than a DSL connection,

Radio Revolution

and as little as one-tenth as much as a third-gener-ation cellular network.33

The WiFi market wouldn’t have taken off with-out standards – both the technical ones defined bythe IEEE and the interoperability testing and certi-fication done by the Wi-Fi Alliance (formerlyWECA), an industry trade group. There is a needfor such industry standards because there is anentire industry of vendors, rather than one serviceprovider and its chosen suppliers, operating in theWiFi universe. Now that the standards are inplace, the market can take advantage of contribu-tions from many competing vendors.

WiFi is now experiencing the next phase ofdevelopment that can occur with dynamic wirelesssystems. It is evolving anddiversifying. The IEEEhas already extended theoriginal 802.11b with sev-eral variants which will bediscussed later. Meanwhile,start-ups are offering newkinds of devices that addfunctionality to the origi-nal short-range WiFiaccess points. Vivato, forexample, has developed asmart, phased-area antennatechnology that can extend the range and capacityof WiFi signals, while remaining completely back-ward compatible with existing equipment.Locustworld in the UK is shipping 802.11 meshnetworking boxes that automatically create ad hocmesh networks with each other. Vendors such asEngim and BroadBeam are offering WiFi switchesthat increase capacity of WiFi infrastructure.Sputnik is shipping low-cost yet powerful self-con-figuring access points. Companies such as Telesymand Vocera are running voice communications overWiFi, opening up whole new market opportunities.

In a traditional static wireless system, changingthe network technology means upgrading thewhole network. Not just end-user devices but allthe core transmission elements must be upgraded.The costs and time frames involved parallel thoseof deploying the system the first time. The transi-tion from analog to digital television (DTV) is a

23

“[T]he unlicensed bands

employ a commons

model and have enjoyed

tremendous success as

hotbeds of innovation.”

— FCC CHAIRMAN MICHAEL POWELL

perfect example. DTV technology has been com-mercially available for more than a decade. In theU.S., the formal transition to DTV began in 1996,and still only a tiny handful of customers canreceive DTV broadcasts. In Japan, high-definitionTV (HDTV) service has been commercially avail-able since the 1980s, but Japan chose an analogstandard. Improving digital technology is nowconsidered the best way to deliver HDTV, and

Japan had to effectively start the entire processover again.

Compare that to the transition from orphanedWLAN standards. Two standards that competedwith 802.11, Europe’s Hiperlan for high-speedWLANs and the HomeRF standard for home net-working, have lost out in the marketplace. Thoughsome equipment has been orphaned, most vendorshave quickly switched to offering 802.11 products.

24


The Unlicensed World

Unlicensed wireless is far more thanWiFi. Dynamic techniques for efficientsharing of the spectrum, combined withthe open field for unlicensed innovation,are creating an explosion of new sys-tems, techniques, and business models.

The Spectrum of Spectrum-Use Regimes“Licensed” and “unlicensed” are gener-ally presented as the two models forspectrum usage. There are actually several variations, not entirely mutuallyexclusive. Put another way, the fact thatWiFi uses spectrum bands dedicated tounlicensed usage doesn’t mean that isthe only mechanism for regulators tocreate more space for unlicensed devicesand systems. Matheson’s electrospacemodel, for example, provides sevendimensions for sharing spectrum, withfrequency only one of the variables.34

Looking at the possibility of a spec-trum commons purely as a matter ofwhat regime to mandate for specific fre-quencies misses the point. Dynamic

wireless techniques, and the systemsthey make possible, ultimately erode thevery rationale for thinking of “the spec-trum” as a physical asset that can andshould be divided into frequencybands.35

For present policy purposes, however,it can still be useful to list prominentmechanisms for sharing spectrum. TheNew America Foundation and otherpublic interest organizations have urgedthe FCC to consider four basic spec-trum use regimes: exclusive licensed,pure unlicensed, shared unlicensed, andopportunistic unlicensed.36

Exclusive LicensedMost spectrum is exclusively licensedtoday. Licensees may be carriers, broad-casters, specialized mobile radio, corpo-rations, the military, or public safetyagencies. The licensee may haveobtained its license for free, or througha competitive mechanism such as anauction. Regardless, it has exclusive con-trol of the frequency band for a period

25

like any other spectrum block. Instead of a serviceprovider gaining the rights to control use of thespectrum, subject to limits set in the terms of thelicense, manufacturers gain the rights to selldevices that conform to FCC-designated standards.The devices themselves must be licensed, generallyon a generic and often self-licensing basis known as“type acceptance.”

An important point here is that unlicensed doesnot mean unregulated. In the 2.4 GHz band, forexample, the FCC mandates power limits and othertechnical requirements. New kinds of equipmentthat use different techniques than those alreadylicensed must receive direct FCC approval. Forexample, Vivato, a start-up that sells a novel “WiFiswitch” based on phased-array antennas, receivedFCC approval in late 2002 for its technology.

Shared UnlicensedUnder certain circumstances, unlicensed devicescan share frequencies with licensed devices. Thisarrangement is sometimes referred to as an ease-ment, by analogy to real property law. It is alsoknown as “underlay,” because unlicensed devicesoperate below the noise threshold of the high-power licensed devices in the band (see Figure 10).

There are two major examples of shared unlicensed use in current FCC rules. The first,Part 15, actually dates back to 1938. This sectionof the FCC’s rules allows devices below a strictpower limit to operate in significant portions of thespectrum. The severe power limits allow for onlyvery short-range devices, which do not producesignificant interference with licensed systems.

A more recent example of shared unlicensed useis ultra-wideband (UWB). The FCC first author-ized this technology in February 2002. It usestransmissions spread across huge frequency rangesat extremely low power, below the detectable noisefloor for other licensed devices in the same band.Using spread spectrum techniques, UWB systemsare able to reconstruct messages and support high-speed transmissions even under such restrictiveconditions. By authorizing ultra-wideband formuch of the spectrum above 3 GHz, the FCCeffectively created a huge new unlicensed ease-ment shared with licensed services.

of years subject to the limitations of its license.Other systems are prohibited from producingharmful interference with the licensee. The broad-cast television bands are good examples of exclu-sively licensed spectrum.

Some spectrum is licensed, but not for a singleuser. Examples include the bands for the amateurradio service, radio astronomy, and privatemicrowave systems. These bands resemble pureunlicensed spectrum, in that one entity doesn’t havetotal control. However, there are still restrictions onwho can transmit, and what kinds of services theycan provide. A technology that could increasecapacity or deliver a new kind of service can’t beused in these bands if it doesn’t meet those criteria.

Dedicated UnlicensedThe opposite regime is to have no exclusivelicensees in the band. For example, the 2.4 GHzIndustrial, Scientific, and Medical (ISM) band isopen solely for unlicensed use. No devices thatoperate in that band—which range from the WiFiaccess points to microwave ovens and cordlessphones—can claim protection against interferencefrom other approved devices in the band.

“Unlicensed” is actually something of a mis-nomer. It implies that the government has notmade the spectrum available to licensees, when infact the spectrum has been allocated and assigned

26


FREQUENCY

POW

ER

HIG

H-P

OW

ER L

ICEN

SED

SIG

NA

L

LOW-POWER UNLICENSED UNDERLAY

FIGURE 10 – UNDERLAY SHARING: In underlay sharing,

low-power, unlicensed devices share frequencies and avoid

interference by operating beneath the noise threshold of

high-power devices in the band.

Opportunistic UnlicensedOpportunistic sharing means taking advantage ofunused spectrum in licensed bands. As noted previ-ously, the official frequency chart paints a mislead-ing picture of how spectrum is actually used. Somefrequencies are allocated but not assigned to anyuser. For example, they may be set aside as “guardbands” between licensed frequencies for older serv-ices such as broadcast television. The guard bandswere necessary because the licensed equipmentwasn’t sophisticated enough to distinguish signalsotherwise. However, today’s smart unlicenseddevices could transmit in the guard bands withoutimpinging on the licensed services.37

In other cases, spectrum may be assigned butnot used in a particular area or for a particularperiod of time. For example, a cellular phonetransmission tower is only active when communi-cating with a handset nearby. When no user is inrange, the spectrum is temporarily available.Other frequencies, licensed nationally, may beused in New York City but not at all in Montana.“Cognitive radios” could detect such holes in thespectrum, switch communications there, and thenmove away as soon as the licensee began transmit-ting (see Figure 11). Furthermore, as the electro-space model shows, there are many ways to slicethe spectrum pie. An “angle of arrival” system, forexample, can opportunistically use “terrestrial”spectrum in bands licensed for communicationwith orbiting satellites overhead.

There is no reason to believe that all the possi-ble mechanisms for opportunistically sharing spectrum have been discovered or implemented.As wireless systems become more dynamic andmore intelligent, they will be capable of coexistingin new ways.

Current Unlicensed ProductsThough the WiFi name is getting a tremendousamount of attention and, as a result, has beenexpanded to include standards other than the orig-inal 802.11b, it is important to keep in mind thatWiFi is not a synonym for unlicensed wireless.WiFi is a local-area networking protocol. It deliv-ers data, such as Internet connectivity and email,across links of no more than a few hundred feet.

Radio Revolution

Local Area Networks (802.11)802.11 refers to the Institute for Electrical andElectronic Engineers (IEEE) working group forwireless Ethernet networking. The IEEE definestechnical standards, but does not certify compli-ance with those standards. In parallel, industryassociations such as the Wi-Fi Alliance createbrand names which vendors are permitted to use ifthey meet compatibility requirements. The termWiFi, a play on HiFi stereo systems, is such abrand name.38 Originally referring only to 802.11b,WiFi now encompasses 802.11 a, b, and g.

There are three widely deployed 802.11 technologies:

z 802.11b – The original WiFi, providing 11 Mbpsconnections using direct-sequenced spread spec-trum modulation in the 2.4 GHz frequencyband.

z 802.11a – A higher-speed standard delivering 54Mbps connections, but using different spectrum(5 GHz) and modulation (orthogonal frequencydivision multiplexing, OFDM) than 802.11b. Asa result, 802.11a systems are not backward com-patible with 802.11b, and require separateradios.

z 802.11g – A backward-compatible high-speedstandard, delivering 54 Mbps through OFDMlike 802.11a, but using the 2.4 GHz spectrum.

27

TIM

E

FREQUENCY

FIGURE 11 – OPPORTUNISTIC SHARING: In opportunistic

sharing, unlicensed devices detect and access licensed

spectrum that is not currently in use, and then move away

as soon as licensees begin transmitting.

The remaining alphabet soup of IEEE 802.11standards are mostly variants of these protocols.For example, 802.11e, based largely on technologydeveloped by Sharewave (now part of chip vendorCirrus Logic), adds quality of service mechanismsto better support video and voice traffic. 802.11iadds a new security protocol to the relatively ineffective WEP encryption in 802.11b. 802.11j isa WLAN protocol for the 4.9 GHz to 5 GHz unlicensed spectrum in Japan.

Metropolitan-Area Networks and Last Mile (802.16)Metropolitan-area networks (MANs) operate overlonger distances than LANs, typically a mile ormore. They are designed to provide relatively highbandwidth to a moderate number of fixed sites,such as homes and businesses, compared to LANs,which connect individual devices. One MAN appli-cation is the so-called “broadband last mile,” whichcan substitute for wired solutions such as digitalsubscriber lines and cable modems.39 However,

28


COMMENTSREPRESENTATIVE COMPANIES

SPECTRUMCAPACITYRANGETECHNOLOGY

Primary wireless LANmarket today

Chipsets: Intersil, Agere,Cisco, IntelEquipment: Cisco,Proxim, Netgear, Vivato,AppleServices: Boingo, Cometa

2.4 GHz 11 Mbps 300 feet 802.11b (WiFi)

Useful for corporate networks, backhaul, andmedia applications

Major 802.11b vendorsplus Atheros, Bermai

5 GHz 54 Mbps 150+ feet 802.11a (WiFi)

Backward-compatiblewith 802.11b devices

Major 802.11b vendorsplus Broadcom

2.4 GHz 54 Mbps 300 feet 802.11g (WiFi)

Originally designed forcable replacement; mar-ket niche unclear

Ericsson, Nokia, Intel,Toshiba, Microsoft,3Com, Motorola

2.4 GHz 1 Mbps 300 feet 802.15.1 (Bluetooth)

High-bitrate personalarea networking formedia devices

XtremeSpectrum,Motorola, TI, TimeDomain, Philips

Wideband(3.1–10 GHz)

110 and 200Mbps

30 feet at 110Mbps or 12 feetat 200 Mbps

802.15.3a (WiMedia)

Low-bitrate personal areanetworking for sensors

Philips, Honeywell, Mitsubishi, Motorola.

900 MHz, 2.4GHz, or wide-band

250 kbps 200 feet 802.15.4 (Zigbee)

Broadband metropolitan-area network connections

Motorola, AlvarionProxim, Fujitsu, Aperto

10-66 GHz for802.16; 2-10GHz for802.16a

70 Mbps 30 miles 802.16 (WiMax)

Ethernet, currently envisioned for licensedspectrum, but may evolve

Cisco, Flarion, HP, Nextel Mobile wireless

3.5 GHz 1 Mbps 15 km 802.20(MobileFi)

TABLE 1 — MAJOR UNLICENSED WIRELESS STANDARDS

MAN systems are also used to provide “backhaul”connections from such last-mile networks to cen-tral aggregation points, point-to-point connectionsbetween facilities, or coverage throughout a campus or other geographically defined facility.Increasingly, WISPs and non-profit communityaccess networks are creating MANs using line-of-sight relays on unlicensed spectrum (at 5GHz) andWiFi (at 2.4 GHz, for the last few hundred feet) tooffer affordable last-mile connections in rural andlow-income areas.40

The IEEE has established a standards group,802.16, for wireless MANs. The original 802.16specification was designed for very high frequen-cies, over 10 GHz. A more recent subgroup,802.16a, is crafting MAN standards for 2-10 GHzfrequencies that, unlike the original standards,don’t require line-of-site visibility. 802.16 envisionssystems delivering 70 Mbps of data over a 30 milerange. An industry alliance called WiMax, includ-ing Intel, Proxim, Fujitsu, Alvarion, Aperto, andNokia, has been formed to promote and ensureinteroperability of wireless MANs.

Though designed as LAN technologies, 802.11aand b are being used by some companies such asEtherlinx as the foundation for MAN systems.Generally these systems use the commodity 802.11physical layer and devices, adding their own mediaaccess control (MAC) layer to boost range.Another option, which several service providers arereportedly considering, is to use standard 802.11devices with enhanced access points from compa-nies such as Vivato that boost the effective range.

Several companies including Motorola (with itsCanopy system), Magis Networks, Proxim, IPWireless, Navini, BeamReach, Aperto, Soma Net-works, and Alvarion offer proprietary products thatare similar to 802.11b and 802.11a. Typically thesesystems offer better performance, reliability, or features such as security that aren’t well imple-mented in the 802.11 standards. Today, most ofthem operate in the 5 GHz band and focus onmarkets such as last-mile residential and small-business connectivity, especially in rural or other-wise underserved areas. Many of these companiesare part of the 802.16 effort. It can be expectedthat, as with WiFi, many proprietary systems will

Radio Revolution

eventually become standards-compliant. A newindustry association, WiMax, hopes to do for unli-censed wireless MANs what WiFi did for LANs.

Other companies are taking a different route todeliver wireless MAN connectivity. Instead oflong-range MANs sending data directly to cus-tomers, they envision wireless mesh networks usingshort-range links to cover neighborhoods. Onestart-up, Skypilot, plans to use standard 802.11bradios in the home connected to rooftop unitsbased on 802.11a with added mesh networkingsoftware. Omnilux plans to use free-space opticstechnology, combined with mesh networking.Free-space optics uses lasers that operate in the visible light range of the spectrum, above the radiofrequencies. It is therefore technically outside thescope of FCC licensing, which applies only tocommunication “by wire or radio.”

Personal-Area Networks (802.15)Moving the opposite direction from MANs, personal-area networks (PANs) are designed forvery short-range connections, no more than a fewdozen feet. WiFi can cover these distances, butbecause WiFi devices are designed to serve largerareas and provide relatively high-bandwidth con-nections, they require more power and have higherequipment costs than would be necessary for close-in, low-speed tasks such as communicatingbetween a mobile phone and a headset.

PAN applications are essentially “cable replace-ment.” They are tasks that people perform todayby stringing wires between devices, but that couldbe done with more freedom if the wires weren’tnecessary. These include scenarios such as printingfrom a laptop computer to a nearby printer, send-ing voice signals between a cordless phone and abase station, and pulling a contact from a PDA to amobile phone. There are also high-capacity cablereplacement tasks involving rich media, such assending music between an Internet-connectedhome server and a home theater or stereo system.

The standards body for PANs is IEEE 802.15.Bluetooth was the first prominent PAN standard,incorporated into IEEE 802.15.1. Based on technology originally developed by Ericsson, it is supported by a private industry standards body that

29

now has more than 2000 members. Bluetooth pro-vides approximately 1 Mbps connections over 30feet by automatically creating network clusters ofnearby devices.

Bluetooth received a great deal of media attention when the consortium was firstannounced, because of its heavyweight backers andexcitement about the potential of all things wirelessat the time. However, interoperability issues andquestions about where Bluetooth really fits in themarket have limited adoption. Bluetooth mobilephones, PDAs, and laptops are now available, andcosts are coming down as volumes increase.

Even shorter range and lower speed than Blue-tooth is Zigbee (802.15.4). The protocol, originallydeveloped by Philips, is optimized for applicationssuch as distributed sensor networks, which mustonly send a few kilobits of data at a time. Zigbee,like Bluetooth, is currently supported by an allianceof companies including Honeywell, Mitsubishi, andMotorola. It works on various unlicensed frequen-cies, providing 20-250 kbps connections over 30 to 200 feet. The big advantage of Zigbee is its low-power consumption and low cost, which are essential for remote monitoring and sensing.

Ultra-wideband (UWB) is more than a PANtechnology. However, its initial applications in thecommunications market are for short-range, PAN-type uses. UWB is a form of spread-spectrumtransmission that uses such a wide band, and suchlow power, that it can “underlay” with licensedusers in the same band. The UWB signal appearsas background noise to other transmitters. Thenature of the technology gives it several otheradvantages, including very low power consump-tion, security, and penetration of walls. Until theFCC’s decision in early 2002 to legalize UWB forcommunications, its primary application was forground-penetrating radar and military uses.

Today, companies such as XtremeSpectrum andTime Domain are building UWB chipsets target-ing short-range, high-bandwidth data applications.UWB systems can deliver 100 Mbps or more overshort distances, which makes them ideal for usessuch as streaming audio or video between mediadevices within the home. Several companiesincluding HP, Kodak, Philips, Motorola, Samsung,

Sharp, Time Domain, and XtremeSpectrum havecreated the WiMedia Alliance to promote thisapplication. Though UWB was late to the PANparty because of its recent approval, it has recentlybeen gaining adherents within the 802.15 group.Most of the proposals for the forthcoming802.15.3a standard, a higher-speed PAN protocol,involve some form of UWB.

Success StoriesToday’s WiFi MarketsThere are four major WiFi markets today: homenetworking, corporate/campus networking, com-mercial hotspots, and public access.

HOME NETWORKING means sharing an Internet connec-tion, or a peripheral such as a printer, among morethan one PC. Vendors such as Proxim, Netgear,Linksys (now part of Cisco), D-Link, and 2Wirehave sold millions of access points and cards to end-users for this purpose. Broadband service providersare now getting into the game, recognizing thatthere is significant demand for home networking asan add-on to high-speed Internet access.

CORPORATE OR CAMPUS ENVIRONMENTS (in both the businessand university sense) are slightly different thanhomes. Except for very small offices, multiple accesspoints are required to cover the facility. These customers generally want security, authentication,and management capabilities to operate the WiFinetwork in conjunction with their existing wirednetworking infrastructure. All the major networkingvendors, such as Cisco, Lucent, and Nortel, nowhave substantial corporate WiFi customer bases.

HOTSPOTS are access points available to anyone withina location, such as an airport, a café, or a hotellobby. Sometimes the hotspots require a fee foraccess. Vendors such as Wayport and Mobilestar(now T-Mobile) deploy hotspots in locations thatreceive significant foot traffic, as both a money-maker and an incentive for more traffic. The best-known hotspot deployment is at Starbucks coffee-houses, now operated by T-Mobile. Aggregatorssuch as Boingo and iPass allow users to pay one feeand access multiple networks of hotspots.

30


Radio Revolution

31

NYCwireless: Evolution of a Wireless User Group

To followers of the wireless broadband revolution,New York has been a hotspot of activity among U.S.cities, and the NYCwireless user group has been lead-ing the movement. The group began as an informalnetwork of early WiFi adopters who placed accesspoints on their apartment windows to share theirbroadband connections with the public parks belowtheir buildings. As the trend gained acceptance, theusers organized to form NYCwireless, a non-profit, volunteer organization, to encourage others to sharetheir broadband and foster an ethic of free publicInternet access across the city.

“New Yorkers live in cramped quarters, and ourgoal has been to get people out of their apartmentsand into the public parks,” says NYCwireless volun-teer Dustin Goodwin. The group considers ubiquitousbroadband access to be a public amenity equivalentto streetlights or water fountains. However, it’s diffi-cult, if not impossible, to provide public parks withwired broadband access because of constructionimpediments on historic or public land. Cheap, andeasily installed WiFi technology allowed apartmentdwellers with good line-of-site to their parks toinstall the 802.11b transmitters and address theproblem for themselves.

Volunteers from NYCwireless have built networksin Bryant Park, Bowling Green Park, and TompkinsSquare Park, among others. This past year, foundingmembers of the group formed a consulting firm,Emenity, to deploy six more public hot spots in lowerManhattan for the NYC Downtown Alliance. The newcompany was started to provide service to commer-cial clients, but their mission of building public accessnetworks remains intact.

Emenity has recently built a public network inUnion Square Park. This project is unique in that itrelies on a wireless backhaul to connect to the Inter-net provided by the commercial wireless broadband

provider TowerStream. Most public access points in the city ultimately access the Internet via a DSLInternet connection.

The efforts of NYCwireless have not gone unno-ticed by broadband service providers. Some providershave slapped “acceptable use” clauses on their sub-scriber contracts in an effort to discourage wirelessbandwidth sharing. One large cable operator hasbeen accused of sending out a WiFi “sniffer” to scourthe city in search of access points leading back totheir customers’ connections to close down the transmitters.

However, as WiFi use has reached a critical mass,more broadband providers are trying to enter thepublic space arena. Speakeasy, Inc., a national DSLreseller, now offers “WiFi Netshare,” a service thatallows users to resell their broadband connections toneighbors, with Speakeasy handling the billing. AndVerizon DSL has built a number of hotspots in NewYork that are free to their DSL home subscribers.

NYCwireless volunteer Dustin Goodwin sees thecommercial efforts to deploy WiFi in the city as adirect response to NYCwireless’ success. While somesee the entrance of commercial players into the public space as a threat to free access, others see the development as an important step to recognizingWiFi as a free public amenity that companies andorganizations should provide as a value-added serviceto their constituents.

Now that wireless broadband has gained a footholdin New York City parks, Goodwin says that NYCwire-less is expanding its mission to resemble a volunteer“Geek Corps” for communities without affordablebroadband Internet. NYCwireless volunteers havetrained residents of a community housing organizationto build and maintain their own wireless network,which will provide more than 50 residents with pri-vate, high-speed connections. This effort is part of agrowing trend among wireless community groupsacross the country to bring affordable broadband tounderserved communities. – Matt Barranca

32


The spread of hotspots has been remarkable.Boingo now has more than 1,200 nodes on its net-work. T-Mobile operates 2,300 hotspots, includingStarbucks coffeehouses, Borders bookstores, Amer-ican Airlines Admirals Clubs, and terminals at fif-teen airports in North America. It has announcedplans to put hotspots in the more than 1,000Kinkos copy shops throughout the U.S.

This is only a fraction of the total. One Websitelists more than 5,000 hotspots worldwide, includ-ing both commercial and community nodes.41 Andthat’s just the beginning. According to PyramidResearch, 1,000 hotels offer WiFi access today,mostly in lobbies and meeting rooms, and 25,000will have it by 2007.42 Cometa, a joint venturefunded by AT&T Wireless, IBM, Intel Capital,and venture capital firms 3i and Apax Partners,plans to build 20,000 hotspots using a wholesalemodel, with its first customer being McDonald’s.

In addition to the access points intended forpublic consumption, most private WiFi nodes donot use any security mechanism, making themavailable to anyone within range. A map of Man-hattan prepared last year by the Public InternetProject, whose volunteers systematically drovethrough the streets with WiFi-sniffing equipment,shows public and home access points already cov-ering most of the island (see Figure 12). Prelimi-nary results from the group’s 2003 survey suggestWiFi density has become significantly greater sincethe original map was published.

PUBLIC ACCESS means providing free connectivity tousers within a particular area. Sometimes this isdone by private groups sharing their own networksor promoting the concept of ubiquitous wirelessconnectivity. In other cases the access is funded bypublic organizations, non-profits, or corporationsas a type of civic amenity.

In all, the number of regular wireless LAN usersis expected to grow sevenfold in the next fouryears, from 4.2 million to 31 million, according tothe Gartner Group.43

Independent Community Access PointsThere are dozens of community WiFi organiza-tions in cities throughout the United States, and

around the world. Typically, these groups are earlyusers of WiFi technologies. They come together inphysical meetings and through online discussionsto share experiences, ask questions, and experimentwith new technologies. In many cases, they installtheir own WiFi infrastructure, with access open toall. They deploy these hotspots where they can orwant to, rather than follow some master plan.

The most active community WiFi groupsinclude SF Wireless in San Francisco; NYCwire-less in New York; SeattleWireless; and the Per-sonal Telco Project in Portland, OR. The infra-

FIGURE 12 – WIFI ACCESS POINTS IN MANHATTAN

structure is typically contributed by members,though increasingly these community groups haveformed partnerships with local businesses and com-munity development organizations.

Hotspots as Civic AmenitiesIn several cities, hotspot deployments are beingfunded by civic organizations or corporate spon-sors as civic amenities, like parks or playgrounds.In Manhattan, Intel and the Bryant Park Restora-tion Corporation supported a project by NYCwire-less to establish a WiFi network in Bryant Park, a

Radio Revolution

popular outdoor gathering place in midtown. TheUniversity of Georgia has funded a network ofWiFi hotspots covering all of downtown Athens,GA. In Long Beach, CA, the Long Beach Eco-nomic Development Bureau partnered with severallocal businesses to establish a WiFi network cover-ing several downtown blocks, with plans to expandit throughout the city’s business district.

For the civic groups involved, the costs of theseWiFi networks are relatively minor, especiallywhen businesses become involved and provide freeInternet bandwidth and other services. Wireless

33

A Community Access Model for the Last Mile

While the success of commercial WISPs has gener-ated much attention, grassroots community accessnetworks or CANs are the originators of the unli-censed movement. Most CANs are groups of like-minded individuals sharing a similar philosophy—thatcitizens should have open, inexpensive, and ubiqui-tous access to the Internet. Using affordable andeasily installed WiFi technology, community mem-bers in Seattle, New York, Austin, San Francisco, Port-land, Oregon, and Athens, Georgia have builtexpanding networks of independently maintainedwireless access points.

Most CANs provide access to public spaces, how-ever some groups have made forays into residentialspace, by connecting neighborhoods with centrallyplaced access points. One such organization is theBay Area Wireless Users Group (BAWUG), an informalgroup of wireless early adopters who began mountingWiFi transmitters on the roofs of their homes to giveneighbors free or shared-cost Internet connections viatheir DSL and cable lines. While the cable and phonecompanies didn’t approve of the practice, consumersdid and access points began popping up all over thecity. There are now more than 25 BAWUG accesspoints in the area.

But BAWUG has not stopped there. Under theleadership of Tim Pozar, a telecommunications engi-neer and one of BAWUG’s founders, the group haslaunched the Bay Area Research Wireless Network(BARWN). BARWN is an active wireless network witha mission to discover the best technical solutions tobring wireless broadband to remote and economicallydisadvantaged communities.

BARWN has set up two centrally located accesspoints atop the San Bruno Mountain and Potrero Hillin south San Francisco, allowing anyone within an 8-mile radius to point a 2.4 GHz antenna at theBARWN towers to share the 11Mbps of bandwidththey provide. Pozar says that a third public accesspoint is soon to be installed on Yerba Buena Island inthe San Francisco Bay, which will link to East Bay andlight-up an underserved area called Treasure Island.

All of these access points are constructed with non-proprietary equipment and open protocols to keepcosts down and to learn what technologies can bemost easily adopted by lower income communities.

As evidence of the network’s stability and flexibility,BARWN is working with the City of San Francisco touse this network for public safety communications—such as earthquake or disaster response. Pozar saysone application for the unlicensed service would be toprovide streaming video of a disaster site to commandcenters to evaluate response tactics. – Matt Barranca

connectivity is becoming a benefit that draws people into downtown areas.

Internet Connectivity for Rural CommunitiesHigh-speed Internet connections are available tomost of the U.S. population today through digitalsubscriber line (DSL) or cable modem service. However, there are still tens of millions of Americanswho live in rural or otherwise underserved areas,where such broadband offerings are not yet avail-able. In some cases, they are unlikely to be avail-able any time soon. Technically minded citizens in

some of these communities have seized upon unlicensed wireless as an alternative route to provide connectivity.

In Laramie, WY, a group of technologists ledby Brett Glass established LARIAT, a non-profitcommunity wireless network. It has been in operation since the mid-1990s, originally usingpre-WiFi unlicensed equipment in the 900 MHzband. A similar effort is MagnoliaRoad.net, acooperative in a rural part of Colorado that isoffering WiFi connectivity to local residents who have no other good broadband option.

34


Revolution in the Rural Last Mile: Unlicensed Spectrum Closing the Technology Divide in Northern Virginia

Despite their proximity to Northern Virginia’s Internetbackbone, many towns in Loudoun County have nobroadband access. The mountainous western regionsof the county are far from the technology infrastruc-ture of Northern Virginia where companies like AOLand VeriSign reside. However, because of license-exempt wireless activity, the technology divide acrossthe county is starting to close.

After the technology bubble of the late 90s burst,Northern Virginia lost as many as 30,000 jobs. Manylaid-off professionals accustomed to broadband con-nections at their work started their own businesses orbegan working from their homes, creating a largedemand for high-speed home services. One start-up,Roadstar Internet, is trying to meet that demand withan expanding rural wireless network.

Started in the autumn of 2002, Roadstar Internetconnects more than 150 rural households and smallbusinesses relying only on unlicensed spectrum. Mostwireless subscribers do not know exactly how theirservice operates. What matters most to users is notthe technology behind the service, but that theirconnections are fast and reliable. The Roadstar net-work is similar to many other WISP efforts, using acombination of point-to-point connections for the

long-distance transmissions, and point-to-multipointtransmissions to connect neighborhood access pointsto subscribers.

The first leg of the network travels 18 miles from amountaintop transceiver using 5 GHz bands andOFDM (Orthogonal Frequency Division Multiplexing)point-to-point technology. OFDM transmissions makeefficient, and secure, use of spread spectrum by divid-ing data into packets and encoding it over multiplefrequencies, without requiring perfect line-of-site.

Long distance point-to-point transmissions are thestandard for rural WISPs seeking to reach larger pop-ulation pockets. Under Part 15 rules for unlicensedusage, the FCC allows operators to make point-to-point connections without reducing Transmitter PowerOutput (TPO) for the 5.725 GHz and 5.825 GHzband. Because of this regulatory latitude for narrowbeam transmissions, providers are able to reach longline-of-site distances with relatively low power.

The Roadstar network makes final, last-mile con-nections within neighborhoods by using modifiedWiFi wireless access points mounted on customersilos, barns, and rooftops. WISPs are able to transmitdistances greater than the 300-foot standards forWiFi technology by creating sectorized cells withhigh-gain, directional antennas. These last-mile con-nections on the 2.4 GHz band are the result of goodplanning and engineering, and typically reach two tothree miles. – Matt Barranca

Meanwhile, Dewayne Hendricks of the DandinGroup is spearheading efforts to provide wirelessInternet connectivity, using WiFi and other technologies, on several Indian reservations in theU.S. and Canada. More than 1,000 commercialWISPs are providing similar wireless broadbandservices, mostly in underserved rural areas acrossthe nation (see sidebar, page 34).44

Internet Connectivity for Low-Income AreasHigh-speed connectivity has important benefits forlow-income and underserved communities. Broad-band Internet access opens the door to educational,informational, job-related, benefits, health, andother materials. However, the costs of wiring low-income facilities such as public housing complexes

Radio Revolution

has traditionally been prohibitive, given that mostresidents cannot afford to pay typical monthlybroadband prices. WiFi is one answer.

In Boston, an MIT graduate student namedRichard O’Bryant led an effort to put free WiFihotspots in Camfield Estates, a 102-unit public housing development in the Roxbury area, withfunding from HP and Microsoft. In Philadelphia, theUnited Way is building two WiFi hotspots in thepoor section of West Philadelphia, which will offerbroadband Internet access for $5 to $10 per month.It plans to give away computers and wireless cards topeople in the community who cannot afford them.In Portland, OR, a non-profit called One Economyis putting WiFi connections into three public hous-ing developments, serving more than 500 residents.

35

Future Scenarios

Expanding the Space of PossibilitiesWe have only scratched the surface ofwhat dynamic wireless systems can do.The growth of WiFi has been so striking, its possibilities so exciting, thatWiFi has become virtually synonymouswith unlicensed wireless and open spectrum. This is a mistake. WiFi is notthe culmination of the wireless story; itis merely the end of the beginning.

WIFI HAS TWO GREAT LIMITATIONS: ITS

PROTOCOL, AND ITS SPECTRUM ENVIRONMENT. The engineers who created the 802.11family of protocols had no idea that WiFiwould take off the way it has, and wouldbe used in so many different deploymentscenarios. They were creating a wirelessEthernet standard, parallel to the wiredEthernet standard that is the basis formost office computer networks today.Thus, WiFi has limited range, and arouting layer that isn’t particularly goodat mesh networking, quality of service,interference management, security, ormany other functions that are importantfor many of its potential markets.

At the same time, the government regulators who established the 2.4 GHzand 5 GHz unlicensed spectrum bandswhere WiFi operates had even less ideaof what was coming. Especially for 2.4GHz, they were looking at “industrial,scientific, and medical” equipment, anddevices such as cordless phones. Therules they created for managing thosebands have been highly successful, butthey were hardly designed to maximizethe potential. Based on our current experience, we can design rules expresslyto promote efficiency and innovationthrough unlicensed wireless technologies.

First and foremost, this means makingavailable more spectrum for unlicenseduses, whether it be through dedicatedunlicensed bands, shared underlayaccess, or opportunistic sharing. Second,it means putting in place minimal rulesfor that spectrum, which may be as littleas power limits, to foster an environmentof efficient cooperative development.

The wireless future, under any sce-nario, is likely to be marked by increas-ingly pervasive but non-uniform connec-

37

38

Future Scenarios

An Unlicensed Education: A Wireless Modelto Connect Rural School Communities

While U.S. school districts have been issued the com-mand to “leave no child behind,” many rural schools arewithout the resources to bring broadband Internetaccess into their classrooms. This last-mile problem pres-ents hardships not only for schools, but also for localhouseholds and businesses unable to fully participate inthe information economy. A public/private partnershiphas been formed in western Pennsylvania to use unli-censed spectrum and the social capital of local schooldistricts to address the last mile on their own. Theefforts of the Broadband Rural Access Information Network (BRAIN) have yielded great results connectingrural areas, and their example could provide a model forrural school communities across the country.

The BRAIN effort began with the vision of a schooldistrict superintendent, Andy Demidont, and the helpof a large regional WISP, Sting Communications. Demi-dont wanted to provide high-speed access to Rock-wood High School and Kingwood Elementary School inmountainous Somerset County. The schools’ existingdial-up accounts were expensive, and rendered connec-tion speeds barely surpassing 14 kbps.

Relying on technical guidance from Sting Commu-nications, and using grant money awarded from theIndividuals with Disabilities Act and E-Rate discounts,the school district installed wireless access points onthe roofs of both schools, turning both schools intostate-of-the-art, wireless hotspots.

In total, Sting Communications installed three tow-ers, creating a pie-shaped hot zone using the 5.8 GHzand 2.4 GHz license-exempt bands. The RockwoodHigh School gymnasium hosts a 100-foot tower that

transmits to a 150-foot tower located at KingwoodElementary School. The two towers share a narrowbeam, point-to-point connection with a third towerowned by the local Seven Springs Ski Resort.

Simply bringing the technology to the area wasn’tthe end goal – using the network to connect theschool with the community is the ultimate design ofthe project. Both schools have put many classroomoperations on-line. Teachers use Palm Pilots and laptops to track student progress and record grades,which are available to parents online.

The project also gives community residents a chanceto purchase access from the school’s network, with theschool district serving as a WISP for the area. Sting hasinstalled access points in many neighborhoods, and thecompany is offering subscription rates between $11and $20 per month, depending on the number of sub-scribers the school can attract. Currently, thirty-fivefamilies have been connected, with an additional 65families expected to be online in the coming months.

From the project’s onset, Sting hoped their school-based approach could be replicated in other rural com-munities. Building on what they have learned in Somer-set County, Sting has built a much larger network inCambria and Clearfield Counties to connect four moreregional school districts. Sting vice president Bob Rolandsays that this new network spans an 1100 square-milearea, and uses both 5 GHz frequency-hopping spreadspectrum and 802.11 connections for the last mile.

For the next phase, BRAIN has applied for an additional $7.4 million grant from the USDA’s RuralUtilities Service to “light-up” a wide corridor betweencentral Pennsylvania and Maryland. If successful, thiseffort could provide a model for building a wide-arearegional network, one school at a time. – Matt Barranca

tivity. No wireless technology, let alone one serviceprovider, can address all the markets and deploy-ment scenarios, from short-range low-bandwidth tolong-distance broadband. Even if unlicensed sys-tems succeed beyond anyone’s wildest dreams, therewill be a need for licensed services for many years.Even if there is soon WiFi in every coffeehouse,

getting it in every dry cleaner will take longer, aswill getting it in every train and airplane.

The downside of such a heterogeneous environ-ment is that everyone is not connected all the time,and any one system or technology will provideonly a small percentage of what connectivity doesexist. The natural impulse in communications is to

try to build all-encompassing networks, but some-times that isn’t the best approach. WiFi hotspotsare spreading because they are cheap, funded byusers or facility owners, and go where there isdemand today. They don’t yet go where there isn’tdemand, but the good news is that users need notpay for the extra cost of putting access points there.As software-defined radios mature, they will makeit possible to stitch together some systems at theend-user device. In general, though, the real question is not how to provide ubiquitous wirelessconnectivity in the abstract, but how to addressconcrete needs and market opportunities.

The scenarios below represent examples ofopportunities that unlicensed wireless technologiescould address. They are relatively straightforwardextensions of existing technology. Most, however,will require either spectrum reforms to expand the space available for unlicensed devices, or at the very least no regulatory actions that wouldhamstring unlicensed devices in the existing areas.

The Last Wireless MileBroadband connectivity to homes is a topic of greatconsternation in the communications industry today.Telephone companies are deploying DSL and cableTV operators are deploying cable modem systems.However, many millions of Americans still have nei-ther available to them, and a greater number haveonly one option. Prices of these broadband services,approximately $50/month, are high compared to therest of the world. These services are generally asym-metric, providing far more bandwidth down to theuser than up from the user to the Internet. Com-bined with terms of service restrictions, this architec-ture limits users from running home servers andother actions. For business users, who typically needhigher bandwidth than homes, the only viable optionis often a traditional T-1 line at $1,000 per month ormore. There is thus great interest in alternatives.

Basic WiFi or its variants, 802.11a and 802.11g,cannot simply be put into service for last-miledeployments. WiFi is a short-range technologydesigned primarily for connections to a nearbyhotspot. Even if every home in a neighborhoodhad a WiFi access point, few of those nodes wouldsee one another and there would be no mechanism

Radio Revolution

to link them together. Even if signals could reach aneighborhood access point, backhaul costs wouldbe significant, because every access point wouldneed a wired connection to a T-1 or larger circuit.

We can, however, envision scenarios for unlicensed wireless last-mile connectivity based ontechnology that is in the market or likely to be soon.For homes, a mesh network configuration could beused to shorten link lengths, increase robustnessthrough alternate traffic paths, and address impedi-ments such as trees. A range-extending technologysuch as Vivato’s phased-array antennas could also beemployed to receive signals from a cluster of nodesin a local area. For businesses or residences wantingmore bandwidth, an 802.16 wireless MAN technol-ogy could be used to deliver tens of megabits persecond over many miles. This same technology, or avariant, could be used to reduce the costs of back-haul, replacing costly wired connections.

An end-user would buy a device, whether a dedicated piece of wireless hardware, a broadband“residential gateway,” or a piece of general-purposehardware such as a laptop. The device could bedesigned to operate with a particular unlicensedwireless network, it might be deployed by a serviceprovider, or it might have “discovery” capability toautomatically locate and connect with nearbyaccess points or wireless end-users.

It is typically assumed today that last-mile broad-band networks are designed to provide access to theInternet (see Figure 13). Certainly, any broadbandcustomer will want to access Internet-based servicesavailable through the World Wide Web, as well as

39

FIGURE 13 – WIRELESS LAST MILE: In one wireless last-mile

scenario, a transmitter tower connected to the Internet

transmits in a point-to-point connection to a community

access point, which makes point-to-multipoint or mesh

network connections to households.

Netbackbone

Transmitterat regionalbackboneconnection

Antennaat localaccess point

global email, instant messaging, and other applica-tions. However, these are not the only things thatan unlicensed last-mile wireless network coulddeliver. There is often value in online communica-tions within a community, especially when thatcommunity has not previously had a high-speed,always-on network. These range from intra-com-munity email to school bulletin boards to decisionsof the city planning board. They could be called

community intranet appli-cations. Because the trafficis local, there is no need toconnect to a backboneprovider. In fact, there maynot be a need for anyprovider at all. If users pro-vide their own wirelessnodes, and no node isoverwhelmed with traffic,the community-wide net-work would function peer-to-peer, much as local areanetworks in businesses do today.

When the networkneeds to connect to theInternet, an economicproblem emerges. Internetbackbones charge for usebased on bandwidth con-

sumed. Even if Internet-based services and contentare a minority of traffic on a wireless communitylast-mile network, there still must be backhaul con-nections to the Internet, and these must be paid for.There is a “free-rider” problem if the node con-nected to the backbone must bear all the costs,independent of whether there is any congestionacross the unlicensed wireless links between thecommunity nodes. It may be necessary to developsome sort of pricing mechanism for end-users insuch a situation. However, pricing could be imple-mented in several ways. It does not require centralproviders or per-packet settlement charges. Forexample, a cooperative could collect dues from allcommunity members and apply those to the back-haul charge, with some limits on each user’s band-width to prevent free riding.

Interoperable Public Safety CommunicationsToday, public safety agencies use many differentwireless communications systems. Many of them useoutdated technology. Few if any can talk to oneanother. In an emergency, if the fire department can’tcommunicate with the police, the consequencescould be disastrous. On September 11, 2001, fire-fighters were trapped in the World Trade Centerbecause they were unable to learn from other publicsafety officers outside that the buildings were aboutto collapse. Stories abound, from 9/11 and othertimes, of firemen using commercial mobile phonesbecause they had better performance and a wideraudience than their expensive private radios. Andwhen these networks go down, everything goesdown with them. Unfortunately, public safety organi-zations are saddled with many legacy communica-tions systems that are costly and difficult to upgrade.

As software-defined radios (SDR) mature, theycould replace the cacophony of devices with a singleset of devices. One phone handset, PDA, or laptopcould tap into any of the existing systems. A fire-fighter arriving on the scene could instantly checkpolice communications as well as data transmissionsproviding essential information directly from dis-patchers, such as building maps. Such a system wouldrequire robust security and authentication mecha-nisms, but these could also be built into the devices.The result would be both lower costs and moreeffective systems for critical public safety services.

Adaptive Mobile PhonesMobile phone networks today are self-containedentities. In the U.S., for example, there are six com-peting national networks. Each has its own networkof transmission towers. If you are within range offive of those towers, but not the one for your service provider, you won’t get service. Things are alittle better in Europe, where universal adoption ofthe GSM standard allows for more roaming agree-ments between carriers, but each carrier still mustmaintain its own complete network.

As mobile phones evolve into SDR devices, thestructure of the business may change. Carriers will be able to share infrastructure much morewidely, because their subscribers will be able totransparently access transmissions from whatever

40

Future Scenarios

“I am hopeful that

unlicensed operations

will…eventually provide

a last-mile application to

connect people’s homes

to the Internet, offering

a real alternative to

telephone wires, cable,

and satellite connections.”

— FCC COMMISSIONER KEVIN MARTIN

tower is closest to them, regardless of what frequency band or encoding mechanism it uses.

Further, there are an increasing number of localconnectivity points, such as WiFi hotspots, that areseparate from the wide-area wireless networks.When a user was within range of a hotspot, adaptivemobile phones could transparently shift communica-tions from voice transmissions over the mobile net-works to voice-over-IP or packet data connectionsthrough the WiFi infrastructure. Tapping these local

Radio Revolution

nodes would reduce costs, avoiding the need to senddata long distances over wireless networks, andwould give users the maximum possible capacity,since WiFi hotspots tend to offer substantiallygreater bandwidth than wide-area data networks.

Personal Broadcast NetworksToday, broadcasting is the domain of the few. Onlycompanies with licenses have access to the airwaves todeliver programming. Using a combination of the

41

Think Globally, Act Locally:Two WISPs Come of Age

At the onset of the unlicensed movement, most WirelessISPs (WISPs) were scaled-up WiFi hotspots built fromadapted, off-the-shelf, 802.11b equipment. But unli-censed devices have advanced beyond WiFi to includefrequency-hopping, non-line-of-sight transmitters, andother technologies such as high-gain directionalantennas that reach distances well over 20 miles. Themost successful WISPs have outgrown the WiFi modeland evolved into sophisticated wide-area networkswith a high standard for service, security, and stability.

AMA*TechTel Communications of Amarillo, Texasis one example. With more than 4,000 users on theirlicense-exempt network, AMA is one of the country’slargest regional carriers of wireless broadband. AMA’s63-tower deployment is a 20,000 square-mile, con-tiguous network providing secure service to numeroustowns, three college campuses, multiple school sys-tems, hospitals, and banks.

AMA relies on equipment manufactured by Alvarionto access the 900 MHz, 2.4 GHz, and 5 GHz unli-censed bands for backhaul and point-to-multipoint con-nections. The network’s expansiveness is due to AMA’spartnership with Attebury Grain, a large grain storagecompany that initially contracted AMA to wirelessly con-nect their grain elevators to the commodities market.After the project, the two companies partnered, usingAttebury’s numerous grain elevators as transmitter tow-ers to provide service to local towns and businesses.

While AMA’s wide-area deployment may be a daunt-ing example for WISP entrepreneurs, this ambitious net-

work isn’t the only successful model. Hundreds of start-up WISPs have relied on skillful engineering to buildsmaller networks for remote communities. Prairie iNet,of Des Moines, Iowa has used this localized approach tobuild a network of unlicensed wireless oases in morethan 120 rural communities in Illinois and Iowa.

Neil Mulholland, Prairie iNet CEO, estimates that toreach their standard for “carrier class” service, the com-pany invests $100,000 in each population center forbase station, customer premise equipment, and a wire-less link to their DS-3 Internet connection. Ideally, thecompany covers their investment in each local networkafter the first 75 customers. The company has morethan 4,000 subscribing customers in total.

Both models – AMA’s wide-area design and PrairieiNet’s localized networks – present varying benefits.Regional utility companies can learn from AMA andleverage their tower infrastructure to provide virtualprivate networks for campus clients and basic servicefor their residential customers. The East Bay MunicipalUtility District in Oakland, California has installed theMotorola Canopy system; Owensboro Municipal Utili-ties in Kentucky and Wheatland Electric Cooperativein Kansas have both installed Alvarion networks; andMidwest Wireless, a rural cellular provider, has built abroadband business using unlicensed spectrum.

Prairie iNet’s localized model has been replicatedin smaller scales in rural areas across the country. In-Stat/MDR, a technology research company, esti-mates that there are up to 1,500 start-up WISPs inoperation. Low start-up costs, numerous equipmentoptions, and high consumer demand account for thisgrowth. – Matt Barranca

techniques described in this paper, it is possible toimagine a world in which anyone can be a broadcaster.

As each user sends out video streams, they wouldbe relayed by other users wherever infrastructurewas unavailable. Cognitive radios would seek outfree space in the spectrum to carry the signals.Content creators could contract with operators of virtual broadcast networks who aggregatedtogether reliable high-speed connectivity to reachan audience, creating a bottom-up divisionbetween different classes of traffic.

Who would want to have their own broadcastnetwork? Some people would want to deliver thekinds of creative programming available on televi-sion today. These personal wireless networks wouldbecome a much more powerful version of the alternative outlets available today, such as publicaccess channels on cable TV systems, public broad-casting stations, low-power FM radio stations, andthe Web. If consolidation in the media distributionbusiness threatened the diversity of voices availableto viewers and listeners, personal broadcast networks would provide a powerful antidote.

But the existing market for heavily produced,

mass-market content would only be a small part ofthe total. Working parents would use personalbroadcast networks to tap into video images oftheir children at home, streamed from Webcams totheir mobile phones. Distance learning coursescould be delivered on demand, or specific instruc-tional modules could be delivered dynamicallywhen and where they were needed. Need tochange a flat tire and don’t know what to do? Need on-the-spot medical advice? Use a personalbroadcast network to watch an instructional videoor establish a videoconference with an expert.

Predicting the future is dangerous. Any scenarioswe envision today will likely miss the specific kinds ofapplications and content that will be popular tomor-row. But that doesn’t matter. The infrastructure ofemerging wireless technologies can be adapted towhatever turn out to be the killer apps. Wireless net-works built using intelligent, dynamic techniquesfrom the computer and networking industries willfeature radical flexibility. Network owners need notpredict uses in order to shape their infrastructurebuild-out, because there will be no owners, infrastruc-ture, or build-out in the current senses of the words.

42

Future Scenarios

WiFi Calling: Campuses Turn to WiFi for Voice Applications

Since the beginning of the wireless movement, col-lege and corporate campuses have been fertileground for extensive WiFi networks. The latest trendfor campus organizations is to bypass local telephonecarriers and use their unlicensed wireless networks tosupport voice applications.

In one example, the University of Arkansas invested$4 million in Cisco’s call-processing software, CallMan-ager, to traffic local calls over the University’s existingWiFi network. The University has reduced their monthlytelephone service fees from $530,000 to $6,000 amonth. At this savings rate, the University shouldrecoup their call-processing investment in six months.

Dartmouth College has a similar program, offeringsoftware to incoming freshman that turns wireless lap-

tops and PDAs into “softphones.” Dartmouth imple-mented their voice over wireless local area network(VoWLAN) when they realized they were spending moremoney billing students for long-distance calls than theywere taking in. The unlicensed VoWLAN provides a sim-ilar quality of service to traditional telephone service,but without the billing and administrative costs.

Hospitals have also been early adopters ofVoWLAN technology. One company, Vocera, has cre-ated an 802.11b communications device that doctorsand nurses wear on their uniform collars to communi-cate with each other remotely. Users speak the nameof the person they’d like to contact into the device,and they are instantly connected through theVoWLAN. The device has the potential to eliminateintercoms and speakers that broadcast announce-ments or pages meant for only one person to anentire hospital floor. – Matt Barranca

Policy Recommendations

Governments should recognize andembrace the tremendous potential ofthe radio revolution. In the U.S., theFCC’s November 2002 Spectrum PolicyTask Force report45 acknowledged thattechnological changes allow for newforms of spectrum access and interfer-ence management. It therefore proposedexpanded use of the commons model. InJune 2003, the Bush Administrationformed a task force to examine govern-ment spectrum use, with a similar mandate to propose reforms.

The FCC has begun or announcedplans for a raft of spectrum reform pro-ceedings. These include: unlicensedsharing of the broadcast spectrum;changes to allow unlicensed access tounderused education spectrum in the 2.5GHz band currently allocated forInstructional Fixed Television Service(ITFS); incentives for unlicensed deploy-ment in rural areas; investigating theimpact of cognitive radio and the unlicensed allocation of extremely highfrequency spectrum (above 50 GHz);

and, a novel method for noise floor base-lines called “interference temperature.”

Because radio signals do not respectpolitical boundaries, spectrum policy isinherently international. Equipmentvendors can better justify the investmentneeded to develop new products whenthey can foresee a global market. Hence, international harmonization ofspectrum bands is valuable, especiallyfor unlicensed bands. Participants at the2003 World Radio Conference agreedto expand the globally unlicensed spec-trum in the 5 GHz range, adding 255MHz to the existing allocation.

Despite these encouraging develop-ments, more needs to be done. Govern-ments cannot merely sit back and waitfor technological development to elimi-nate spectrum scarcity. Spectrum policiesaround the world remain centered onwhat the FCC calls a “command-and-control” model, more appropriate for theindustrial era than for the coming age ofdigital networks. Though governmentregulation of the radio spectrum was

43

designed to manage and mitigate physical scarcity ofthe airwaves, that regulation now creates scarcity.

As a general matter, the federal governmentshould recognize that the objective of spectrumpolicy is not to minimize interference, but to maxi-mize usable capacity. It should move away fromcommand-and-control toward more flexibleapproaches, recognizing that there is value in adiversity of legal regimes. In experimenting withexclusive property rights and secondary markets, it should balance the potential gains against the

opportunity cost of mak-ing that spectrum availablefor unlicensed sharing inthe future. Wirelessdevices will continue tobecome increasinglysophisticated and inexpen-sive. Policy decisionsshould take into accountthe possibility that what isimpractical now may becommonplace in thefuture...if there is room for innovation.

Specifically, policy ini-tiatives should proceed in four areas:

z MORE DEDICATED UNLICENSED SPECTRUM

The FCC and other government agenciesshould continue their efforts to identify addi-tional spectrum for unlicensed use. Some spec-trum may become available through return ofanalog broadcast television licenses as part of thedigital TV transition,46 or from guard bandsestablished to protect obsolete devices. Otherspectrum could be shifted from government ormilitary use, or through restructuring of under-utilized bands such as the UHF television andMMDS/ITFS frequencies.47

Unlicensed spectrum at low frequencies,below 2 GHz, is particularly important. Transmissions at those frequencies can more eas-ily penetrate trees, walls, and other obstacles, akey consideration for last-mile broadband appli-cations. They also use less power, and thus makemore efficient use of batteries, a key considera-

“Increasing demand for

spectrum-based services

and devices are

straining longstanding,

and outmoded,

spectrum policies.”

— FCC SPECTRUM POLICY TASK FORCE

44


tion for portable consumer electronic devicessuch as PDAs. Contrary to the recommendationof the FCC’s Spectrum Policy Task Force, newunlicensed allocations should not be limited tospectrum above 50 GHz. However, as muchspectrum as possible at such extremely high frequencies should be made available for unlicensed use with minimal restrictions.

z SHARED UNLICENSED UNDERLAY

The FCC should advance its proceeding on unli-censed sharing of broadcast and other licensedspectrum bands. It should also make good on itscommitment to review its rules governing ultra-wideband and low-power FM radio, both of whichare subject to significant restrictions. Interferencewith licensed services is a legitimate concern, buthypothetical fears unsupported by the evidenceshould not trump innovation and competition.

The FCC should work with companies andstandards bodies in the private sector to identifyother technical means for unlicensed sharing.The interference temperature concept outlinedin the FCC Spectrum Policy Task Force reportis promising but undeveloped. Identifying moreclearly the boundaries between high-powerlicensed and low-power unlicensed services inthe same bands would benefit both types of sys-tems. However, the boundaries should not be solow as to preclude viable unlicensed devices.

z OPPORTUNISTIC SHARING

The FCC should allow for spectrum sharingalong the other dimensions of the electrospacemodel. Depending on the nature of the licenseduse, the government can make under-utilizedfrequencies available for non-interfering sharedaccess by time, geographic area, direction of sig-nals, or other variables. By establishing guide-lines for cognitive radios that can sense andrespond to the local spectral environment, thegovernment could in effect create “virtual white-space” within spectrum that is already licensed.The Defense Advanced Research Agency’s(DARPA) XG sharing technology provides onemodel for doing this. The XG moniker standsfor “neXt Generation.”

The FCC should pursue its proposal in theSpectrum Policy Task Force report to examinestandards for receivers. Existing rules focusentirely on transmitters, but as discussed in thispaper, interference is a phenomenon that mani-fests itself on the receiving end. Poorly definedspectrum licenses create incentives to build thecheapest, least robust receivers allowed. A poor-quality receiver may not affect the service forwhich it is licensed, but it will create scarcity bylimiting the space for opportunistic and sharedunlicensed devices around it.

z EXPERIMENTATION AND RESEARCH

Because spectrum has long been highly regu-lated and controlled by a small group oflicensees, opportunities for experimentation havebeen limited. WiFi happened almost by accident,

Radio Revolution

because the 2.4 GHz “junk band” spectrum wasso congested with devices such as microwaveovens that it was considered unusable forlicensed systems. Because the band was unli-censed and open to public access, engineers andcompanies interested in wireless local area networking could use it for their commerciallyand technically unproven technologies.

The government should ensure that its rulesfor unlicensed bands allow for the broadest pos-sible experimentation. It should also ensure thatit provides ample opportunities for experimentalauthorizations that go beyond the existing rules,so that innovators can develop novel techniques.A large block of high frequency spectrum (above50 GHz) that is currently unassigned could bedesignated as an open space for completelyunregulated unlicensed activity, as Tim Shepard

45

A Universal Communications Privilege for the “Supercommons”

Any set of prophylactic rules for wireless communi-cation, no matter how liberal, will prevent somepotential transmissions. The current block allocationsystem of government-licensed frequencies leaveshuge quantities of spectrum fallow, even thoughtoday’s devices could exploit the unused capacitywith no harm to existing systems. Even with unli-censed allocations, underlays, and opportunisticsharing, some non-interfering transmissions will beprecluded inefficiently. The optimal arrangement ofwireless communications devices depends on loca-tion, time, pre-existing systems, the desired service,and technological capabilities. No set of rules deter-mined ahead of time can possibly take all these factors into account.

The very idea of spectrum as a scarce resourcedivided into frequencies is becoming less and lessuseful. As devices become smarter, the effective com-munications space outside designated frequencybands increases. This “supercommons” cannot be

exploited fully so long as only certain wireless systems are legally permitted to operate.

An alternate approach would be to establish abaseline universal privilege to communicate.48 Any-one could engage in any form of wireless transmis-sion they chose, so long as they did not impinge onother systems. Instead of prophylactic interferencerules enforced by the government, a liability regimecould be used to avoid and resolve disputes. Thosesuffering harm from other transmissions could sue fordamages in court, using principles drawn from thecommon law doctrine of torts. Various safe-harborand backstop mechanisms could be used to ensurethe system operated efficiently. Equipment vendorsand wireless device owners would have incentives touse intelligent mechanisms to increase capacity.

The supercommons regime could operate alongsideexisting licensed and unlicensed frequencies. It wouldexpand communications capacity without threateningcurrent allocations. Access to the airwaves wouldbecome a distributed, real-time decision rather thanthe result of a government pre-approval process. Inno-vation and investment would increase dramatically.

and Yochai Benkler have proposed. This wouldallow for real-world exploration of ideas thatmay later migrate to lower frequencies wherepropagation characteristics are better.

The government should also continue tosupport and publicize research on groundbreak-ing wireless technology. Among the key ques-tions for exploration are: whether technical

“etiquettes” are useful to facilitate efficientspectrum sharing among unlicensed devices,and if so, what those etiquettes should be; whatare the fundamental information theoretic limits of radio communications; and how dolarge numbers of unlicensed devices using dif-ferent transmission mechanisms perform in the real world.

46


Conclusion

A powerful lesson from the history ofcommunications and computing is that afew simple trends have extraordinaryeffects over time. The shift from analogto digital networks, for example, is revo-lutionizing all forms of communicationand media. That transition has beengoing on for many years, and hasn’t yetfinished. Similarly, growing intelligenceof computing devices at the edges ofnetworks exerts a powerful force that ismagnified over time.

The paradoxical fact is that wirelesscommunications are both more mun-dane and more remarkable than we cur-rently believe. The mundane aspect isthat spectrum isn’t somehow special andremoved from the forces affecting otherindustries touched by the relentlessimprovements in computing and datanetworking. Spectrum isn’t a domainwhere normal market and technologicalforces go out the window, replaced byiron scarcities. In fact, it’s not a place ora thing at all. It’s a mental construct weuse to aid understanding. We are rapidly

reaching the point where that mentalconstruct does more harm than good.

The real question now is not whetherbut when. There is absolutely no ques-tion that wireless devices will continueto become more powerful, and thatenterprising technologists will find newways to multiplex and interweave radiosignals. Whatever obstacles regulatorsor incumbents throw up will ultimatelybe routed around. The only immutablebarriers are physical, and though theyundoubtedly exist, we are nowhere nearreaching them.

Through errors of both omission andcommission, however, governments candelay and weaken the revolutionarychanges that could bring into being thescenarios described in the previous sec-tion. If no more unlicensed spectrum ismade available through dedicated open-access bands, low-power underlay, andopportunistic sharing, it will take thatmuch longer to overcome the scarcitiesthat define the market today. Such delayswould have economic and social costs.

47

The last time a new networking and communi-cations paradigm took hold, it was called the Inter-net. There too, it was possible to discern signs andpossibilities years before the full commercial real-ization of the network. The U.S. government,through both affirmative steps (such as protectingnascent data services from domination by tele-phone companies) and conscious rejection of

unnecessary regulation, laid the groundwork forthe Internet to emerge as a powerful business andsocial force whose impacts are still being feltthroughout the economy.

The next Internet is before us. It is an Internet ofthe air, in some ways even more powerful than theInternet of wires. If we put aside our preconcep-tions and outmoded notions, we can make it real.

48

Conclusion

Baran, Paul, “Visions of the [sic] 21st CenturyCommunications: Is the Shortage of Radio Spectrum forBroadband Networks of the Future a Self Made Problem?”Keynote Talk, 8th Annual Conference on Next GenerationNetworks (November 9, 1994). Available at http://www.eff.org/GII_NII/Wireless_cellular_radio/false_scarcity_baran_cngn94.transcript (challenging the traditional conception ofspectrum scarcity)

Black, Jane, “The Magic of Wi-Fi,” BusinessWeek Online, March18, 2003.

Benkler, Yochai, “Overcoming Agoraphobia: Building theCommons of the Digitally Networked Environment,” 11 Harvard Journal of Law and Technology 287 (1998).(describing the benefits of a spectrum commons)

Benkler, Yochai, “Some Economics of Wireless Communications,”16 Harvard Journal of Law and Technology 25 (2002). (analyz-ing the economics of wireless systems under property rightsand commons)

Carter, Kenneth R., Ahmed Lahjouji, and Neil McNeal,“Unlicensed & Unshackled: A Joint OSP-OET WhitePaper on Unlicensed Devices and their Regulatory Issues,”Federal Communications Commission, Office of StrategicPolicy, Working Paper Series No. 39 (May 2003). Availableat http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-234741A1.pdf (surveying unlicensed wireless systemsand the FCC’s policy towards them)

Crawford, Sabrina, “Wireless Wheels: Police Now UsingInternet to Solve Local Crime Cases,” The San FranciscoExaminer, September 18, 2003.

Clarke, Arthur C., Profiles of the Future: An Inquiry into theLimits of the Possible, (London: Victor Gollancz, 1962).

The Economist, “Watch this Airspace,” The Economist,Technology Quarterly, June 22, 2002. (describing four dis-ruptive technologies that “could shake up the wirelessworld”)

Faulhaber, Gerald R., and David Farber, “Spectrum Management:Property Rights, Markets, and the Commons,” AEI-Brookings Joint Center, Working Paper 02-12 (December2002). Available at http://rider.wharton.upenn.edu/~faulhabe/SPECTRUM_MANAGEMENTv51.pdf (argu-ing that commons and property rights can coexist through“non-interfering easements”)

Federal Communications Commission, Spectrum Policy TaskForce Report, ET Docket No. 02-135 (November 15, 2002).Available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-228542A1.pdf (proposing an overhaulof spectrum regulation)

Radio Revolution

Federal Communications Commission, “First Report andOrder: Revision of Part 15 of the Commission’s RulesRegarding Ultra-Wideband Transmission Systems,” ETDocket 98-153 (April 22, 2002). (providing rules for ultra-wideband operation)

Fleishman, Glenn, “Take the Mesh-Networking Route: MeshNetworks Offer an Agile, Cost-Effective Alternative,”InfoWorld, March 7, 2003. Available at http://www.infoworld.com/article/03/03/07/10mesh_1.html (describing mesh net-working)

Gartner Group, Wireless LAN Equipment: Worldwide, 2001-2007,[report], Gartner Group, (January 2003).

Gelsinger, Patrick, “Wireless Fidelity (Wi-Fi): Leaping Acrossthe Digital Divide,” Keynote Address to the UnitedNations Information and Communication TechnologiesTask Force (June 26, 2003). Available at http://www.w2i.org/pages/wificonf0603/speaker_presentations/W2i_Gelsinger_Presentation.pdf

Gilder, George, “The New Rules of Wireless,” Forbes ASAP,March 29, 1993. (explaining the possibility for spectrumsharing without exclusive rights)

Gilder, George, “Auctioning the Airwaves,” Forbes ASAP, April 11, 1994.

Hazlett, Thomas W., “The Wireless Craze, the UnlimitedBandwidth Myth, the Spectrum Auction Faux Pas, and thePunchline to Ronald Coase’s ‘Big Joke’: An Essay onAirwave Allocation Policy,” 14 Harvard Journal of Law andTechnology 335 (2001). (attacking the commons position andadvocating exclusive property rights in spectrum)

In-Stat/MDR, It’s Cheap and It Works: Wi-Fi Brings WirelessNetworking to the Masses, [report], In-Stat/MDR,(December 2002).

Johnston, James H., and J.H. Snider, “Breaking the Chains:Unlicensed Spectrum as a Last-Mile Broadband Solution,”New America Foundation, Spectrum Series Working Paper#7 (June 2003). Available at http://www.newamerica.net/index.cfm?pg=article&pubID=1250

Keizer, Gregg, “Wireless LANs Set For Growth Spurt,”Information Week, March 27, 2003.

Konston, Kalle, “In Pursuit of a Wireless Device Bill of Rights,”Presentation to the September 18, 2002 Meeting of theFCC’s Technical Advisory Council (TAC). Available athttp://www.fcc.gov/oet/tac/Kalle_Kontson_9.18.02_bill_of_Rights_Final.ppt

49

Bibliography

Lessig, Lawrence, The Future of Ideas: The Fate of the Commons ina Connected World, (New York: Random House, 2001).(describing the benefits of an open wireless commons forinnovation)

Margie, Paul, “Efficiency, Predictability, and the Need for an Improved Interference Standard at the FCC,” PaperPresented at the 31st Annual Telecommunications Policy Research Conference (TPRC), Arlington, VA(September 2003).

Matheson, Robert, “The Electrospace Model as a Tool forSpectrum Management,” Paper Presented at the 6th

Annual International Symposium on Advanced RadioTechnologies (ISART), Boulder, CO (March 2003).

New America Foundation, et al., “Additional Spectrum forUnlicensed Devices Below 900 MHz and in the 3 GHzBand,” Comments to the Federal CommunicationsCommission Notice of Inquiry, ET Docket No. 02-380(April 17, 2003). Available at http://www.newamerica.net/Download_Docs/pdfs/Pub_File_1212_1.pdf

New America Foundation, et al., “Amendment of theCommission’s Rules to Facilitate the Provision of Fixed and Mobile Broadband Access, Educational and OtherAdvanced Services in the 2150-2162 and 2500-2690 MHz Bands,” Comments to the Federal CommunicationsCommission Notice of Proposed Rulemaking, WT Docket No. 03-66 (September 10, 2003). Available athttp://www.newamerica.net/Download_Docs/pdfs/Pub_File_1350_1.pdf

New America Foundation, and Shared Spectrum Company,“Dupont Circle Spectrum Utilization During Peak Hours,”[report], New America Foundation (June 2003). Available athttp://www.newamerica.net/Download_Docs/pdfs/Doc_File_183_1.pdf

Noam, Eli, “Spectrum Auctions: Yesterday’s Heresy, Today’sOrthodoxy, Tomorrow’s Anachronism: Taking the NextStep to Open Spectrum Access,” 41 Journal of Law andEconomics 765 (1998). (proposing an “open access” modelfor wireless systems)

Reed, David, “Comments for FCC Spectrum Policy Task Force on Spectrum Policy,” Comments to the FederalCommunications Commission Spectrum Policy Task ForcePublic Notice, ET Docket No. 02-135 (July 8, 2002).(describing the technical changes making a spectrum com-mons viable)

Shepard, Timothy, J., “Spectrum Policy Task Force SeeksPublic Comment on Issues Related to the Commission’sSpectrum Policies,” Comments to the FederalCommunications Commission Spectrum Policy Task ForcePublic Notice, ET Docket No. 02-135 (July 8, 2002).

Singer, Michael, “Analysts: Wi-Fi a ‘Positive Disruption’,”802.11 Planet, April 3, 2003.

Snider, J.H., and Max Vilimpoc, “Reclaiming the ‘VastWasteland’: Unlicensed Sharing of Broadcast Spectrum,”New America Foundation, Spectrum Series Issue Brief #12(July 2003). Available at http://www.newamerica.net/index.cfm?pg=article&pubID=1286 (proposing shared useof broadcast spectrum)

Stone, Amey, “Wi-Fi: It’s Fast, It’s Here — and It Works,”Business Week Online, April 3, 2002. (overview of the WiFimarket and its implications)

Weinberger, David, “The Myth of Interference,” Salon, March12, 2003. Available at http://www.salon.com/tech/feature/2003/03/12/spectrum/print.html (explaining David Reed’s ideas)

Werbach, Kevin, “A Layered Model for Internet Policy,” 1Colorado Journal on Telecommunications and High TechnologyLaw 37 (2002).

Werbach, Kevin, “Open Spectrum: The Paradise of theCommons,” Release 1.0, November 2001. (summarizing thecommons position and its implications for the communica-tions industry)

Werbach, Kevin, “Open Spectrum: The New WirelessParadigm,” New America Foundation, Spectrum SeriesWorking Paper # 6 (October 2002). Available athttp://www.newamerica.net/index.cfm?pg=article&pubID=1001 (explaining the benefits of the spectrum commonsapproach)

Werbach, Kevin, “Supercommons: Toward a Unified Theory of Wireless Communication,” 82 Texas Law Review(forthcoming March 2004). Draft available athttp://werbach.com/research/supercommons.pdf

Woolley, Scott, “Dead Air,” Forbes, November 25, 2002.(describing efforts to reform spectrum policy based ontechnologies that allow spectrum sharing)

Community WiFi Directorieshttp://www.seattlewireless.net/index.cgi/SimilarProjectLinkshttp://www.personaltelco.net/index.cgi/WirelessCommunitieshttp://www.afcn.org/http://freenetworks.org/moin/index.cgi/http://www.communitywireless.org/

50

Bibliography

1 Albert Einstein reportedly offered an entertaining descriptionof radio with a similar “negative” character: “You see, wiretelegraph is a kind of a very, very long cat. You pull his tailin New York and his head is meowing in Los Angeles. Doyou understand this? And radio operates exactly the sameway: you send signals here, they receive them there. Theonly difference is that there is no cat.”

2 Arthur C. Clarke, Profiles of the Future: An Inquiry into the Lim-its of the Possible (London: Victor Gollancz, 1962).

3 The word paradigm, since its popularization in ThomasKuhn’s The Structure of Scientific Revolutions, has becomedull from over-use. A paradigm shift isn’t merely a newtechnology or a new business opportunity. It’s a change incognitive framework. Our paradigms are the unconsciousscaffolding we use to associate concepts and information.Most of the time they can safely operate in the background,but occasionally they break down. True paradigm shifts arerare, but this rarity increases their significance.

4 This analogy was first developed by Tim Shepard in his com-ments to FCC’s Spectrum Policy Task Force. See TimothyJ. Shepard, “Spectrum Policy Task Force Seeks PublicComment on Issues Related to the Commission’s SpectrumPolicies,” Comments to the FCC’s Public Notice, ETDocket No. 02-135 (July 8, 2002).

5 Capacity may not always be expressed in raw throughput suchas bits per second. Some services have particular require-ments. For example, live voice communication needs lowlatency (delay), while high-quality video broadcastingrequires limited packet loss. An environment that allowsthese uses may be more valuable than one that does not,even if total carrying capacity is lower.

6 For a list of some other multiplexing mechanisms, see RobertMatheson, “The Electrospace Model as a Tool for Spec-trum Management,” Paper presented at ISART 2003 con-ference, Boulder, CO, March 2003.

7 This approach is known as Time Division Multiple Access(TDMA), and is the basis of the popular GSM standard forsecond-generation wireless phone systems.

8 See New America Foundation and Shared Spectrum Com-pany, “Dupont Circle Spectrum Utilization During PeakHours,” (July 2003) (finding a significant amount of spec-trum is unused, even in dense urban areas during peakhours).

9 Transitioning from analog to digital is another means ofexpanding capacity by changing the system design. Digital“2G” mobile phone networks support more simultaneouscalls than analog “1G” systems, and they can support addi-tional features such as data transmission and messaging. Yetthe 2G systems actually use less spectrum.

10 The layered model also has important implications for regu-lation. See Kevin Werbach, “A Layered Model for InternetPolicy,” 1 Colorado Journal on Telecommunications and HighTechnology Law 37 (2002).

Radio Revolution

11 The seven layers, from bottom to top, are: physical, datalink, network, transport, session, presentation, application.The OSI model is broadly used as a conceptual tool, butnot always as a formal part of network design. In the1980s, efforts to formally require adherence to the OSIstandard failed.

12 Separate from the technological critique developed here,there is a long and distinguished history of economic argu-ments against the current structure of spectrum regulation,generally advocating private property rights as an alterna-tive to government licensing. However, the spectrum-shar-ing techniques the new dynamic wireless paradigm makespossible remove the need for exclusive rights to minimizeinterference. See Yochai Benkler, “Overcoming Agorapho-bia: Building the Commons of the Digitally NetworkedEnvironment,” 11 Harvard Journal of Law and Technology287 (1998).

13 The FCC has oversight over private spectrum, and NTIA isthe lead agency for government spectrum. Several otheragencies, including the Defense and State Departments,play significant roles in spectrum policy.

14 The process of deciding which frequencies shall be availablefor which uses is known as allocation. Determining who isentitled to use those frequencies is called assignment. Inmany cases, spectrum bands are assigned to more than oneuser, or to one “primary” user and one or more “second-ary” users.

15 See Ronald H. Coase, “The Federal Communications Com-mission,” 2 Journal of Law and Economics 1 (1959).

16 This became an issue in the FCC proceeding authorizingultra-wideband devices to underlay under existing licensedservices. Sprint PCS argued that, because it had spent bil-lions of dollars purchasing spectrum licenses and buildingout its mobile phone network, it had an expectation ofexclusive control over the frequencies. The FCC rejectedthis claim, holding that even for flexible PCS spectrum, thelicense was not absolute. A licensee may have a reasonableexpectation to operate free from harmful interference forthe limited term of the license, but the government maygive non-interfering users access to the same frequencies.See “FCC First Report and Order: Revision of Part 15 ofthe Commission’s Rules Regarding Ultra-Wideband Trans-mission Systems,” ET Docket No. 98-153 (April 22, 2002).

17 There are two different patterns to prevent interference innearby cities. There is no channel 3 in New York, but nochannel 2 in Philadelphia, to ensure the signals are receivedproperly.

18 In addition, because only a handful of big city markets havemore than 13 channels on the air, TV channels 30 to 69 (234MHz of prime spectrum) are in use in only about 10 percentof the nation’s TV market, on average. See J.H. Snider andMax Vilimpoc, “Reclaiming the ‘Vast Wasteland’: UnlicensedSharing of Broadcast Spectrum,” New America Foundation,Spectrum Series Issue Brief #12 (July 2003).

51

Endnotes

19 For an excellent discussion of the FCC’s evolving regulationof interference, see Paul Margie, “Efficiency, Predictability,and the Need for an Improved Interference Standard at theFCC,” Paper presented at TPRC 2003, Arlington, VA,September 2003.

20 Harmful interference is defined as “Interference whichendangers the functioning of a radio navigation service orof other safety services or seriously degrades, obstructs orrepeatedly interrupts a radio communication service operat-ing in accordance with the Radio Regulations.” 48 CFR§97.3(a)(23).

21 See Matheson, supra note 6.22 See Kalle Konston, “In Pursuit of a Wireless Device Bill of

Rights,” Presentation to the September 18, 2002 Meetingof the FCC’s Technical Advisory Council (TAC).

23 See “FCC First Report and Order: Revision of Part 15 of theCommission’s Rules Regarding Ultra-Wideband Transmis-sion Systems,” ET Docket No. 98-153 (April 22, 2002).UWB was previously used for non-communications appli-cations such as ground-penetrating radar, especially by themilitary.

24 Not to be confused with defunct digital subscriber lineprovider Northpoint Communications. Because itsapproach is so novel, Northpoint has had to endure amulti-year struggle to get its technology approved by the FCC.

25 In a pure mesh, every receiver is also a transmitter and arepeater. However, this is not a requirement. The Internet,for example, has many different kinds of routers and net-works, some of which are connected hierarchically with“edge” and “core” devices.

26 See Sabrina Crawford, “Wireless Wheels: Police Now UsingInternet to Solve Local Crime Cases,” The San FranciscoExaminer (September 18, 2003).

27 Any digital wireless system involves some measure of soft-ware. SDR involves using software to control most of thefunctions that traditionally were handled through RF hard-ware.

28 See George Gilder, “The New Rules of Wireless,” ForbesASAP (March 29, 1993). See also George Gilder, “Auction-ing the Airwaves,” Forbes ASAP (April 11, 1994).

29 See Kevin Werbach, “Open Spectrum: The New WirelessParadigm,” New America Foundation, Spectrum SeriesWorking Paper #6 (October 2002). See also Kevin Wer-bach, “Open Spectrum: The Paradise of the Commons,”Release 1.0, (November 2001).

30 See Yochai Benkler, “Some Economics of Wireless Commu-nications,” 16 Harvard Journal of Law and Technology 25(2002).

31 See Michael Singer, “Analysts: Wi-Fi a ‘Positive Disrup-tion,’” 802.11 Planet, (April 3, 2003).

32 See It’s Cheap and It Works: Wi-Fi Brings Wireless Networkingto the Masses, [report] In-Stat/MDR (December 2002). Seealso Wireless LAN Equipment: Worldwide, 2001-2007,[report] Gartner Group (January 2003).

33 See Patrick Gelsinger, “Wireless Fidelity (Wi-Fi): LeapingAcross the Digital Divide,” Keynote Address to the UnitedNations Information and Communication TechnologiesTask Force (June 26, 2003).

34 See supra note 6.35 See Kevin Werbach, “Supercommons: Toward a Unified

Theory of Wireless Communication,”82 Texas Law Review___ (forthcoming March 2004), draft available at http://wer-bach.com/research/supercommons.pdf

36 See New America Foundation et al., “Additional Spectrumfor Unlicensed Devices Below 900 MHz and in the 3 GHzBand,” Comments to the FCC Notice of Inquiry, ETDocket No. 02-380 (April 17, 2003).

37 See J.H. Snider and Max Vilimpoc, “Reclaiming the ‘VastWasteland’: Unlicensed Sharing of Broadcast Spectrum,”New America Foundation, Spectrum Series Issue Brief #12(July 2003).

38 WiFi stands for “wireless fidelity” only in this marketingsense. HiFi systems actually have higher sound fidelity;there is nothing special about the fidelity of WiFi systems.

39 See James H. Johnston and J.H. Snider, “Breaking theChains: Unlicensed Spectrum as a Last-Mile BroadbandSolution,” New America Foundation, Spectrum SeriesWorking Paper #7 (June 2003).

40 Id.41 See http://www.nodedb.org42 See Jane Black, “The Magic of Wi-Fi,” BusinessWeek Online,

(March 18, 2003).43 See Gregg Keizer, “Wireless LANs Set For Growth Spurt,”

Information Week, (March 27, 2003).44 See James H. Johnston and J.H. Snider, “Breaking the

Chains: Unlicensed Spectrum as a Last-Mile BroadbandSolution,” New America Foundation, Spectrum SeriesWorking Paper #7 (June 2003).

45 See Federal Communications Commission, Spectrum PolicyTask Force Report, ET Docket No. 02-135 (November 15,2002).

46 See New America Foundation et al., “Additional Spectrumfor Unlicensed Devices Below 900 MHz and in the 3GHzBand,” Comments to the FCC Notice of Inquiry, ETDocket No. 02-380 (April 17, 2003).

47 See New America Foundation et al., “Amendment of theCommission’s Rules to Facilitate the Provision of Fixed andMobile Broadband Access, Educational and OtherAdvanced Services in the 2150-2162 and 2500-2690 MHzBands,” Comments to the FCC Notice of Proposed Rule-making, WT Docket No. 03-66 (September 10, 2003).

48 See Werbach, supra note 35.

52

Endnotes

radio revolution: the coming age of unlicensed wirelesswerbach.com/docs/radiorevolution.pdf ·...

Documents