ITU-Trends in Telecommunication Reform 2006

20

06

I n t e r n a t i o n a l T e l e c o m m u n i c a t i o n U n i o n

I n te rnat iona lTe lecommun icat ionUn ion

reformTRENDS INTELECOMMUNICATION

2006R E G U L A T I N G

I N T H EB R O A D B A N D

W O R L D

Printed in SwitzerlandGeneva, 2006

ISBN 92-61-11431-8

*28388*

TR

EN

DS

IN

TE

LE

CO

MM

UN

IC

AT

IO

N R

EF

OR

M

20

06

© ITU 2006

International Telecommunication UnionPlace des Nations

CH-1211 Geneva, Switzerland

First printing 2006

All rights reserved. No par t of this publication may be reproduced, by any means whatsoever, without the prior writtenpermission of the International Telecommunication Union.

Denominations and classifications employed in this publication do not imply any opinion on the par t of the InternationalTelecommunication Union concerning the legal or other status of any territory or any endorsement or acceptance of any boundary.

Where the designation “country” appears in this publication, it covers countries and territories.

ISBN 92-61-11431-8

ALSO AVAILABLE FROM ITU

PUBLICATIONS

Trends in Telecommunication Reform: Licensing in an Era of Convergence, 2004/05 (6th Edition) ..............................95 CHFTrends in Telecommunication Reform: Promoting Universal Access to ICTs, 2003 (5th Edition) ..................................90 CHFTrends in Telecommunication Reform: Effective Regulation, 2002 (4th Edition) ...........................................................90 CHFTrends in Telecommunication Reform: Interconnection Regulation, 2000-2001 (3rd Edition) .......................................90 CHFTrends in Telecommunication Reform: Convergence and Regulation, 1999 (2nd Edition) .............................................75 CHFGeneral Trends in Telecommunication Reform 1998: World Volume I ...........................................................................75 CHFGeneral Trends in Telecommunication Reform 1998: Africa Volume II ..........................................................................65 CHFGeneral Trends in Telecommunication Reform 1998: Americas Volume III ...................................................................55 CHFGeneral Trends in Telecommunication Reform 1998: Arab States Volume IV ...............................................................45 CHFGeneral Trends in Telecommunication Reform 1998: Asia Pacific Volume V ................................................................60 CHFGeneral Trends in Telecommunication Reform 1998: Europe Volume VI .......................................................................72 CHFCollection of five Regional reports (Volumes II-VI) ......................................................................................................297 CHFCollection of Regional and World reports (Volumes I-VI) .............................................................................................372 CHFThe Arab Book: Telecommunication Policies for the Arab Region, 2002 ......................................................................50 CHFThe Blue Book: Telecommunication Policies for the Americas, 2000 (2nd Edition) ......................................................50 CHF

Please contact the ITU Sales Service:Tel.: +41 22 730 5111Fax: +41 22 730 5194E-mail: [email protected]: www.itu.int

Note: Discounts are available for all ITU Member States, Sector Members and Least Developed Countries.





W O R L D


Inte rnat iona lTe lecommun icat ionUn ion

v1.0

1

ii

This report was prepared by a team led by Doreen Bogdan-Martin and Susan Schorr. The team included: John Alden, William Bratton, Tracy Cohen, Yang-Soon Lee, Olli Mattila, Michael Best, John Muleta, John Palfrey, Bjorn Pehrson, Audrey Selian, Wu Wei Shi, Russell Southwood and Nancy Sundberg. We specially thank Tania Bezago, our GREX Advisor, who provided additional research and Esperanza Magpantay, BDT, who contributed ICT statistics.

The cover was designed by Stéphane Rollet.

The report was edited by John Alden. The regulatory tables were prepared by Kevin Munn and Nancy Sundberg. The ITU Publications Composition Department was responsible for production of the report.

The report has benefited from the comments and advice of ITU staff including: Tim Kelly, Colin Langtry, Fabio Leite, Ricardo Passerini and Robert Shaw.

We would especially like to thank for their comments: Martin Cave (Warwick Business School, University of Warwick); Claudia Sarrocco (Organisation for Economic Co-operation and Development); Suresh Ramasubramanian (Outblaze); John Haydon (Australian Communications and Media Authority); Tom Dale (Australian Department of Communications, IT and the Arts); Ewan Sutherland (International Telecommunication Users Group); Derek Bambauer, David Abrams, Jonathan Zittrain (the Berkman Center); and César David Moliné (Indotel, Dominican Republic).

These contributions, together with the support from ICT ministries and regulators, and others who have provided data and background ma teri al, are gratefully ac knowl edged. Without their support, a report of this nature would be impossible.

iii

FOREWORD

Hamadoun I. TouréDirector

Telecommunication Development Bureau7 March 2006

We are proud to present the seventh edition of Trends in Telecommunication Reform, an integral part of our dialogue with the world’s information and com-munications technology (ICT) policy-makers and regulators. This 7th edition has been released at a time of remarkable transformation of the informa-tion and communication technology (ICT) sector, fuelled by a combination of technological, market, policy and regulatory developments. These changes include unparalleled numbers of voice telephone subscribers, the rise of IP-enabled networks and Voice over IP (VoIP) services, initial-yet promising-deployment of fixed line broadband and broadband wireless access (BWA) services and intelligent radio devices. At the same time that developed countries are busy planning for the deployment of next gen-eration networks and visualize a world of ubiquitous networks, most developing countries have expanded their continuing quest to provide universal access to basic voice services to include universal access to broadband internet services. Are developing coun-tries making any progress in this quest? How can regulators harness the potential of new technologies and innovative business models to foster ICT sector development?

In 1984, the Maitland Commission’s ‘Missing Link’ report identified the challenge of bringing ba-sic tele com mu ni cation services within easy reach of all the world’s people by the early part of the 21st century. The Missing Link report has been the touchstone for the work of the Telecommunication Development Bureau (BDT) of the International Telecommunication Union (ITU) since its incep-tion. In 2002, mobile cellular communications were heralded as the answer to the missing link. Second generation mobile services have been rolled out to more people in the developing world than the Maitland Commission could ever have imagined possible-although the challenge of universal access to basic services remains. The success of mobile communications is linked to policy and regulatory reforms as well as innovative business models and new technological applications, such as short mes-sage services (SMS).

Twenty years later, the World Summit on the In-formation Society (WSIS) set even more ambitious targets, to extend the internet to all the world’s vil-lages by 2015 as the foundation for building the In-

formation Society. The WSIS also identified the vital role the regulatory framework plays in enabling the Information Society. The global community of na-tional communications regulators met on the eve of the second phase of WSIS to develop a new vision of a regulatory framework to promote the deploy-ment of broadband internet communication services worldwide, in developing and developed countries alike. At that same meeting, known as the ITU Glo-bal Symposium for Regulators (GSR), regulators recognized that full participation in the Information Society requires access to broadband internet serv-ices.

Low cost technologies exist today that can pro-mote broadband access and enable developing coun-tries to “leapfrog” over older technologies to advance into the broadband future rapidly. Many businesses stand ready to start providing broadband services in developing countries-mirroring the keen interest second-generation mobile service providers dem-onstrated in developing countries throughout this decade. Although broadband technologies and busi-ness models hold great promise, the pace of broad-band take-up hinges on the regulatory framework. In many countries, today’s broadband ‘missing link’ is the regulatory framework. Regulators have an unprecedented opportunity to speed the uptake of broadband to enable the Information Society. To-day’s broadband challenge requires new thinking, and an end to business as usual. This publication is designed to enable regulators and policy-makers to meet this challenge.

The report has been prepared by the ITU’s Tele-communication Development Bureau. The authors have benefited from comments and input from a range of people, inside and outside of ITU. The views expressed in the report, however, are those of the authors and do not necessarily reflect the opin-ions of ITU or its members.

iv

1 MARKET AND REGULATORY TRENDS IN THE ICT SECTOR ...............................................................................11.1 What will it take to bring broadband to the masses? ............... 1

1.2 ICT infra struc ture deployment ................................................ 2

1.3 Market Trends in Privatizations, Investments and Services ....... 6

1.4 Competition .......................................................................... 9

1.5 National Regulatory Authorities, Who Rules? ........................ 12

1.6 Regulatory Challenges of VoIP ............................................ 14

1.7 Spectrum Management ....................................................... 16

1.8 Dealing with Spam ............................................................. 16

1.9 Conclusion ......................................................................... 18

2 WHAT IS “BROADBAND”? ............................................ 212.1 What Do We Mean When We Say “Broadband”? ................... 21

2.2 A Short Broadband Taxonomy ............................................. 22

2.3 Conclusion ......................................................................... 26

3 UNDERSTANDING BROADBAND TECHNOLOGIES ..... 273.1 How Broadband Networks Are Designed ............................... 27

3.2 A Wire-line Broadband Roadmap ......................................... 29

3.3 Broadband Wireless Access (BWA) Networks ....................... 33

3.4 A Decision Framework for Broadband ................................... 41

3.5 Power requirements for broadband ....................................... 44

3.6 Conclusion ......................................................................... 49

4 THE ROLE OF THE REGULATOR IN BROADBAND

DEVELOPMENT ............................................................... 514.1 Introduction ........................................................................ 51

4.2 The Importance of Broadband in Developing Countries ......... 54

4.3 Key Issues in Promoting Broadband in Developing Countries 54

4.4 Providing Incentives for Network Investment ......................... 56

4.5 Broadband Licensing .......................................................... 58

4.6 Alternative Approaches to Broadband Deployment ............... 60

4.7 Competition and Industry Regulation.................................... 63

4.8 Increasing Broadband Awareness ......................................... 69

4.9 Conclusion ......................................................................... 71

CONTENTS

5 BROADBAND SPECTRUM MANAGEMENT ................ 755.1 Introduction ........................................................................ 75

5.2 The Economics of Broadband Wireless Access ..................... 76

5.3 The Technology Revolution ................................................. 78

5.4 Adapting Spectrum Regulatory Models for BWA ................... 78

5.5 Defining Best Practices ....................................................... 83

5.6 Case Study: BWA Spectrum Allocation in Mauritius.............. 87

5.7 Conclusion ......................................................................... 90

6 VOIP AND REGULATION ................................................ 916.1 VoIP: Regulatory Evolution or Revolution? ............................ 91

6.2 The Pace of VoIP Market Development ................................. 95

6.3 Grappling with Change: Regulators’ Responses to VoIP ....... 96

6.4 Crafting New Regulatory Approaches to VoIP ...................... 100

6.5 End User and Consumer Issues ......................................... 105

6.6 Conclusion ....................................................................... 108

7 STEMMING THE INTERNATIONAL TIDE OF SPAM ..1117.1 The Spam Problem ............................................................ 111

7.2 An Outline of a Model Law ................................................. 115

7.3 Codes of Conduct ............................................................. 119

7.4 Education and Awareness ................................................. 122

7.5 Conclusion ....................................................................... 123

8 MAKING BROADBAND WORK FOR ALL ....................1278.1 What Is Broadband? ........................................................... 127

8.2 Why Should I Care about Broadband?................................. 127

8.3 How Can I Get Broadband? ................................................ 128

8.4 What Can Regulators Do? .................................................. 128

REGULATORY TABLES ..........................................................131

Table 1 – Countries with a separate Regulatory Authority .............133

Table 3 – Status of the main fixed-line operators .........................169

Table 4 – Level of competition ....................................................203

PAGE

PAGE

v

PAGE

PAGE

LIST OF TABLES .....................................................................229

LIST OF BOXES ......................................................................230

GLOSSARY OF TERMS ..........................................................215

LIST OF FIGURES ..................................................................228

1

Trends in Telecommunication Reform 2006

CHAPTER 1

1 MARKET AND REGULATORY TRENDS IN THE ICT SECTOR

regulators can aim to make local communities and non-govern-mental organizations aware of the technologies and broadband provisioning opportunities they could seize, and also coordinate with other government and public institutions, such as univer-sities, to drive demand for broadband-enabled health, educa-tion and government services. At the same time, regulators will strive to revise outdated regulatory frameworks designed for an earlier era. The new regulatory framework could be described as a ‘less means more, old meets new’ approach. Less regula-tory intervention means more business opportunities. Time-tested regulatory principles such as transparency and open competition will be applied to new technologies and the new regulatory issues they raise. And the promotion of wireless broadband technologies will require flexible and innovative spectrum management practices.

Of course, broadband also poses new challenges. The rapid rise of voice over IP (VoIP) – hastened by the spread of broadband – is turning the old telecom business model on its head. VoIP service providers have introduced a new business model, providing voice services for free or bundled as a part of a triple play package. What effect will this have on the busi-ness plans of traditional telecom operators? And what do these developments mean for the current regulatory framework designed for the old business model?

VoIP is not the only broadband-related challenge facing the ICT sector. There is also the scourge of spam, which clogs email inboxes and leads to internet fraud as well as poses net-work security problems such as the spread of internet viruses and worms.

This year’s Trends in Telecommunication Reform contains eight chapters addressing each of the broadband-related chal-lenges and opportunities to enable regulators to harness the potential of broadband to build a safe and secure Information Society for all:• This chapter provides an ICT market and regulatory over-

view to set the stage for the following chapters;• Chapter Two defines what policy-makers and technolo-

gists mean by the term “broadband;”• Chapter Three explores broadband technology at a more

technical level and looks ahead, providing a roadmap for

1.1 What will it take to bring broadband to the masses? Bringing broadband to the masses is one of the major

challenges facing the global ICT community. Addressing this challenge requires new thinking, and an end to business as usual. The 7th edition of Trends in Telecommunication Reform is designed to enable regulators and policy-makers to meet this challenge.

What does an end to business as usual mean? Service pro-viders are now offering the triple play of voice, internet and broadcast as convergence moves from a dream to a part of every day life, starting in the developed world and spreading to the developing world. Music lovers download songs from the internet onto their MP3 players, working parents order their groceries online and come home at night to watch their favorite TV shows broadcast over the internet, while interna-tional travelers book their flights and reserve their hotel rooms on the web. Broadband is about more than travel and enter-tainment. International organizations like ITU use a broad-band-enabled e-learning platform to provide capacity building services to its membership and work with governments to deploy broadband-enabled e-government services.

Wireless broadband technologies offer the prospect of faster rollout of services, as well as portability and mobility. Many broadband technologies can also be deployed incremen-tally, as demand develops, rather than requiring expensive net-work-wide upgrades. This means that a full range of players, large and small, private and public, can harness the power of these technological developments to become ICT service pro-viders and close the broadband divide that exists between devel-oping and developed countries, and between rural and urban areas within countries. For example, already two developing countries, Mauritius and TFYR Macedonia, have announced plans to become entirely wireless broadband nations.

All of these market and technological developments are exerting pressure on the current regulatory framework. How will regulation change? Broadband regulation means a new vision of reduced regulatory burdens, innovative incentives, and coordinated efforts by all links in the broadband value chain to unleash commercial deployment opportunities. Regu-lations can be carefully tailored to open the door to both large and small-scale broadband providers. Broadband-promoting

In many countries, today’s broadband ‘missing link’ is the regulatory framework. Regulators have before them an unprecedented opportunity to speed the uptake of broadband to enable the Information Society.

CHAPTER 1


2

regulators to plot the rapid technological changes that are resulting in new opportunities and services;

• Chapter Four examines the role of regulators in promot-ing broadband;

• Chapter Five, recognizing the key role that wireless tech-nologies are likely to play in the promotion of broadband in developing countries, examines spectrum management practices to promote broadband;

• Chapter Six looks at current regulatory treatment of Voice over IP (VoIP) and the regulatory road that lies ahead for VoIP;

• Chapter Seven addresses enforceable codes of conduct for ISPs as a new legal tool regulators could deploy in their fight against spam, and

• Chapter Eight offers a conclusion and a look ahead.

1.2 ICT infra struc ture deployment

The state of current ICT infra struc ture deployment is key to understanding the new technologies that can be used to promote broadband access in developing countries.1 Today, the majority of the world’s broadband subscribers are found in developed countries, which have upgraded existing fixed line telephone and cable TV infra struc ture to provide broadband services. Access to basic communications in the developing world has largely been achieved through mobile communica-tions. It is unlikely, therefore, that developing countries will follow the same migration path to broadband services as devel-oped countries. While developing countries will no doubt also deploy some fixed line broadband services, broadband wireless access is expected to play a key role for developing countries seeking to foster the Information Society.

As shown in Figure 1.1, the growth of mobile lines con-tinues to outpace the growth of fixed lines, and this is particu-larly evident in developing countries, a trend attributed to the introduction of prepaid mobile services, rapid and cheaper network deployment, a competitive environment as well as the fact that mobile services provide access to a range of new applications such as short and multimedia messaging services (SMS and MMS). By the end of 2004, the world counted some 1.8 billion mobile subscribers (including both second and third generation mobile subscribers), or 28 per cent of the world’s population. Some 58 per cent of these mobile subscribers were located in developing countries. Further information about 3G subscribers is provided in section 1.2.2.

The number of fixed line subscribers worldwide, had reached 1.2 billion lines, and a penetration rate of 19 per cent. As shown in Figure 1.2, both fixed line and mobile penetra-tion rates are lowest in Africa and Asia. Figures 1.2 also demon-strates that mobile teledensity rates outpace fixed teledensity in every single region.

By year end 2004, there were an estimated 840 million internet users in the world, representing 13.2 percent of the total population.

1.2.1 Global Fixed Line Broadband Growth

As previously noted, the total number of internet users (broadband and dial-up) around the world continues to increase, having reached about 840 million at the end of 2004. The total number of fixed-line broadband subscribers had reached nearly 160 million at the end of the 2004. Broadband internet subscribers represented approximately 2.5 percent of the world’s population, and 38 per cent of all internet subscrib-ers worldwide in 2004.

Figure 1.1: The Number of ICT Users Worldwide, 1994-2004Fixed-line and mobile subscribers and internet users

Source: ITU World Telecommunication Indicators Database.

CHAPTER 1 3


The vast majority of today’s broadband users are in the developed world. Globally, Asia, Europe and the Americas represent no less than 99 percent of all broadband subscribers, the majority of which are in the wealthier countries of North America, Western Europe and Asia (Figure 1.3). By contrast,

Africa is home to only a fraction of broadband subscribers, and many African countries have not yet launched high-speed Internet services. Although Africa has the fewest numbers of broadband subscribers globally, the number of subscribers increased some thirty times in the two-year period from 2002

Figure 1.2: Fixed vs. Mobile Teledensity by Region (per 100 subscribers)


Fixed telephone line penetration by region over the period of 1994-2004 Mobile cellular penetration, by region, 1999-2004

Figure 1.3: BroadbandDistribution of broadband subscribers by region, 2004


CHAPTER 1


4

Figure 1.4: Global Distribution of Internet and Broadband Subscribers, 2004

Source: ITU World Telecommunication Indicators Database, and OECD, the Organisation for Economic Cooperation and Development

Figure 1.5: Top 25 Broadband Subscribers, Non-OECD Countries (2004)


CHAPTER 1 5


to 2004, signally a healthy trend. A host of other developing countries from other regions are also showing signs of vigor-ous broadband growth.

These healthy trends are reflected in Figure 1.4, showing that approximately 25 per cent of all broadband subscribers are in non-OECD countries, fuelled largely by China, which has the second largest number of broadband subscribers in the world, after the United States. Figure1.5 shows the top 25 non-OECD countries measured by number of broadband subscrib-ers.

Indeed, attention should be focused on the considerable new broadband deployment activity in nations throughout the developing world, from the Arab States to Southeast Asia. For example, TE Data and ISP Nile Online have rolled out DSL internet lines. In Chile, carrier Telsur has initiated a broadband development project, which has garnered a total investment of USD 20 million in the last five years. Brazil has launched triple play services including broadband, as has the Indian operator MTNL.

Still, developing countries need to continue to grow their broadband subscribers to compete with today’s top broadband-enabled economies. Figure 1.6 provides a snapshot of the top ten countries in terms of broadband subscribers, and their growth over the past four years.

Of the total broadband subscriber base, 32 per cent of subscribers are using cable modems for access (51 million subscribers), 62 per cent are using DSL (98 million subscrib-ers), and 6 per cent (9.1 million subscribers) are using another technology (for example, satellite, fibre-to-the-home and Eth-ernet LANs). Figure 1.7 depicts the breakout of broadband access platforms by region. It is fair to assume that while there may be other ways of achieving “broadband status” (for exam-ple, a VSAT connection), most broadband subscribers today

are connected via DSL or cable modems. It is clear from the data that DSL features most prominently in all regions of the world except North America where cable modem technology remains (at least for a now) more dominant. Meanwhile, the majority of people – more than 60 per cent of all internet sub-scribers – still use dial-up connections.

1.2.2 Global Wireless Broadband Growth

IMT-2000 technologies, known popularly as 3G mobile, are also starting to sprout broadband subscribers. By January 2005, 56 of the world’s economies were offering commer-cial 3G services and the total number of reported subscribers accessing 3G technologies was 150 million, close to 60 per cent growth from the previous year,2 and just shy of the 160 mil-lion fixed-line broadband subscribers. Of these 150 million 3G subscribers, 100 million are located in just three countries, the United States (49.5 million), the Republic of Korea (27.5 million) and Japan (25.7 million). As described more fully in Chapters 2 and 3, the dominant 3G technologies deployed to date are W-CDMA and CDMA-1x. Figure 1.8 shows the top ten 3G mobile markets worldwide in 2005, broken down by these two standards.

As operators determine their migration preferences toward the achievement of 3G networks, one open question is whether the dominance of the GSM standard in 2G telephony will continue to prevail in its more advanced 3G form, as some industry groups predict. The breakdown of 2G standards use by subscribers is highlighted in Figure 1.9.

Many countries are also looking at other wireless broad-band technologies. Some countries are already extending broadband connectivity through such BWA technologies, and a couple of these have begun to do so on a nationwide basis. (See Box 1.1).

Figure 1.6: Subscriber Growth in the Top 10 Broadband Countries (2000-2004)


CHAPTER 1


6

1.2.3 The Poor Pay More for Less BroadbandAs previously shown, broadband penetration is higher in

high-income countries. And yet, subscribers in low-income countries pay more for less capacity. Given the fewer broad-band users per capita in low-income countries, how does the price and capacity of the service vary from country to country? The average fee paid by subscribers in low-income countries for broadband service is USD 291 per month, compared with a mean price of USD 18 a month in upper-income countries.

While the poor pay more for their broadband service, on average, they also receive less bandwidth. The average high-speed downlink capacity in upper-income countries is 3.8 Mbit/s, compared with an average for low-income coun-tries of 712 kbit/s.

In summary, while penetration is very small in low-income countries, broadband subscribers there pay consider-ably more money for inferior service. Perhaps further research and market studies could be undertaken to address whether there are objective factors – such as the need to recover higher network construction costs over a smaller initial subscriber base – to explain why customers in lower-income and middle-income countries are paying more for less.

1.3 Market Trends in Privatizations, Investments and ServicesPrivatization of state-owned incumbent operators often

affects the pace of regulatory reform, reducing or eliminating concerns regulators and policy-makers may have about protect-ing legacy operators from the pressure of new market entrants or new technologies that can provide lower cost and more innovative services. Where privatizations have not occurred, or where only partial privatizations have been implemented, there

is generally a greater tendency to balance the concerns of legacy operators (for revenues and market share) with those of end-users (for low cost services). This delicate balance often results in limitations on regulatory reforms. Privatization usually gives regulators and policy-makers greater freedom to focus on the interests of end users, and accelerate the reform process.

Although many governments now recognize the value of privatizing their fixed line operators, privatization activity has stalled over the past two years, with no new privatization of state-owned fixed line operators either in 2004 or 2005, as shown in Figure 1.10. A number of initial and further privati-zations, however, are on the horizon, including in Tunisia and Pakistan. Likewise, a number of companies are busy negoti-ating acquisitions or arranging financing to expand their net-work holdings in foreign markets. For example, one of China’s leading mobile operators has held discussions concerning its expansion into India; another mobile operator from Hong Kong, China is planning expansion into India, Indonesia, and Vietnam.

The ICT sector has also witnessed significant investment in infra struc ture equipment over the past year, signalling signs of recovery following the market collapse in the early years of the decade. According to some analysts, the total global market in 2004 for wireless infra struc ture was more than USD 39 bil-lion, a 12 per cent increase in operator spending from 2003 levels.3 Expenditures on W-CDMA equipment likely crossed the USD 10 billion threshold in 2005, largely based on spend-ing generated by operators in China, Germany, Japan, Spain, the United Kingdom and the United States.4 In fact, analysts believe that China alone generated about 16 per cent of W-CDMA-related spending in 2005.

The world was also witnessing the awesome growth of companies benefiting directly from the growth of the internet.

Figure 1.7: Broadband Platform Distribution (by Region) 2004


CHAPTER 1 7


Box 1.1: Broadband Wireless Nations

A small set of countries has announced plans to vie (unoffi cially, of course) to become the world’s “fi rst broadband wireless nation.” Mauritius and TFYR Macedonia both are deploying fi xed wireless broadband networks across the bulk of their countries, using technologies that track the emerging 802.16 WiMAX standards.

In Mauritius, a small African island nation of 1.2 million people, the wireless broadband network reportedly already covers 60 per cent of the island and 70 per cent of the population. By the end of 2005, Mauritius intends coverage to reach a full 90 per cent of the country.1

In TFYR Macedonia, a country of 2 million people, local network provider On.NET is deploying a broadband wireless network across the country using Motorola’s Canopy radio system. The project is a unique partnership between the donor community, the Macedonian Government, and the private sector. The Government of China has donated thousands of personal computers to be used in primary and secondary schools. Complementing that donation, The U.S. Agency for International Development (USAID) is providing broadband internet connectivity to 460 primary and secondary schools and 71 other sites through 2007. In October 2005, the project partners announced an important milestone: 95 per cent of the country’s population was within reach of a broadband wireless signal.

The project will guarantee a substantial countrywide customer base. This has created a business case for On.NET to make a signifi cant investment in a pervasive national wireless network. On.NET is free to sell capacity to additional corporate or consumer subscribers throughout the country. Furthermore, in metropolitan areas, On.NET is deploying a mesh-based network to offer pervasive hotspot connectivity in the country’s population centres.

1 This is being done using wireless technologies developed by Texas based Navini Networks.

Figure 1.8: Top 10 3G Mobile Markets Worldwide, 2005

Source: ITU, The Internet of Things.

The 2005 December holiday season, for example, saw many retailers celebrating record online sales volumes.

Traditional tele com mu ni cations and cable TV operators are also benefiting from the rise of the internet and conver-gence. The past year has seen considerable emphasis on the launch of triple play services combining video (television), voice (telephone) and broadband (internet access) offerings from the same provider. The triple play trend is most notable

in the Americas and the Asia-Pacific region, and also in some of the Arab States. Many tele com mu ni cation operators in Brazil, for example, are investing in triple play packages.5 The Uruguayan state-owned tele com mu ni cation company, Antel, planned to launch triple play trials in 2006,6 while Chilean pro-vider VTR-Metropolis invested in a plan to achieve nationwide triple play coverage by 2010, serving 2 million homes.7 Service launches were also set for June 2006 in Colombia.

CHAPTER 1


8

Figure 1.9: 2G Standards Usage, by subscribers, world and by region (2004-5)

Sources: GSM Association, CDMA Development Group, Mymobile.com News, NTT DoCoMo Corporate website, KDDI Corporate website, Vodafone Japan; Source: The Diffusion Group Report, “U.S. Mobile Markets: Analysis & Forecasts.”

Notes: TDMA subscriber numbers are included on the high end of analyst estimates.

Today’s mobile subscribers are served by cellular systems based on just a handful of major standards. The two dominant standards are GSM and CDMA, with GSM the most popular second-generation mobile stand-ard in the world. The GSM Association estimated a global user base of more than 1.5 billion as of the first quarter of 2005--or three-quarters of the 2G subscriber base. Approximately 31 million users are signing on each month, with growth driven chiefly by developing markets such as those in Africa, Latin America, Eastern Europe and Asia. GSM is also gaining popularity in the United States, reaching 55 million subscribers in 2005. CDMA is the second most popular 2G standard, with the CDMA Development Group reporting more than 270 million users. TDMA has a far smaller subscriber base, and even this appears to be declining. Globally, reports varied between 75 and 100 million TDMA subscribers, with numbers in the United States declining some 23 per cent in 2004. The Personal Digital Communica-tions (PDC) standard comprised the smallest segment of global mobile users (chiefly dominating the Japanese landscape). Many PDC users have been migrated over to 3G services (through NTT DoCoMo) or had their service terminated entirely, due to KDDI’s decision to inaugurate CDMA 1X service.

CHAPTER 1 9


In Asia, for example, Pakistan Telecommunications Com-pany Limited (PTCL) announced plans to launch a triple play service, featuring telephony, cable TV and DSL services over the same line.8 Nearby, in India, state-owned telephone com-pany Mahanagar Telephone Nigam Ltd. plans to launch that country’s first ‘triple play’ service, in Delhi and Mumbai; with the state-owned sister company BSNL covering the rest of the country.9 Bahrain was set to receive its first triple play voice, video and data service from incumbent telephone company Batelco, which struck a deal with an equipment supplier to deploy infra struc ture nationwide.10

In Eastern Europe, Romanian competitor Astral Telecom completed a USD 1 million upgrade to its cable-based IP net-work to provide an enhanced triple play, with voice, video and broadband data services.11 Spain’s dominant fixed-line operator, Telefónica, launched a triple play package of voice telephony, broadband internet and multi-channel TV services over tradi-tional copper lines, reaching at least 100,000 subscribers.12 Ital-ian fixed-line tele com mu ni cation operator Unidata launched its new “ADSL2+” service, enabling users to download voice, video and data at speeds of up to 24 megabits per second (Mbit/s).13 Expanded service upgrades were also under way in France and Germany.14

Mobile operators and equipment manufacturers are also benefiting from convergence. One mobile operator recently unveiled a personal digital assistant (PDA) that combines 3G mobile technology with a miniature laptop personal computer design and functionality, enabling mobile voice and internet through 3G, Wi-Fi, and a Windows mobile operating system.15

1.4 Competition

Over the past decade, the introduction of competition into the second-generation mobile sector has been one of the key factors linked to its success in increasing teledensity in developing countries (along with prepaid cards and cheaper network deployment costs). Many developing countries now seek to replicate this success with broadband services. The level of competition that is authorized by countries (based on the number of competitive players), as well as the kinds of services opened to competition, will remain key to these strat-egies. Competition in international services, in particular, the international gateway, and leased lines, for example, is vital to ensuring low cost internet access.

Figure 1.11 illustrates that competition is authorized in basic telephone services (defined as local, long distance and international) as well as in leased lines in more than 60 per cent of countries worldwide. Competition is authorized in around 90 per cent of countries for services such as VSAT, Cable TV, internet access, and IMT2000 (3G) offerings, and by more than 80 per cent of countries for DSL and fixed wire-less broadband services. The region with the lowest levels of competition is the Arab States, where less than 40 per cent of markets have been opened to competition. Europe is the most competitive region, while Africa is nearly evenly split between monopoly and competitive conditions. Competition prevails in the Americas, Asia and, indeed, around the world as a whole

– by a ratio of 3 to 2.

1.4.1 Interconnection is Key to Competition

It is one thing for countries to authorize competition in the provision of ICT services. It is another to ensure that new market entrants are actually licensed or otherwise authorized to provide services and are able to compete in the market on

Figure 1.10: Privatizations, 1991-2005, World

Source: ITU World Telecommunication Regulatory Database.

CHAPTER 1


10

a level playing field. One of the key issues in ensuring a level playing field is a fair and transparent inter con nection regula-tory framework. Figure 1.12 illustrates the extent to which inter con nection agreements, or Reference Interconnection Offers (RIOs), are made public, recognized as one of the best practices in inter con nection regulation today, ensuring that all competitors are aware of and can benefit from the same inter-con nection rates. Transparency in inter con nection agreements is highest in the Americas, Asia-Pacific and Europe. In Europe,

pricing is public in approximately 72 per cent of countries, but this is not the case in other regions. Worldwide, nearly 60 per cent of countries do not make inter con nection agree-ments public, although about 58 per cent make pricing infor-mation available. This is one area regulators could assist, for example, by publishing RIOs on their websites. ITU provides links to published RIOs as part of the Regulatory Profiles sec-tion of its TREG website (http://www.itu.int/ITU-D/treg/pro-fi les/Interconn.asp).

Figure 1.11: Status of Competition Worldwide, 2005Level of competition in selected services and networks, World


Level of competition in basic services, per region

CHAPTER 1 11


1.4.2 Regulating Local Loop Unbundling

Opening up access to the local loop for competing opera-tors is gaining ground as a means to foster full competition. At the end of 2005, 72 countries worldwide had required unbun-dling (See Figure 1.13). The debates about unbundling mirror discussions about competition in general. One argument is that mandating unbundling creates disincentives for future infra struc ture investments. In the past, the local loop was considered a “natural monopoly,” so it is not surprising that many incumbents feel they should be able to retain the full benefits of their investments. This argument is often coupled

with the assertion that incumbents are in the best position to channel their profits into large-scale expansion of last-mile networks. Incumbents have not, however, always lived up to those promises. Local-loop unbundling, therefore, is increas-ingly recognized as being important not only for competition in traditional telephone services but also to prevent the incum-bent’s monopoly from spilling over into the domain of broad-band internet.16 Therefore, the goal of unbundling has often been to make network components available to new market entrants with greater commitment, expertise and incentives to improve services.

Figure 1.12: Status of Public Interconnection Agreements and Pricing Information, 2005


CHAPTER 1


12

1.4.3 Internet Exchange PointsInternet Exchange Points (IXPs), also known as network

access points (NAPs), are a way to maximize the existing infra-structural base for internet service provision in developing and transitioning countries and foster competitive provision of internet services. They allow Internet Service Providers (ISPs) to exchange traffic between their networks through mutual peering agreements at minimal cost, with better efficiency (in terms of bandwidth and latency) and increased speed. Net-works can inter con nect directly, via the exchange, rather than through third-party networks, and the cost of diverting traf-fic to upstream providers or to faraway places is minimized by keeping it within and between adjacent ISPs. For many devel-oping countries, IXPs keep traffic local or regional, instead of routing internet traffic to hubs like New York or London

– just to reach an email account on a different ISP in the same country or to exchange internet traffic with a neighbouring country. Figure 1.14 illustrates the number of countries, per region, that participate in national or subregional IXPs.

A study prepared for the 2004 Global Symposium for Reg-ulators presented several issues related to creating a conducive environment for IXPs, including the importance of clearing the regulatory obstacles that exist at subregional levels, ensur-ing competition at the level of the international gateway, open-ing up VSAT use, allowing for co-location of equipment in an incumbent operator’s facilities, and ensuring that licensing or other additional burdens are minimized or eliminated.17

1.5 National Regulatory Authorities, Who Rules?

The establishment of a separate regulator is one of the most visible signs of sector reform. Separate regulatory agen-cies cannot compensate for boom and bust cycles in financial markets and the macro-economy, but they can certainly go a

long way in laying the grounds for a favourable investment cli-mate and promoting market opportunities.18 It is well docu-mented that the degree of actual and perceived autonomy from government control and industry influence is a key indicator of the effectiveness of a regulator. Regulatory procedures should be clear, transparent, and predictable. Regulators, in enforc-ing these procedures, should be accountable, and should have sufficient credibility and authority to enforce the relevant laws and regulations.

Over the last five years, the number of regulatory authori-ties worldwide has increased by approximately 36 per cent, with new regulators on the verge in Samoa and Liberia. Regu-latory activity is most notably on the rise in the Arab States, Africa and Asia. Figure 1.15 depicts the growth of these regula-tors worldwide.

Meanwhile, figure 1.16 shows the regional variation in countries that have functional regulators in place, breaking out the percentage of countries in each region that have established regulatory authorities. It is evident that the presence of regula-tors in the Americas and in Africa (relative to the total number of countries in each region) is higher than in the Arab States or in Asia. In absolute numbers, the region with the highest number of regulatory authorities is Europe.

Merely establishing a regulator – while a step in the right direction – does not ensure effectiveness or relevance. Priori-ties have to be established, the necessary statutory framework in which the regulator will operate must be created, proce-dures put in place, and relationships between stakeholders established. The way in which these processes work together and are respected within a general context of good governance will determine the effectiveness of the regulator and ultimately the success of the market.19

Figure 1.13: Countries Requiring Local Loop Unbundling, 2005


CHAPTER 1 13


Some countries possess more collegial, cooperative and intertwined policy formulation climates than others. Some governments permit informal decision-making processes, while others mandate use of formal mechanisms. Retroactive applications of influence can reverse or undermine regulatory decisions. More fundamentally, some countries may consider their regulatory bodies independent and separate, even when in reality they may be subordinate to a ministry or subject to

oversight from a higher authority. In the end, the most impor-tant element is effectiveness.20

Another key question facing policy-makers and regula-tors is what kind of regulatory body to create. Worldwide, most countries have what can be termed a “single-sector” regulatory authority for tele com mu ni cations. However, with the conver-gence of different communication services and technologies

“converged” regulatory agencies have started to emerge. Coun-tries such as Austria, Finland, France, Italy, the Netherlands,

Figure 1.14: Number of Countries with National and SubRegional IXPs, by region, 2005


Figure 1.15: Growth of Regulators Worldwide, 2005


CHAPTER 1


14

Saudi Arabia, Singapore, South Africa and the United King-dom have taken this path.21 There is evidence that a unitary, converged regulatory structure may yield more efficient inter-nal ad min is tration, institutional flexibility and lower regulatory costs.

Interestingly, multi-sector regulatory authorities have also emerged in the last few years, giving oversight not only of the tele com mu ni cation sector, but also of other industry sectors with common economic and legal characteristics (for example, water, energy or transportation). Costa Rica, Gambia, Germany, Jamaica, Latvia, Luxembourg, Niger and Panama have chosen this model, which has long been standard for public utility commissions in individual states in the United States.22 One of the arguments for such a structure is centralizing and optimiz-ing limited regulatory skills, as well as reducing costs through economies of scale. However, there are know disadvantages of this model including dilution of sector specific expertise, fail-ures of a multi-sector authority cascading through multiple regulated sectors, increased risk of political capture, delays in reform, etc.

The fourth approach is when countries have opted to have no national regulatory authority per se, and the functions of sector regulation rely on competition and antitrust rules.

1.6 Regulatory Challenges of VoIP Many of the world’s carriers have been persuaded to

deploy IP-based networks that can carry both voice and data. In this way, operators are able to invest in a single network that can be used more efficiently for many different forms of traf-fic. Many of these operators have started to offer VoIP to their customers.

Recognition of this shift in practice and service is wide-spread and international. In fact, international VoIP increased by 35 per cent from 2003 to 2004 (See figure 1.17). For exam-ple, 20-25 per cent of all historic operators in Africa were using VoIP to carry part of their international traffic in 2004.23

Telkom Kenya currently offers a VoIP-based international serv-ice. Five African carriers – in South Africa, Botswana, Angola, Namibia and Uganda – have announced that the introduction of IP-based networks is imminent, while Mexican incumbent Telmex (amongst various others carriers) has already imple-mented IP for the majority of its core network. Among the Arab States, Oman’s incumbent Omantel has committed itself to creating an end-to-end IP communication services network.

VoIP is also being offered by new market players, which are often viewed as a threat to traditional PSTN operators. VoIP may be offered by ISPs, in internet cafés and, more recently, by companies from abroad that have no local presence, yet whose impact on local market conditions is felt quite strongly. The impact of VoIP is felt from the loss of outgoing international retail traffic (as customers search for the lowest cost interna-tional rates) to reductions of incoming international settlement traffic (as traffic from VoIP customers abroad skirts the inter-national settlement rate system).

Regulators participating in the 2005 ITU Global Sym-posium for Regulators (GSR) recognized that although VoIP poses increasing challenges to legacy operators, it also brings new opportunities to end users for more affordable services.24 In many ways, the rise of VoIP has crystallized the delicate bal-ancing act that many regulators have been performing as regula-tory reform has been implemented ever more widely. Weighing in on one side of the scale are the commitments of the World Summit on the Information Society (WSIS) to encourage low cost access to ICT services, while the other end of the scale balances the desire to protect incumbent operators-especially when incumbents remain at least partially government owned.

It is not surprising, therefore, that the rise of VoIP has prompted an array of regulatory responses, from outright bans to full legalization. The question of whether to allow or to prohibit VoIP, however, is only one of many issues prompted by the rise of VoIP. Other issues include developing regulatory frameworks for the inter con nection of circuit-switched and IP-based networks as well as transitioning to inter con nection in

Figure 1.16: Separate Regulators, by Region, 2005


CHAPTER 1 15


Table 1.1: Newly Created Regulatory Authorities, 2005

Country Regulator Year created Structure FunctionsCape Verde Instituto das

Comunicações e das Téconologias de Informação

2004 Collegial body of 3 members

Licensing, interconnection, price regulation, technical standards, radio frequency allocation, numbering, type approval, monitor service quality, universal service, quality of service

Trinidad & Tobago Telecommunications Authority of Trinidad and Tobago

2004 Collegial body of 11 members

Interconnection, price regulation, Radio frequency allocation, numbering, monitor service quality, quality of service standards, and licensing and universal service with the sector Ministry

Thailand National Telecommunications Commission (NTC)

2004 Collegial body of 7 members reporting to House Senate

Interconnection, price regulation, technical standards, radio frequency allocation, numbering, type approval, monitor service quality, quality of service

Qatar Supreme Council for Communication & IT

2004 Headed by Secretary General

Mandate to create a legal and regulatory environment that promotes development of ICTs

United Arab Emirates

Telecommunications Regulatory Authority

2004 Collegial body of 5 members reporting to UAE Telecom Supreme Committee

Licensing, interconnection, price regulation, technical standards, radio frequency allocation, numbering, type approval, monitor service quality, universal service, quality of service standards

Niger Autorité de Régulation Multisectorielle (ARM)

2004 Headed by a president with a sector director for each sector it regulates.

Transport, energy, water, and telecommunication

Iran Communications Regulatory Authority

Law passed in 2003 and operational

in 2005

Collegial body of 7 members reporting to Minister of ICT

Licensing, interconnection, technical standards, radio frequency allocation, numbering, type approval, monitor service quality, universal service, quality of service standards

Afghanistan Afghanistan Telecom Regulatory Authority (ATRA)*

Law passed in December

2005. Operational

mid-Jan 2006

All regulatory functions, including licensing & compliance, spectrum planning & assignment, numbering, ensuring network interconnection, promoting competition and consumer protection

TFYR Macedonia Agency for electronic communications

2005 Collegial body of 5 members reporting to Parliament

Licensing, interconnection, price regula-tion, technical standards, radio frequency allocation, numbering, type approval, monitor service quality, universal service, quality of service standards

* The ATRA was established from the merger of the Telecom Regulatory Board, created in 2003, with the State Radio Inspection Department. See http://trb.gov.af/trb.htm


a fully IP-based world. There is also a range of issues arising from the fact that VoIP customers can use the same VoIP serv-ice nomadically, rather than from one fixed line location or one mobile terminal. The nomadic nature of VoIP is prompting challenges to traditional practices in numbering and emergency services. The variety and intricacies of the regulatory treatment of VoIP are explored in Chapter 6.

The rise of VoIP is emblematic of the crossroads at which the ICT sector now finds itself. Regulatory practices and wisdom built upon the experiences of the heavily regulated PSTN era are now meeting head on the largely unregulated internet world. Which model will apply as these two worlds converge? Or will hybrid or entirely new regulatory models be developed? Chapter 6 begins to identify the issues and prac-

CHAPTER 1


16

tices facing regulators today and in the years ahead as IP net-works replace circuit switched infra struc ture.

1.7 Spectrum Management

A pragmatic framework for managing spectrum resources effectively includes ways of facilitating deployment of innova-tive broadband technologies. Spectrum management to pro-mote broadband access is discussed more fully in Chapter Five. Strategically balancing the use of unlicensed and licensed spec-trum, for example, is becoming a key component of effective regulation.

Across regions, a lack of consistency among national spec-trum policies – particularly on unlicensed “commons” models

– is becoming an issue. Lack of coherence from country to country in the fees and costs associated with spectrum access may be manageable in the short-term, but it is likely to be fun-damentally problematic in the long term, as innovations sweep through the market in disregard of state borders. Gradual con-vergence of regional policies and fees has been encouraged, particularly through actions of the European Commission, for example.

The disparity between massive fees paid for 3G licences (in Europe) and the unlicensed use of wireless local area net-works (WLANs) looms as a potential regulatory issue. It will loom larger if both technologies converge into a single broad-band wireless market. For now, it appears that licensing as a regulatory mechanism is being applied in limited ways to the realm of WLANs. Figure 1.18 shows that the majority of countries in all regions do not to require spectrum licences for WLANs. Perhaps in light of the numerous, uncontrolled and rather random deployments of WLANs as part of small-scale development projects (often launched in lieu of fixed-line infra-

struc ture), developing countries in Africa, the Americas and Asia appear less eager to regulate WLANs than their counter-parts in Europe, where WLANs are just one potential solution to broadband connectivity.

1.8 Dealing with Spam

Over the last decade, the unbridled growth of spam has gained increasing attention, not only due to its inconvenience and cost, but perhaps even more importantly, because spam often carries viruses and worms or poses other network secu-rity issues, or is used a vehicle for fraudulent behaviour. Today, there is general agreement about spam’s core characteristics, including that it consists of unsolicited electronic messages sent in bulk. “Spammers now employ a variety of advanced upload methods such as open mail relays, insecure Web proxies, malformed CGI scripts and zombied clueless-user machines.”25 This means that spam messages tend to be identi-cal and are sent indiscriminately to selected recipients. Most experts involved in the fight against spam counsel in favour of a multi-pronged approach, including technical solutions, legal and regulatory actions, end-user education and international cooperation.

According to some analysts, spam accounted for around 70 per cent of all e-mail traffic by mid-2005 (see Figure 1.19). The costs associated with spam are difficult to determine, although it is logical to assume that it puts pressure on ISPs in terms of reduced bandwidth and increased storage costs – not to men-tion the burden of dealing with customer complaints. The European Commission has estimated the cost of spam to inter-net users worldwide to be around EUR 10 billion a year, and a recent study has estimated the loss of productivity due to spam messages to be at USD 1,930 annually for each employee.26 In

Figure 1.17: Growth of International Traffi c

Source: ITU World Telecommunication Indicators Database and TeleGeography Global Traffic Statistics 2006 report. Telegeography Research is now part of PriMetrica, Inc. (see www.primetrica.com)

CHAPTER 1 17


Figure 1.18: Are Licences Required for WLANS, 2005?


Figure 1.19: Spam as Percentage of Emails Worldwide, 2003-05

Source: Message labs.

marked contrast, the costs of startup and operation for spam-mers are extremely low, and the architecture, based on Simple Mail Transfer Protocol (SMPT), allows them to work anony-mously.

The success of legislating and making policies effective in countering spam has been limited thus far. In 66 per cent of all countries, there is no single, identifiable entity responsible for combating spam (see Figure 1.20). Only thirty-two countries have passed anti-spam legislation. As a region, Europe has the greatest focus on anti-spam measures, although international

attempts at stand ardizing business practices – or at least har-monizing ISPs’ approaches in countering spam – are growing.

To date, anti-spam laws have focused mainly on tracking down and prosecuting spammers. Such anti-spam laws require considerable investigative and enforcement resources, the very resources that often are in short supply in developing countries. While anti-spam laws targeted at spammers remain an essential tool in the anti-spam arsenal, their use by developing coun-tries may more likely be as the foundation for international cooperation. Anti-spam authorities with more experience and

CHAPTER 1


18

Figure 1.20: Spam Regulation, 2005


resources may seek to work with regulators in developing countries in tracking down and prosecuting spammers. Having an enforceable anti-spam law in place as part of a coordinated international effort will facilitate action against spammers acting (and hiding) across multiple jurisdictions.

But the time may also be ripe for anti-spam authorities to expand their efforts to include working with ISPs, who can be instrumental in fighting spam. Chapter 7 therefore looks not only at the components of anti-spam laws targeted at spam-mers, but proposes the establishment of enforceable codes of conduct to be developed by ISPs, and then approved and enforced by regulators. Such a system of ‘managed self-regu-lation’ would require ISPs to prohibit their customers from using that ISP as a source for spamming and related bad acts, such as spoofing and phishing, and not to enter into peering arrangements with ISPs that do not uphold similar codes of conduct. Rather than continue to rely upon chasing individual spammers, regulators in the most resource-constrained coun-tries in particular would be more likely to succeed by working with and through the ISPs that are closer to the source of the problem, to their customers, and to the technology in ques-tion. The regulator’s job would be to ensure that ISPs within their jurisdiction adopt adequate codes of conduct and then to enforce adherence to those codes.

While some ISPs can be expected to resist even such light-handed regulation, the advantage is that it places all ISPs on a level playing field. Under current practices, responsible ISPs find themselves bearing the brunt of the costs of spam. This explains why some ISPs have begun suing spammers for dam-ages, an option that may not be available in all jurisdictions. The goal of managed self-regulation is to reduce spam in a way that protects responsible ISPs. ISPs that implement responsi-ble, effective anti-spam measures should be rewarded for their

good behaviour. One means of rewarding those responsible ISPs is for regulators to hold their irresponsible competitors accountable. Regulators can also make consumers aware of the good works of the best ISPs, for example, by certifying ISPs that enforce their codes of conduct and allowing such ISPs to use the regulator certification in their advertising. As with many other tele com mu ni cation-related policy issue that is salient across national borders, the importance of consistency, shared strategic approaches and international cooperation is paramount.

1.9 Conclusion

Since the release of the 2004 edition of Trends in Tele-communication Reform, the global ICT sector has retained its dynamism, although potential challenges to continued growth loom large ahead. The prevailing trend of liberalized markets, privatized operators, separate regulatory authorities and new regulations aimed at tackling the phenomena of convergence are very real. It is unlikely that countries today will attempt to shut down competition and bolster the formation of new monopolies, although some operators upgrading to next gen-eration networks will have to be watched closely to counter such tendencies. Regulators will continue to play a key role in ensuring that the values of transparency, cooperation, and market competition now being championed in the ICT sector become more real and apparent over time.

As institutions and frameworks are gradually being re-designed, a clear message is being disseminated: the role of reg-ulators is a critical catalyst to the process of reform in the ICT sector. Broadband internet access (whether through fixed lines or wireless) is becoming increasingly relevant to the demands of subscribers in developed and developing countries alike.

CHAPTER 1 19


The essence of voice telephony is being transformed. Next-generation networks are being designed and developed, even as 3G services begin to gain widespread acceptance. Through all of these trends, one thing appears certain: the sector is tending

toward a more open, competitive, and transparent model, in which governments, operators, development agencies, educa-tional institutions, civil society groups, and end users all have equally important stakes.

1 A comprehensive picture of the state of current ICT infra struc ture deployment may be found in the 2006 World Telecommunication Development Report, being released at the same time as this publication.

2 3G Today Newsletter. January 2005. Volume 2. Issue 1. See: http://www.3gtoday.com/wps/portal/!ut/p/kcxml/04_Sj9SPykssy0xPLMnMz0vM0Y_QjzKLN4r3DAb-JmMUbxBub6keiijjCBXw98nNT9b31A_QLckMjyh0VFQEBN7t9/delta/base64xml/L3dJdyEvUUd3QndNQSEvNElVRS82XzJfSjI!?newsletterId=1180

3 Aytar, Ozgur, Pyramid Research cited in CommsDay Global, November 5, 2004, p.4.4 “Pain or Gain: The Year Ahead for Mobile”, 3G Analyst Predictions Lucent Newsletter, Issue 18, February 2005. 5 “Brazil Telecom Launches Triple Play”, Telegeography Commsupdate, October 14, 2005. 6 “Antel Trials 3G and Triple play Services”, Telegeography Commsupdate, September 28, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=9228. 7 “Chilean companies invest heavily in triple-play networks”, Telegeography Commsupdate, September 13, 2005. Link: http://www.telegeography.com/cu/article.

php?article_id=9004. 8 “PTCL To Launch Triple Play Service”, Telegeography Commsupdate, May 3, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=6995. 9 “MTNL Launches triple-play (India)”, Telecomfinance.com, September 2005. Link: http://www.telecomfinance.com/nodes/Main-Pages.html. 10 “Batelco to Launch Triple Play”, Telegeography Commsupdate, June 20, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=7690. 11 “Astral Telecom carries out USD1 million voice-over-cable upgrade”, Telegeography Commsupdate, September 21, 2005. Link: http://www.telegeography.com/cu/

article.php?article_id=9116. 12 “Telefónica talks up triple play”, Telegeography Commsupdate, October 14, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=9465. 13 “When in Rome… download at 24Mbit/s”, Telegeography Commsupdate, June 22, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=7736. 14 “Swisscom runs into May triple play delay”, Telegeography Commsupdate, May 31, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=7433. 15 “Orange Launches Laptop Style 3G Phone/PDA For Business”, Businessmobile.com, September 22, 2005, Link: http://www.businessmobile.com/index.php?name

=News&file=article&sid=51. 16 Regulators from West Africa recognized the importance of local loop unbundling in a set of guidelines on intercon nection agreed by the West African Telecommu-

nications Regulators Assembly (WATRA) http://www.itu.int/newsarchive/press_releases/2005/12.html 17 Russell Southwood, “Via Africa: Creating local and regional IXPs to save money and bandwidth”, International Telecommunication Union, 2005, p. 23, Link: http://

www.itu.int/ITU-D/treg/publications/AfricaIXPRep.pdf.18 Melody, W.H. “Stimulating Investment in Network Development: Roles for Telecom Regulation”, World Dialogue on Regulation of Network Economies, March

2003, p. 9. 19 Forthcoming ITU/infoDev ICT regulation toolkit, Module on legal and institutional aspects of regulation, 2006. Link: http://www.ictregulationtoolkit.org/20 For further reading on effectiveness, consult the ITU 2002 Trends in Telecommunication Reform: Licensing in an Era of Convergence. http://www.itu.int/ITU-D/

treg/publications/Trends02_ExecSummary.pdf21 Forthcoming ITU InfoDev Toolkit, Organizational and Institutional Approaches to Regulation, Chapter 6, p. 9-10.22 See http://www.itu.int/ITU-D/treg/index.html. See ARESEP (Costa Rica) http://www.aresep.go.cr/cgi-bin/menu.fwx; OUR (Jamaica) http://www.our.org.jm/; Ente

Regulador de los Servicios Publicos (Panama) http://www.ersp.gob.pa/default.asp. inks to U.S. State PUC’s can be found at http://www.dps.state.ny.us/stateweb.htm.

23 Muleta, John. VoIP Chapter X, pp. 24 See http://www.itu.int/ITU-D/treg/Events/Seminars/2005/GSR05/Documents/chairmansreport.pdf25 Race, Jeffrey, “You needn’t eat spam (or worms) The real reasons why spam still exists today – and what to do about it”. Free Software Magazine, Issue 6, August

2005.26 Nucleus Research “Spam: The Serial ROI Killer”. Link: http://www.nucleusresearch.com/index.html

21


CHAPTER 2

Authors: Michael Best, Georgia Tech; and Bjorn Pehrson, Swedish Royal Institute of Technology

There is a lot of discussion around the world about the advent of broadband technologies and services – and justifiably so. The long-awaited broadband revolution now appears to be gaining momentum, in developing countries as well as devel-oped ones. This chapter seeks to take the measurements of this broadband growth by defining what policy-makers and tech-nologists mean when they say “broadband. This will set the stage for Chapter 3, which explores broadband technology at a more technical level. Chapter 3 also looks ahead, providing a roadmap for regulators to plot the rapid technological changes that are resulting in new opportunities and services, on a nearly daily basis.

The range of broadband systems that are in the market and commonly employed today can be organized into three broad families of technologies:• Broadband wire-line networks, including DSL over

twisted-pair copper cables, cable modem services over cable TV systems, and fibre networks;

• Broadband wireless solutions, including 3G mobile services, wireless LANs (WLANS) and other fixed and mobile wireless access solutions; and

• Non-terrestrial options, including satellite systems employing very-small aperture terminal (VSAT) dishes.

This chapter will discuss the types of equipment, infra-struc ture and software that are needed to deploy each and their viability for rural and underserved areas of developing coun-tries. The picture that emerges is that broadband development is no longer the exclusive preserve of developed countries and high-income communities. Increasingly, developing countries are employing low-cost technologies – many of them wireless

– to strategically introduce broadband capabilities.

Moreover, it is not just large, incumbent operators that have opportunities to deploy broadband facilities and services. There are opportunities for non-traditional operators – such as universities, local governments and community-based groups

– to invest in broadband capabilities and link their systems to the broader internet. The combination of non-traditional entrepreneurs, local governments and major incumbents may provide the formula to drive broadband development in many countries.

2.1 What Do We Mean When We Say “Broadband”?

Like many terms used in today’s fast-moving technology sector, the term broadband is not well defined. The word was originally used in the network engineering community to sig-nify transmissions carrying multiple channels simultaneously. This was contrasted with baseband, which involved transmit-ting on only a single channel at any one time. Today, however,

“broadband” is used much more frequently to indicate some form of high-speed internet access or data transmission.

Deciding which networks provide sufficient capacity to be called broadband is open to debate. There have been many attempts to associate the term with a particular speed or set of services, but in reality, broadband is a moving target. Internet speeds are increasing constantly, and with each new advance, marketers eagerly emphasize just how blazingly fast the latest connection speeds are.

The speed of a network is usually expressed in terms of its data transmission rate, which is measured in kilobits or mega-bits per second (this is also known as the bandwidth of the net-work).

What is not up for debate is that today’s dial-up internet access speeds, topping out at about 56 kilobits per second (kbit/s), are not broadband connections. But beyond that, defining broadband seems to be subjective. The U.S. Federal Commu-nication Commission (FCC) has defined broadband as start-ing at 200 kbit/s. The OECD sets the bar at 256 kbit/s, and the ITU defines broadband as a combined capacity (upstream and downstream) totalling 256 kbit/s or more.

A good example of the broad range of definitions can be found within a single country. The Swedish IT Commission (1994-2004) has defined broadband as supporting a formida-ble 5 megabits per second (Mbit/s) upstream and downstream. But the Swedish government has said that broadband starts at 2 Mbit/s, up and down, while Sweden’s incumbent operator, Telia, defines it as at least 500 kbit/s up and down. In Swedish metropolitan area networks (MANs), meanwhile, 10 Mbit/s up to 100 Mbit/s has become the standard. While there may be a diverse array of parameters in the marketplace, for the pur-poses of this report, the term broadband refers to data rates that correspond to the ITU’s definition in this section.

2 WHAT IS “BROADBAND”?

CHAPTER 2


22

Beyond precise data-rate thresholds, perhaps a more useful way to define broadband is by discussing what one can do with it. This certainly includes fast internet downloads from the Web. But it also should include compact disc-quality streaming audio, fully interactive voice services such as VoIP, some level of interactive video chat services (if not full-capabil-ity video conferencing), and at least reasonable-quality stream-ing video services (if not full, DVD-quality video on demand). Note that receiving this wish list of services and applications is not simply a matter of available bandwidth. Interactive applica-tions such as VoIP also require little or no latency (delays), low error rates, and minimal jitter (the result of data arriving at its destination out of order).

2.2 A Short Broadband Taxonomy

As noted in the introduction to this chapter, broadband networks can be divided into three major categories: broad-band wire-line networks, broadband wireless networks, and non-terrestrial (that is, satellite) networks.1 These categories reflect the different development paths for broadband technol-ogies that are now available in the marketplace. The following subsections trace those different paths and briefly describe the current state of development of standards and capabilities.

2.2.1 Broadband Wire-Line Networks

In general, internet access obtained its first foothold as an outgrowth of wire-line networks. The earliest internet access was, of course, through the public switched telephone network (PSTN), via dial-up modems or leased telephone lines, to an internet service provider (ISP). Data communications using a dial-up telephone connection require an analogue modem at both ends of the telephone line. Traditional modems encode data in the same frequency band as a voice call (up to 4 kilo-hertz). The user can either speak or send data. Data rates vary between 2.4 kbit/s and 56 kbit/s, depending on the quality of the analogue copper telephone line, whether or not the net-work operator’s central office switch is digital, whether the switches are clock-synchronized, and whether the switches are connected using modern links, such as fibre or microwave.

From an economic point of view, dial-up modems are cheap and typically integrated into personal computers. Carrier-class dial-up servers required at the ISP end are quite expensive, while lower-class equipment that can be used by rural entre-preneurs on a small scale is cheap (less than USD 500) and supported by open-source software. From the end user’s point of view, the cost typically includes the local telephone call tariff, plus the ISP’s internet access fee, which may be flat-rated or connection-time based (in many areas of the United States, the local calling rate is also flat). Dial-up service continues to be a standard form of internet access in many countries, but, as previously mentioned, it does not meet most definitions of broadband. Dial-up service has paved the way, however, for the broadband technologies discussed in the following subsections.

2.2.1.1 Digital Subscriber Line (DSL)

DSL is actually a family of technologies that provide a dig-ital connection in an unused part of the frequency spectrum of the telephone network’s copper-wire subscriber line.2 The circuit-switched voice and packet-switched data connections can thus be used independently of each other. DSL technol-ogy provides a significant enhancement to the already installed PSTN base, protecting the sunken investment value of the copper network. For that reason, DSL is one of the most popu-lar broadband technologies. According to the ITU’s World Tele-communication Indicators database, DSL is used to provide over 60 per cent of worldwide home broadband connectivity.

The bandwidth that DSL systems can provide has been increasing. There are now systems that can provide transmis-sion speeds from 256 kbit/s up to 1.2 Mbit/s upstream and 512 kbit/s to 28 Mbit/s downstream. DSL bandwidth limitations are caused by the attenuation of signals at higher frequencies. The amount of attenuation on any network depends on the qual-ity of the copper lines and their installation. Moreover, DSL’s viability in any given location depends on the distance between the subscriber and the exchange, which usually must be within 5 kilometres.

To deploy DSL, equipment must be added at both ends of the subscriber line. At the user end, a DSL modem and a cheap passive splitter must be installed. The passive splitter plugs into the existing telephone socket and splits the incoming signal between the telephone and the DSL modem. On the other side of the modem, the user can directly connect a computer or set up a LAN via a customer premises gateway (often for less than USD 100). The DSL modem converts signals from the data format used in the local area network environment (mostly an Ethernet LAN or IEEE802.3 signal), into a digital audio stream. On the operator side, before the subscriber line is connected to the telephone exchange, the DSL circuits are separated and terminated in a digital subscriber line access multiplexer (DSLAM), which aggregates the digital connections from different users and feeds them to the ISP network.

2.2.1.2 Cable TV System Networks

Cable TV (CATV) networks use coaxial cable to reach all users in a point-to-multipoint topology. Initially, of course, CATV was employed only for distribution of television chan-nels in a tree-structured network, created by using passive splitters. Broadband communication over cable TV networks is accomplished by transferring data via unused bandwidth in the cable, in a way similar to what DSL does over the PSTN. The standard is Data over Cable Service Interface Specification (DOCSIS).3 The basic data rates are 54 Mbit/s downstream and 3 Mbit/s upstream. An ISP connects to the cable compa-ny’s central office (known as the head end by CATV operators) and uses the cable network to connect to users.

As with DSL, equipment has to be installed at both the head end and at the customer premises. On the operator’s side, a cable modem termination system (CMTS) is installed at the head end. It separates the digital communication channel from the television circuits, aggregates connections from different users and feeds them into the ISP network. On the user’s side,

CHAPTER 2 23


a splitter and a cable modem must be installed at the home (most cable modem subscribers are residential). The splitter divides the incoming signal between the TV set and the cable modem. On the other side of the modem, the user can con-nect a computer or a residential gateway via an Ethernet port and a USB telephone, allowing VoIP service if provided by the ISP or other provider.

From an economic point of view, broadband over cable is favourable in areas where there is already an existing cable network. The tree-structured, point to multipoint technique has, however, several severe disadvantages compared to point-to-point solutions:• For the regulator, it is less attractive, since it may prevent

local loop unbundling on some technical layers of the net-work;

• For the operator, it is more complex to plan, manage and upgrade; and

• For the user, the performance depends on how much traf-fic there is from other users connected to the local trans-mission lines.

Users in a neighbourhood (typically, 100 to 2000 homes) share the available bandwidth provided by a single coaxial cable line. Therefore, connection speeds can vary between 10 Mbit/s and a few kbit/s, depending on the volume of traffic from other users. While most networks share a fixed amount of bandwidth between users, cable networks are generally spread over larger areas and require more attention to performance issues. The broadcasting technique also raises concerns regarding security and privacy. To address these concerns, the DOCSIS standard includes encryption and other privacy features that are sup-ported by most cable modems.

2.2.2 Broadband Wireless AccessThe ITU defines broadband wireless access (BWA) as

encompassing either mobile or fixed access technologies that provide connections at speeds higher than the pri-mary rate (for example, 2 Mbit/s). Up until now, wireless broadband has somewhat lagged behind the development curve of its wire-line cousins. Throughout the 1990s, mobile net-works, for example, were primarily viewed as voice offerings, and broadband fixed wireless systems (many of which relied on line-of-sight technologies) did not enjoy the same take-up rates in developed countries that DSL and cable modem serv-ices did. Wireless broadband options are, however, becoming increasingly more available and more functional (see Chapter 3). The following subsections trace the development paths of mobile and fixed wireless terrestrial systems.

2.2.2.1 Mobile Services: GSM and CDMA

Second generation (2G) GSM networks were initially implemented in Europe and then Asia. They have since been installed across much of the world. The original GSM systems support only very limited data capacity (well below that required to be labelled broadband). On the path toward broadband, the next step for GSM-based data services has been

“2.5G” capability. This is provided using technologies known as general packet radio service (GPRS) and the enhanced data GSM

environment (EDGE). GPRS offers maximum speeds of 171.2 kbit/s, while EDGE can triple those rates. A typical cell radius for these 2.5G networks is 500-1 000 meters, meaning that to achieve complete coverage, every point in the coverage area can be no more than 1 km away from an antenna. GPRS and EDGE technologies allow mobility at vehicular speeds and can handle seamless handoffs between cells.

Meanwhile, upgrades from 2G to 2.5G networks can be expensive, since they generally require both software and hardware changes at the base station. They may also require improvements to backhaul networks, as well as increased con-nectivity at the core network. In addition, in many settings, high costs for spectrum licensing (for example, in the 2 GHz band) add to the upgrade costs. Subscriber handsets also need to be upgraded to support 2.5G, and handset replacement cycles have often been a central component of adopting new technologies. While estimating the cost is difficult, common im ple men ta tions have experienced costs per base station above USD 100,000. In other words, upgrade costs can be on the order of USD 50 or more, per subscriber, in low-population density areas.

The second major category of 2G cellular technologies is the code-division multiple-access (CDMA) IS-95 family. These systems were developed by the U.S. company Qualcomm and are used primarily in the United States. They do not employ the time-division multiple-access (TDMA) modulation approach of 2G GSM. Instead, they carry multiple transmissions simul-taneously by filling the channel with packets encoded for their specific destination devices.

There is a family of upgrades for CDMA networks called CDMA2000. This includes the CDMA2000 1x system, which supports data rates up to 307 kbit/s. The CDMA Development Group (CDG) reports their CDMA2000 subscribers as “3G” users. Some operators and equipment manufacturers, however, believe that the real equivalent to 3G is CDMA2000 1xEV (for

“evolution”), which is a higher-speed version of 1x. Within this set of technologies are CDMA2000 1xEV-DO (data only) and 1xEV-DV (data/voice). Recent versions of EV-DO and EV-DV support 3.1 Mbit/s downstream and 1.8 Mbit/s upstream theo-retical data rates. Real-world rates are about half that speed.

The upgrade path from a 2G CDMA IS-95 network to a 3G CDMA2000 network is perhaps like the path from GSM to GPRS or EDGE. Certainly, it can require similar attention to handset, backhaul, and core network upgrades. But some industry experts (and many CDMA advocates) have argued that the base station upgrade requirements CDMA IS-95 to CDMA2000 are easer and cheaper than those for GSM to GPRS/EDGE.

It should be noted that, particularly in the case of GSM systems, the 2.5G systems are merely way stations on the migration path to full 3G systems (such as W-CDMA), which are discussed in detail in the next chapter.

2.2.2.2 Broadband WLAN Technologies

To date, the most common broadband wireless WLAN standard is Wi-Fi (IEEE 802.11). The 802.11 family currently includes six over-the-air modulation techniques that all use the

CHAPTER 2


24

same protocol. The most popular (and prolific) techniques are those defined by the ‘a’, ‘b’, and ‘g’ variations to the original standard. Wi-Fi uses different bands of radio spectrum. Perhaps the most optimal use of Wi-Fi is to transform and extend the capacity of standard wired Ethernet networks in public areas like meeting rooms, training classrooms and large auditoriums. In many places, of course, it has been adapted to provide local-area “hotspot” internet access. Wi-Fi generally provides a maxi-mum of 54 Mbit/s data transfer to a maximum range of about 50 meters. The extension of Wi-Fi signal strength has emerged as a niche development market in its own right, fostering the growth of companies like AirMagnet (used in Antarctica to extend Wi-Fi signals), among others.

In many North American and European countries, one can find WLANs in public cafes, airports, university cam-puses, conference venues and other central, high-traffic loca-tions. This is also slowly emerging in non-OECD countries. For example, China’s largest shopping centre, the Super Brand Mall in Shanghai, features the most extensive public WLAN system in Shanghai. There are 67 access points, covering six of the building’s 10 floors.

WLANs not only draw crowds, they also bolster the pro-gressive, technological image that organizations, businesses and institutes often seek to create. Moreover, WLANs contribute to a sense of community (particularly in academic settings) and provide new opportunities for generating access-related revenues from people who are simply “passing through.” For example, Switzerland offers a hot-spot service available in every major city, allowing people in transit to benefit (at a small fee, of course) from wireless broadband connectivity wherever available network signals are detectable.

2.2.2.3 The Advent of WiMAX

Unlike Wi-Fi, which began life as a short-range network-ing technology, WiMAX is an infra struc ture technology, on the order of DSL or cable modem technology. It can provide the same functionality as a router providing the backbone access to a location. Individuals may connect to a WiMAX modem via a wired Ethernet or Wi-Fi connection.

Based on a technique called orthogonal frequency division multiplexing (OFDM), WiMAX is more bandwidth-efficient than 3G technology or Wi-Fi. It supports high-throughput broadband connections over long distances. WiMAX can be used for a number of applications, including “last mile” broad-band connections, hotspots, cellular backhaul, and high-speed enterprise connectivity for businesses.

The prospect that mobile users may connect to WiMAX “hot zones” directly is still in development, and mobile access to WiMAX networks is not yet deployed. Meanwhile, genuine, standards-compliant hardware has yet to appear, even for fixed WiMAX installations.4 Figure 2.1 illustrates the growth forecast for WiMAX equipment sales through 2009. There is currently very little in the way of a track record by which to gauge the cost-effectiveness of WiMAX as a technology.

Even the potential of WiMAX, however, raises interesting regulatory questions. For example, how will the technology be received in countries that have already auctioned off expensive

3G licences? Governments might not have as strong an incen-tive to promote WiMAX, particularly if it could enable compe-tition with 3G services.5

Leaving aside the issues stemming from unlicensed usage of spectrum for WiMAX (for example, potential interfer-ence), the growth of WiMax could depend on market forces. It is not yet clear how much supply there will be from equip-ment manufacturers, nor is there a full picture of the level of demand from consumers. If consumers want to use their PCs, and these PCs are unable to leverage 3G spectrum to achieve broadband connectivity in their neighbourhoods, WiMax might be able to fill in the gap for ubiquitous wireless internet access. On the other hand, the market model could simply be scattered, shorter-range Wi-Fi hotspots, in tandem with mobile data services.

Some transitional economies are moving faster than others in providing wireless broadband connectivity. Considerable pre-liminary deployment activity is under way throughout Africa, Asia and the Arab World. For instance, the Malian telephone company Ikatel, a subsidiary of France Telecom, has contracted a vendor to supply and deploy a WiMAX network in the capi-tal city of Bamako.6 Meanwhile, in Saudi Arabia, broadband wireless technology has been deployed to enhance data transfer rates in the kingdom’s urban areas. Asian mar-kets have been particularly active; repre senta tives from China, Japan and Korea (Rep.) have reached an agreement to jointly develop future 3G technologies. Large-scale interna-tional manufacturers are also doing their part. Intel, for instance, had announced its plans for WiMAX trials in Malaysia, the Philippines and Thailand before the end of 2005, and in Indonesia and Vietnam before the end of 2006.7

2.2.3 Non-Terrestrial Broadband Platforms

Non-terrestrial systems (mostly satellite systems in the current marketplace) are generally regarded as complemen-tary to terrestrial broadband networks – and particularly useful (or necessary) in remote areas where no terrestrial infra struc-ture exists. This view is reinforced by the high costs, limited bandwidth and longer delays generally associated with satel-lite systems. Still, the number of subscribers for internet serv-ices provided by satellite operators is increasing. Aside from TV distribution services and telephone connections, satellite capacity is primarily used to connect internet service providers (ISPs) to internet backbone facilities. It is less commonly used for connecting individual users to an ISP, although this is cer-tainly a growing market in rural and underserved areas.

Whether broadband internet access via satellite is the right choice depends on the costs and performance of the alternatives. The cost per megabyte of satellite access is decreasing, albeit slowly. While the cost of operating the large, low earth orbit (LEO) constellations limits their margins, regional geostation-ary systems should be able to reduce their costs significantly. Prices on the order of USD 0.10 (10 cents) per megabyte have been predicted. The open standard for digital video broadcast-ing, with a return channel via satellite (DVB-RCS) is increas-ing its market share and is expected to benefit both users and

CHAPTER 2 25


industry through lower costs, customer choice and equipment inter op erability. The market is, however, still dominated by a few proprietary vendor solutions.

A satellite terminal consist of two parts:1 An outdoor unit consisting of a transceiver and an antenna

that is placed in direct line of sight to the satellite, and2 An indoor unit that provides the interface between the

transceiver and the end user’s communications system (a computer or a local area network).

The aperture of the antenna decreases at higher frequen-cies, due to the reduction in parabolic antenna beam-width at higher frequencies. Due to this, terminals operating at higher frequencies are called Very Small Aperture Terminals or VSATs.

Most satellite systems for television broadcast and broad-band data communications in operation today use portions of the C-band and the Ku-band of spectrum. In the C-band,

antennas are typically 2-4 metres in diameter, while in the Ku-band they can be smaller than 1 metre, which also makes it easier to direct the antenna. The lower-frequency parts of the spectrum that the C-band and Ku-bands represent cannot accommodate the data rates and traffic volumes demanded. This has forced commercial satellite system operators to con-sider the Ka-band and V-band as well. But these higher fre-quencies present other challenges, including rain attenuation, fading and signal scattering.

Independently of the system type, non-terrestrial wireless broadband network topologies fall into the following catego-ries:

– Bent pipe star topology – This is characterized by a large gateway earth station that transmits one or more high-data-rate, forward-link broadcasts to a large number of small user terminals. These broadcasts contain address information that allows each user terminal to select those

Figure 2.1: WiMAX Growth, Forecast Sales, 2004-5

Source: Sky Light Research, Sept. 29, 2005.

Worldwide Forecasted Sales of WiMAX Equipment2005 $23 million2006 $207 million2007 $566 million2008 $1.1 billion2009 $1.5 billion

Box 2.1: Wi-Fi Beyond Hotspots

The explosive growth of Wi-Fi hotspots that provide wireless local connections in business complexes, homes, and public spaces such as coffee shops and airport lounges has made Wi-Fi a household word. The IEEE 802.11 family of standards was de-signed by the data networking community for indoor, short-range, nomadic (can support walking speeds) uses. Wi-Fi has weak handoff capability between access points. But the vision of Wi-Fi networks simply as a replacement technology to the physical cables in the home or office LAN now seems constrained.

The amazing growth of Wi-Fi has driven economies of scale, such that the price for access points and end-user systems can be as low as USD 50 or even less. And this attractive pricing, along with the spectrum licence exemptions that many countries offer for Wi-Fi use, has led researchers to explore ways that Wi-Fi chipsets could be used in other networking environments.

A number of university research projects have been exploring modifications to 802.11 systems that would make them better suited to long distance, point-to-point backhaul networking. This includes work at the Indian Institute of Technology campuses in Kanpur and Chennai, as well as at the University of California, Berkeley in the United States. These projects have attempted long-range (tens of kilometres) point-to-point backhaul hops using 802.11 systems.

The principle problem with the 802.11 standard, in this context, lies with the way multiple radios on the network contend for transmission capacity. In technical jargon, the MAC layer uses a Carrier Sense Multiple Access (CSMA) protocol, which is ill-suited for wide-area networks, although it is well-suited for local area networks. Research projects have developed new Wi-Fi protocols better suited for long-distance, point-to-point networking. What may be emerging is a “rural network” extension to the 802.11 family that could directly compete with WiMAX.

More likely, however, is that Wi-Fi networks will complement WiMAX (and related networks). Wi-Fi hotspots can provide nomadic broadband for the “last mile” (or “last metre”) distribution networks, while WiMAX or other technologies can take over for backhaul and trunking.

CHAPTER 2


26

transmissions intended for it. In the return direction, the remote user terminals transmit in bursts at low-to-medium data rates to the gateway.

– Bent pipe point-to-point – This topology calls for a dedicated duplex connection, set up between a large gate-way earth station and a single user terminal.

– On-Board Processor (OBP) Switching – In this topol-ogy, the satellite rather than the gateway is the central node in a star network. The satellite is connected to the gate-way by one or more high-data-rate trunks. The on-board processor de-multiplexes the uplink trunk into several downlinks for different geographical areas, usually deter-mined by the footprint pattern. The forward downlinks contain messages for large numbers of user terminals, and the destinations are identified by message headers. In the return channel, the uplink transmissions from user termi-nals in one or more cells are multiplexed onto a downlink trunk to the gateway.

The benefits of the “bent pipe” versus OBP are under dis-cussion. In the bent-pipe system, the received signal is retrans-mitted without processing. While this scheme is less complex and has performed well in the first phase of internet develop-ment via satellite, OBP promises better bandwidth efficiency

and true mesh connectivity. Several companies are conducting OBP trials8.

2.3 Conclusion

Broadband networks are no longer a dream for the future. Increasingly, they are at the centre of ICT development today. The growth of converged, “triple play” offerings is coming to dominate business and regulatory developments. This is par-ticularly true in developed countries, but it is also a powerful factor in developing ones. A full range of broadband technolo-gies can and are being deployed in rural and underserved areas of developing countries. Increasingly, regulatory conversations around the world are laced with the fundamental question of how can we ensure that all communities participate in, and benefit from, the deployment of broadband capabilities?

The remainder of this edition of Trends seeks to answer that question, and others that face regulators around the world who are working to recalibrate their regulatory frameworks to address broadband issues. The next chapter aims to help regulators understand the technical dimensions of broadband technologies and begin to discern how they will progress and evolve, globally and in their own markets.

1 Chapter 3 will discuss the potential for stratospheric broadband platforms, which rely on sub-space aircraft such as airships to relay transmissions to and from the ground.

2 The DSL family is often denoted by the abbreviation xDSL, in which the x denotes that there are various forms of the technology. Those forms are explored in more detail in Chapter 3.

3 See: //http://www.cablemodem.com/specifications/re information about DOCSIS.4 Gruman, Galen, Wireless Broadband’s long and winding road: Truly pervasive, high-speed mobile data services won’t happen overnight. Info World. Vol. 27, issue

39. September 26, 2005. 5 Gruman, Galen, “Wireless Broadband: The Long and Winding Road”, Infoworld, September 29, 2005. Link: http://www.infoworld.com/pdf/special_report/2005/

39SRwamobile.pdf. 6 “Ikatel to launch Wi-MAX”, Telegeography’s Commsupdate, November 29, 2005. Link: http://www.telegeography.com/cu/article.php?article_id=10078. 7 “Intel to Trial Wi-MAX in Indonesia, Malaysia, Philippines, Thailand, Vietnam”, PricewaterhouseCoopers Telecom Direct News, September 26, 2005.8 These include Astrolink, SpaceWay and EuroSkyWay.

27


CHAPTER 3

3 UNDERSTANDING BROADBAND TECHNOLOGIES

Each wave of technological development offers new promise in the battle to bridge the Digital Divide. Most new technologies are cheaper to deploy than legacy copper net-works, and at the same time, they can deliver a full range of ICT services, from voice to broadband applications and serv-ices. Many new technologies can also be deployed incremen-tally, even locally, rather than on the large scale of traditional tele com mu ni cation networks.

Not only does this make deployment more affordable, it opens the door to a whole new range of possible broadband providers that can drive demand for broadband services. The types of broadband providers enabled by new technological developments include regional or private network operators, small and micro entrepreneurs, as well as public institutions such as universities, schools, libraries, post offices, local gov-ernment offices, health facilities, and non-governmental organ-izations active in developing countries. This chapter identifies and analyzes the promising new technologies that can help promote broadband access in developing countries, particularly in rural and underserved areas.

3.1 How Broadband Networks Are Designed

First, it is useful to introduce two concepts that help explain different dimensions of networks: layered architecture and network topology. These terms are important because they help to understand the dynamics that determine incentives, viabil-ity and the potential for collaboration between entities that are engaged in broadband deployment. In order to help regulators make informed decisions about the array of technical options before them, the following sections break out the various aspects and layers of networks, providing definitions and tech-nical specifications.

There are also useful comparisons between different types of platforms, such as DSL, CATV, “broadband over power line” (BPL) solutions, and fibre optical networks. These form the basis of a “wire-line broadband roadmap,” outlined in Sec-tion 3.2, for regulators and policy-makers. That section also provides detailed explanations of the various link layers (point to point vs. point-to-multipoint access networks) and system types. Attention is also given to wireless broadband systems, as

well as options for non-terrestrial wireless broadband networks when wire-line solutions are not feasible or applicable.

3.1.1 Layered Network ArchitectureThe concept of layered network architecture divides a net-

work at any specific point into layers, each of which adds value to the physical medium of communication. A layered architec-ture based on open standards is useful for several purposes:

– Technical: to define physical and logical interfaces required to connect different subsystems.

– Commercial: to define the conditions under which a user or a provider of value-added services can get access to serv-ices provided by a specific provider at a specific level.

– Regulatory: to identify the value chains and define the roles of actors providing services in different layers.

It may be necessary to regulate in order to promote com-petition and to allow service providers to buy services they need in order to provide value added services at a higher level. Rights-of-way, spectrum licensing, access to essential resources and local loop unbundling, are all examples of areas in which such regulation may be needed. The following subsections explain the network layers in greater detail.

3.1.1.1 The Physical Layer

The physical layer identifies the transmission medium – such as radio spectrum or copper or fibre wires – and speci-fies the mechanical and electrical interfaces that connect to the medium for communication purposes. The most important media include:• Wireless spectrum: Existing spectrum regulations are often

too restrictive, from a public policy perspective, prevent-ing new, innovative actors from entering the market. The attempt to loosen these restrictions is clearly reflected in the increasing reliance on licence-free spectrum in the dynamic development of Wi-Fi (unlicensed usage is explored more fully in Chapter 5 on broadband spectrum management).

• Wire-line infra struc ture: This includes fibre, copper cable and coaxial cable. Wire and cable are essentially unlim-ited in supply and can be deployed and made available at reasonable cost. In some countries, the provisioning of

Authors: Michael Best, Georgia Tech; and Bjorn Pehrson, Swedish Royal Institute of Technology

CHAPTER 3


28

passive infra struc ture is an independent business, while in others it is an integrated part of vertically integrated operations. Wire-line regulation, too, can be more restric-tive than necessary to promote broadband access. Access to the physical medium is essential for entities that have particular choices of transmission system or that want to compete at the link level and above (for example, by introducing a new competitive transmission technology). Regulatory measures to promote access to both wire-line and wireless broadband infra struc ture are explored more fully in the next chapter on regulators’ role in promoting broadband.

3.1.1.2 The Link Layer

Access to link layer services is essential to service provid-ers (or private network operators) that seek to build their own networks without having to operate their own transmission systems. With the physical layer as a foundation, the link layer adds procedures for digital data transmission over the physical medium, either point-to-point or point-to-multipoint. The link layer includes access both to wireless spectrum and wire lines, as well as to error control (such as automatic repeat requests when check sums do not match), forward error correction (based on the inclusion of redundant information coded in a way that transmission errors can be corrected directly rather than via retransmission requests), etc. Different link-level technologies also have different properties that are important for users, such as capacity, performance, security or privacy.

The properties of link-level technologies are different and may be of different significance to operators, users and regulators wanting to strike a balance between producer and consumer interests. Links are implemented using transmission equipment, and the most important ones to be discussed here include:• The most commonly used Wide Area Network (WAN)

technologies;• Technologies for data access over legacy networks such as

digital subscriber lines (DSL) over the Public Switched Telephone Network (PSTN);

• Cable modems over cable TV networks;• Broadband over power line (BPL) transmission; and• A range of both wired and wireless local and metropoli-

tan area network (LAN/MAN) technologies, all of which use the Ethernet frame data format according to the IEEE 802.3 standard1.

An Ethernet frame is the unit of data that is transmitted between network points on an Ethernet network. Examples of such wired networks include Ethernet with data rates from 10 Mbit/s to 100 gigabits per second (IEEE802.3). Examples of wireless link level equipment include Wi-Fi (IEEE 802.11) and WiMAX (802.16).

3.1.1.3 The Network Layer

The network layer provides mechanisms for addressing and forwarding data. This chapter assumes that Internet Proto-col (IP) is used for this purpose. The network layer is imple-

mented by network elements such as routers interfacing to different link level technologies, link level switches and mul-tiplexers, etc, to connect network hosts (servers and terminals). This is the level at which all ISPs provide services.

3.1.1.4 The Transport Layer

The transport layer provides end-to-end connections between user applications in network hosts. The central trans-port protocols include the connection-oriented Transport Con-trol Protocol (TCP) and the User Datagram Protocol (UDP).

3.1.1.5 The Application Layer

In the application layer, the communication parts of a user application, such as email or file transfer, Web access or data-base access, are implemented.

3.1.2 Network TopologyThe network topology concept divides networks into

functional parts, including access, backbone and service net-works and traffic exchange points where different service providers exchange traffic. Each of these functional parts is composed of the layers briefly described in Section 3.1.1. The following subsections describe the functional parts of network topologies.

3.1.2.1 Access Networks

Access networks are the links between end users and the service providers’ networks, whether they are the first mile or last mile (or last metre). First mile refers to a topology in which the user or a local service provider – or perhaps even an apart-ment building company2 – owns the access network and con-nects to service providers using its own upstream links. A last mile access topology denotes the operator’s ownership of the access network.

Different link-level technologies used in access networks have different properties, different sets of strengths and weak-nesses and different value assessments by regulators, users, operators and network owners. Depending on the geographical context, the access network could be a LAN or a MAN. As will be discussed in a later section, the technical solutions might be different depending on who owns the access network. In open regulatory environments, it could be owned by anyone

– a service provider, an organization, a municipality, or a user agent or a neutral agent, such as a real estate owner or inde-pendent operator.

The physical layer infra struc ture available for access net-works includes the PSTN, cable TV networks, electrical power networks, radio spectrum, and increasingly, fibre to the neigh-bourhood, office or home (this is often known as FTTP, for

“fibre to the premises”). On the user premises, the connections to the residential or office gateway might employ dedicated wiring, legacy copper telephone wiring, power-line or wire-less facilities. The most common link level technology on user premises is a local area network (IEEE 802), wired Ethernet or wireless (Wi-Fi).

In an environment with more than one operator, local loop unbundling is an important first step towards an open

CHAPTER 3 29


market. But this option is often frustrated in practical terms by the fact that a dominant operator owns the infra struc ture and is being forced to lease it, essentially, to a competitor. A more developed market structure may involve independent ownership of the infra struc ture, with mechanisms for service providers to obtain direct access to users without intermediary gatekeepers in the way.

3.1.2.2 Backbones

Backbones consist mainly of long-haul links that ISPs can use to expand their service networks geographically, to get transit to the Internet and to connect to regional traffic exchange points. In some countries, an open backbone market has emerged from the existence of parallel or complementary fibre infra struc tures deployed by different owners, including telecom operators, power utility companies, railways, pipeline companies and even municipal or regional governments.

3.1.2.3 Service Networks

A service network contains the ISP’s servers, which offer Internet access to users via access networks, as well as transit to the next tier of ISPs and peering with neighbouring ISPs in the same tier.

3.1.2.4 Traffic Exchange Points

Traffic exchange points are used by operators to exchange traffic through peering directly between service networks rather than indirectly, via transit through their upstream providers. An Internet exchange point (IXP) consists, in its simplest form, of a link layer switch over which ISPs peer to exchange IP traf-fic. The exchange point improves network performance by keeping local traffic local and minimizing transit costs for the connected ISP. This is particularly important in areas where the backbone consists mainly of satellite links with long delays, high bandwidth prices and the remote ends on different con-tinents.

This is a particular problem in Africa, which functions as a large number of VSAT “islands” on the shores of other Internet backbone “continents.” One solution to this isolation is to inter con nect IXPs, even via satellite links if better alterna-tives are lacking. This would cut the number of satellite hops by making direct hops between IXPs rather than two transit hops (sometimes more), plus perhaps a few transcontinental and overseas passages depending on where the different transit links happen to terminate. The main drawback of distributing an IXP geographically by connecting local IXPs is the bundling of the switching function and the long-haul link, which could lead to unfair competition in situations where link capacity is expensive.

In communities that only have VSATs available as gate-ways, it makes sense to have a local IXP keeping local traffic local before sending it upstream via the satellite channel. This is also valid for remote local communities that can take advan-tage of limited alternative fibre facilities – for example, links used for metre monitoring in pipelines (water, oil, gas) or power lines. IXPs can thus start appearing in the local access networks, not just between service networks. The importance

of IXPs is explored more fully in the joint ITU-IDRC Report Via Africa: Creating local and regional IXPs to save money and band-width3.

3.2 A Wire-line Broadband Roadmap

To an increasing extent today, wherever possible, dedi-cated broadband networks are being deployed. Such networks are typically based on Ethernet over fibre – in backbones, to the curb, block or neighbourhood, and to the home or office. They can interface to any other link level technology to take advantage of existing infra struc ture.

At the same time, new technologies have been developed to provide broadband services over legacy networks, such as DSL over the PSTN, hybrid fibre/coaxial (HFC) for cable TV networks and broadband over power line (BPL).

The different link level technologies involved are dis-cussed and analyzed in this section, in order to provide:• A description of likely broadband upgrades based on exist-

ing infra struc tures; • A discussion of differences in quality of service and trans-

mission rates among the various solutions;• An indication of the kinds of infra struc ture investments

that are required to deploy each technology (as a proxy for cost); and

• A means of assessing the suitability of different network options to various socio-economic and geographical con-texts.

3.2.1 Upgrades to the PSTN

3.2.1.1 Dial-up and ISDN

Upgrading from the PSTN to ISDN requires a digital network to the user premises. That, in turn, entails investment in equipment, both at the central office and at the user end. If already installed, ISDN is still an alternative for Internet access in areas where more advanced services such as DSL, cable modem or fibre networks cannot be used. But if ISDN is not already in place, DSL appears to be a better investment because it facilitates cheaper and higher-quality broadband service. Unfortunately, ISDN is incompatible with some DSL systems over the same infra struc ture, so end users that have upgraded from PSTN to ISDN actually may have to downgrade again, before they can then upgrade to DSL.

Compared to dial-up service on the PSTN, ISDN is an improvement – from both a bandwidth and reliability point of view. Operators offer ISDN services including two channels as a basic rate offering and 24 or 30 channels (depending on basic PSTN type), as a primary rate offering. Available user equipment consists of simple routers interfacing with the ISDN modem, including 2 or 24/30 channels and an Ethernet interface for a local area network on the customer premises. The equipment supports automatic opening of new channels as needed when the traffic increases, thereby providing from 64 kbit/s up to 1.5 to 2 Mbit/s connections. It is also possible to connect different channels to different destinations.

CHAPTER 3


30

On the operator side, the ISP typically leases a primary rate connection (PRI) from the telecom operator. From the end user’s point of view, the usage includes the call fee to the telecom operator for each channel while connected, and the ISP fee, which may be flat or connection-time based. Even in the best of circumstances, with 56-64 kbit/s maximum bandwidth per connection at voice tariffs, neither dial-up nor ISDN systems are able to offer competitive broadband services.

3.2.1.2 Digital Subscriber Line (xDSL)

ADSL

Asymmetric Digital Subscriber Line (ADSL) is the most wide-spread DSL technology. The data channels use one frequency band for a low-speed upstream channel (25 KHz to 138 KHz) and another for a high-speed downstream channel (139 KHz to 1.1 MHz). Data transmission speeds vary, based mainly on the distance between the subscriber and the central office. Some users cannot be reached at all by ADSL, because the dis-tance to the central office is too great. In Denmark in 2004, for example, about 5 per cent of households could not be reached by any ADSL services, and only 70 per cent of the population could access a 2 Mbit/s connection. More recent ADSL stand-ards, such as ADSL2 and ADSL2+, promise improved capacity and coverage.

VDSL

Very high-rate Digital Subscriber Line (VDSL)) is similar to ADSL, except that it is optimized for shorter distances (300-1 500 metres). Existing systems offer bandwidth capacities of up to 52 Mbit/s by including more high-frequency bandwidth in the copper cables and by deploying more efficient modula-tion. To extend its range, VDSL requires deployment of a fibre optical backbone network to the curb, block or neighbourhood (street cabinet). It also needs a power supply at the street cab-inet, which is not required for service over the PSTN. This increases deployment costs significantly. VDSL also has other limitations, including interference from ADSL and AM radio services. VDSL2, a standard under development, promises to achieve bit rates of up to 100 Mbit/s.

Uni-DSL

One DSL for Universal Service (UDSL or UniDSL) is a new variant of DSL, integrating all earlier DSL variants. It promises aggregated bit rates of up to 200 Mbit/s, including 100 Mbit/s symmetrical connections. While Uni-DSL gives operators the flexibility to offer a range of connections, the higher data rates cannot be offered on the existing PSTN infra struc ture. Uni-DSL would require a fibre backbone infra struc ture and would use only the part of the existing subscriber line closest to the user premises.

3.2.2 Upgrading CATV networks

A Hybrid Fibre/Coaxial (HFC) network combines a con-ventional coaxial cable TV network with fibre optic cables installed between the head end and the curb, block or neigh-bourhood (interfaced by converters). An HFC network may carry a variety of signal types, including analogue TV, digital TV, telephone, and data. It increases the competitiveness of

cable operators in a manner similar to the way that Passive Opti-cal Network (PON) upgrades reinforce the telephone industry (PON is discussed in greater detail in section 3.2.4.2).

3.2.3 Broadband over Power Line (BPL)

Power line communication systems that use the existing electrical power grid as a local loop for delivery of broadband services are often referred to as Broadband over Power Line (BPL). The typical power grid comprises generators, high-volt-age lines (155-765 kilovolt or kV), substations, medium-volt-age lines (1-40 kV), transformers and low-voltage lines (up to 400 V). High-voltage lines are unsuitable for BPL since there are too many electromagnetic disturbances (noise).

Several organizations are working on standards to ensure coexistence and inter op erability between technologies, as well as compliance with electro-magnetic compatibility (EMC). The Institute of Electrical and Electronics Engineers, Inc. (IEEE) has started work towards a “Standard for Broadband over Power Line Hardware” (P1675),4 which is intended to provide electric utilities with a comprehensive standard for installing the required communication hardware on distribu-tion lines. The standard is targeted for completion in mid-2006. There are also working groups within the Special International Committee for Radio-Electric Disturbances (CISPR),5 which has produced relevant directives including EN55022 (Euro-pean) and CISPR22 (international).

In Europe, standards include the low-power voltage 240-volt usage and frequencies from 30 kHz to 150 kHz. In North America, corresponding standards include the 120-volt grid and set of frequencies above 150 kHz, as well. Power utility companies often use frequencies below 490 kHz for their own telemetry and equipment control purposes. BPL uses medium-voltage power distribution lines (access BPL) and low-voltage in-house wiring (in-house BPL).

Access BPL uses modems and couplers, which are induc-tive injectors wrapped around the power lines. Typically, a fibre optic network connection from an ISP is terminated in an opto-electric converter and connected to a BPL modem at the util-ity substation, where the high-voltage lines are transformed to medium-voltage distribution networks. The traffic is fed into, and extracted from, the distribution lines via couplers.

The carrier supporting the communications signals can share the same line with the electrical signals because they operate at different frequencies. This is known as Frequency Division Multiplexing (FDM) of telecom and electrical power, with the BPL signal using frequencies between 2 MHz and 80 MHz.

Repeaters amplifying the signal and regenerating data must be installed about every 300 metres between the power station and the customer’s premises. Signals are terminated in a device just before the transformation to low-voltage lines (110/220 V) used inside the premises. BPL internal wiring facilitates home networking by enabling devices plugged into wall outlets in a building to communicate with each other over the existing wiring. One formal industrial standard that serves as an indus-try reference point is HomePlug (www.homeplug.org), which

CHAPTER 3 31


offers specifications to operate in the frequency range of 4.5-21 MHz. Some systems in Europe operate at 10-30 MHz. In-house BPL and access BPL are not dependent on each other, so either system can be employed with other technologies. Examples of Access BPL system manufacturers include:• Ilevo [www.ilevo.com] – offers products providing 200

Mbit/s and 45 Mbit/s bandwidth. Each power outlet is an access point to the power line network. The Ilevo sys-tems include a head end connecting the power grid to an upstream ISP through any standard link-level technology, different types of repeaters and a modem at the customer premises. The frequency band used is 1-30 MHz.

• Amperion [www.amperion.com] – offers products that deliver Internet connectivity via a wireless link called Pow-erWi-Fi (IEEE 802.11b) to an Ethernet port, instead of via the in-house wiring. These systems operate over 3-35 kV medium-volt lines, and they provide up to 24 Mbit/s of throughput per injection point, depending on line quality and equipment spacing.

3.2.4 Fibre NetworksSo far, this chapter has discussed ways to extend exist-

ing networks or modify infra struc tures to provide broadband Internet access. This section takes another course: using fibre technology to deploy broadband networks offering data as well as voice and video services. In addition to a technical overview, this section will discuss strategies for deployment of fibre in developing countries – particularly in rural and underserved areas, where penetration of wired telephone networks is low and, in many cases, decreasing due to the popularity of mobile telephony.

The wireless networks in many developing countries are based on microwave backbones, which provide little support for broadband applications. Many rural and underserved areas lack any broadband communication infra struc ture. Neverthe-less, these regions do have users that demand broadband serv-ices. The lack of legacy infra struc ture can clearly be turned into a strength by leapfrogging over the need to build expen-sive, older-technology communication networks.

Instead, fibre deployment can be coordinated and even shared with other infra struc ture-dependent sectors, namely power utilities (on which many ICTs rely to operate), rail-ways, pipelines and roads. Many developing countries have developed such strengths in recent years, after building politi-cal awareness through national ICT policies and infra struc ture plans. The most striking examples include Laos, Rwanda and Tanzania. The availability of infra struc ture creates new oppor-tunities – but only if the regulatory environment allows entre-preneurs to take advantage of them.

3.2.4.1 The Physical Layer of a Fibre Network

An optical fibre is a hair-thin thread of glass that trans-ports light waves with very low diminution over long distances. Fibre is deployed in cables. Standard cables contain 24, 40 or 96 fibres. Cables can be deployed underground in conduits, under water as submarine cables or strung between poles or pylons. The cost of deploying fibre stems mainly from civil

engineering work involved. The marginal cost of adding more fibre cores in a cable is generally very low compared with costs of other types of infra struc ture.

Power utilities deploy fibre primarily for supervision, control and data acquisition (SCADA) functions of managing the power grid. But they are increasingly adding more fibre (at a very low marginal cost) to lease to other parities. In new installa-tions, utilities normally use a special ground wire with a fibre cable in the core (an optical power ground wire or “OPGW”). Or, where they already have power lines, they wrap fibre around the transmission lines, in a process known as “SkyWrap.” The-oretically, then, every power grid substation – including those in rural and underserved areas – can become a point of pres-ence for access to fibre.

Signalling over optical fibre is accomplished by lasers as transmitters and photo diodes as receivers. Standard data rates are 1 gigabit per second (gbit/s), 2.5 gbit/s, 10 gbit/s or 40 gbit/s in each stream. A 100 gbit/s prototype was presented at the 2005 European Conference on Optical Communication (ECOC). Wavelength division multiplexing makes it possible to have up to 96 parallel data streams in a single fibre. The maximum total capacity in a single fibre is currently in the 1-10 terabits per second (tbit/s) range.

3.2.4.2 Optical Networks

The physical layer of optical fibre consists of two sub-layers: (1) the passive fibre itself, without any signals; and (2) an active optical network. In the passive fibre network, passive optical splitters can be installed, creating a tree-structure infra-struc ture, similar to what is done with coaxial cable in cable TV networks. This infra struc ture is called a passive optical net-work (PON). It is used to implement point-to-multipoint links, providing “broadcast” transmissions to end-users. The term

“passive” in this context means that the transmission of signals from the central office to the customer premise equipment does not require an external power source. Instead, PONs use light waves for data transfer. The use of such trees to establish links is discussed further in subsection 3.2.4.3. below.

Active optical networks are point-to-point infra struc tures. They include active (powered) optical components that pro-vide routing, grooming and restoration of signals at the wave-length level, as well as wavelength-based services, by using wavelength division multiplexing.6 This creates a purely optical network infra struc ture, before involving the electrical or digital domains on the link level.

Not all fibre networks are designed using an optical net-work sub-layer in the physical layer. The main arguments for introducing an optical sub-layer on top of the passive fibre, before adding the digital communication link, are:

• Better utilization of the installed optical fibre base, including the possibility to resell capacity on a wavelength basis rather than reselling an entire fibre. This argument mainly con-cerns the already-installed fibre base, since the marginal

CHAPTER 3


32

cost of adding more fibre cores when deploying a new cable is low.

• Better network restoration capability after network failures, since optical networks can perform protection switching faster and more economically.

• Reduced costs for the entire communication system, since the distributed wavelength routing scheme decreases the cost for cross-connects and only wavelengths that inject or tap traffic at a node need an electrical network element at that node.

3.2.4.3 The Link Layer of Fibre Networks

The dominant wire-line link level technology in fibre access networks is Ethernet (IEEE 802.3). The 10/100 Mbit/s Ethernet capability is standard in all new computers, including laptops. A 1 gbit/s Ethernet capability is the standard interface in most networking components used in access networks.

Regarding backbones, 10 gbit/s Ethernet has been in oper-ation for quite some time in high-end network components. A few 40 gbit/s backbones are in operation, and 100 gbit/s back-bones soon will be deployed. In the backbone, the Synchro-nous Digital Hierarchy (SDH) and Synchronous Optical Network (SONET)7 standards still dominate, because of the availability of robust and reliable carrier-class equipment in those stand-ards. The considerably cheaper and less complex Ethernet technology is, however, making its way into the backbone, reducing both capital and operating expenditures. Moreover, Internet Exchange Points are now Ethernet-based. So the now 30-year-old Ethernet technology is finally expanding into all parts of the network topology.

Point-to-point access networks establish independent links between the user premises and service networks. They offer maximum flexibility to all stakeholders, regardless of who owns and operates the involved links. What can be termed the operator-neutral model has been developed in Sweden. It is neu-tral in the sense that the passive access network infra struc ture is often owned by housing companies, condominiums or tenant organizations and, in some cases, by municipalities.

In this model, access networks are connected to a shared access network backbone. Any service provider can then con-nect its network gateway and offer services using the access network. The access network backbone is designed so that individual users in a housing area can select service provid-ers independently of each other. Low-cost, standard Ethernet multiplexers or switches are used to aggregate links from user premises to the access network gateway, while preserving the provider selection.

The operator-neutral model suits rural and underserved areas well, since there are few traditional operators that are likely to see a profitable business there. Given an adequate reg-ulatory environment and appropriate technologies, local entre-preneurs that know the local market opportunities can provide local services and connect users to the network points of pres-ence (POPs) of the larger service providers.

Turning to point-to-multipoint networks, the IEEE 802.3ah Ethernet in the First Mile standard and the ITU-T G.984

GPON standard have made Ethernet the preferred protocol also among the traditional, vertically integrated telecom opera-tors. As discussed in Section 3.2.4.2, a passive optical network, or PON, is a point-to-multipoint technology similar to cable networks, but fibre-based. By introducing passive optical split-ters and couplers, a tree structure is created from an optical line terminator at the central office to optical network terminals (ONTs) at a number of customer premises. Downstream data is broadcast to the terminals, each of which looks for a match-ing address at the protocol transmission unit header. Upstream traffic is coordinated using a TDMA protocol, in which dedi-cated transmission “time slots” are granted to each terminal.

The main fibre can operate at 155 Mbit/s, 622 Mbit/s (Broadband PON or BPON managing up to 16 ONTs) and 1.25 gbit/s or 2.5 gbit/s (Gigabit PON or GPON managing up to 32 ONTs). Bandwidth allocated to each customer from this aggregate bandwidth can be static or dynamically assigned in order to support voice, data and video applications. The ter-minal can provide all the appropriate interfaces. A single fibre, meanwhile, can serve 16, 32, or more buildings through the use of passive devices to split the optical signal and PON pro-tocols to control the transmission of signals across the shared access facility.

From an economic point of view, PON saves on the cost of fibre and equipment at the central office/head end, compared to using point-to-point connections. Any savings should, how-ever, be weighed against the weaknesses of point-to-multipoint technologies, which are similar to those of broadband cable networks. For a regulator pushing an open regulatory regime, the lock-in effect of PON is less attractive, since the topology of the physical medium makes it impossible to separate users and thus prevents local loop unbundling on the physical and link levels. There is no unique path between the central office and a single user, because passive splitters are used to build the physical infra struc ture, as in a cable network. Active optical networks, on the other hand, use switches, so end users can be separated.

For the operator, meanwhile, a PON is more complex to plan, manage and upgrade than point-to-point links. For the user, the performance depends on traffic from other users (because of bandwidth sharing) unless traffic control of indi-vidual connections is introduced. The broadcasting technique also raises concerns about security and privacy. Policy-makers and regulators seeking to promote new fibre backbone deploy-ment should weigh carefully the costs and benefits of these two fibre network options in order to best meet the ICT develop-ment goals of their country.

3.2.4.4 Fibre Deployment in Developing Countries

Fibre deployment is taking off in developing countries. The technologies exist and are not that expensive. In many of the developing countries deploying fibre, universities are at the forefront, establishing national research and education net-works (NRENs). Bangladesh, India and Pakistan, for example, all have national fibre backbones. NREN examples include:

CHAPTER 3 33


• The Pakistan Education and Research Network (PERN) connects all public universities in Pakistan.

• Laos has a national fibre infra struc ture reaching all pro-vincial and district capitals. There is an Internet exchange point (IXP), to which all ISPs, except the incumbent, are connected. The universities are in the process of setting up an NREN, and they participate in an European Union funded regional academic backbone programme, dubbed TEIN2, that connects NRENs in the ASEAN countries.

• For a reported cost of USD 50 million, a Kenyan network provider has commissioned the installation of a 1,140 km optical fibre network by the end of 2006. The optical fibre technology will give mobile operators, ISPs and fixed-line operators a core network extending from the coastal city of Mombasa in the southeast to the country’s western border.

• In Rwanda, Terracom is in the process of deploying a fibre network to all schools and other priority groups. Rwanda has connected all of its ISPs at an (IXP). Meanwhile, aca-demic institutions are in the process of organizing an NREN.

• Both Tanzania and Mozambique have a mix of multiple fibre owners. Both have IXPs and are in the process of establishing NRENs.

• Malawi and Zambia rely on power utilities for fibre deployment. Both are in the process of deploying IXPs, and the universities are in the process of setting up NRENs in both countries.

• Bolivia has a national fibre backbone, an IXP and an expanding NREN.

3.3 Broadband Wireless Access (BWA) Networks

Wire-line broadband deployments, of course, are not the only options. A range of terrestrial wireless solutions are on offer or on the horizon. The development of BWA solutions has been marked both by significant technological progress and substantial “weeding out” of some market players. BWA solu-tions can be divided into four primary sub-categories, two of which are also often labelled as 3G or IMT-2000 technologies (W-CDMA and CDMA2000). The other two major families, WiMAX and 802.20, are growing in use and maturity in parallel with the development of their standards. While the 3G systems have evolved out of the mobile telephony sector, the other two technologies have emerged from the data networking sectors and often offer fixed wireless solutions (although they, too, are moving towards mobility).

This section explores the migration paths of the four families of broadband wireless access technologies, picking up from the 2G and 2.5 solutions explored in Chapter 2. The technologies in this chapter represent the immediate future of broadband development.

3.3.1 The GSM Migration PathThe 3G migration path from GSM, known as wideband

CDMA (W-CMDA), is an incremental upgrade from GSM’s 2.5G networks. W-CDMA networks, sometimes called UMTS, have been initially deployed in Europe and Japan. Standard W-CDMA systems can support up to 2 Mbit/s, while an enhanced version, called HSDPA, allows downlink rates of up to 14 Mbit/s by using higher modulation rates and other advanced techniques. With 2.5G base stations and available backhaul networks in place – and with spectrum secured – the upgrade

Table 3.1: Relative Functionalities of Broadband Wireless Access Technologies

Technology Used frequencies Supported data rates Cell radius Notes

EDGE 850 / 900 / 1 800 / 1 900 MHz Up to 384 kbit/s 1 km Deployed globally; extensive availability of

terminals

WCDMA 1 900 / 2 100 MHz Up to 2 Mbit/s 0.4-2 km Large-scale deployment; widely-supported by vendors

HSDPA 1 900 / 2 100 MHz Up to 14 Mbit/s 2 km Enhancement to WCDMA; limited deployment in Japan

1xRTT 450 / 850 / 950 / 1 800 / 1 900 / 2 500 MHz Up to 144 kbit/s Up to app. 50 km Widely used as fi xed wireless solution

1xEV-DO 2.3 GHz Up to 2.4 Mbit/s Up to 15 km Deployment mostly concentrated in North Asia; widely supported by vendors

UMTS-TDD 1.9 / 2 / 3.4-3.5 GHz Up to 7 Mbit/s 29 km Deployed in a few countries, e.g. New

Zealand, Australia, Portugal, etc.

WiFi 2.4 GHz Up to 11 Mbit/s 100 m Widely deployed globally; backed by major vendors with range of terminals available

WiMAX 3.5 GHz Up to 75 Mbit/s Up to 50 km with line of sight

Believed to be optimal solution for fi xed wireless access and for pushing broadband into rural areas

CHAPTER 3


34

Box 3.1: A Radio Transmission Primer

All wireless networks communicate via electromagnetic energy, which is described by wavelength or frequency. Since elec-tromagnetic energy traces a sine wave, the frequency specifies the number of humps (or troughs) per second of time and this unit is called the Hertz (or just Hz). Radio waves between 30 MHz (million Hz or megahertz) and 20 GHz (thousand million Hz or gigahertz) are usually used for data. Lower frequencies are used mostly for broadcast services such as FM and AM radio (though in some cases these are also used for data). Energy much above 20 GHz is not very suitable for data over long distances, as it is easily absorbed by water vapour in the atmosphere.

The frequency of electromagnetic energy tells a lot about its usefulness for communications. Relatively low frequency energy – AM radio broadcasting frequencies, for example – travels as ground waves, literally hugging the earth’s surface. This allows the radio waves to travel past the horizon dramatically, extending their distance. These waves are not easily absorbed or reflected by objects such as trees, buildings, and the like. Relatively low frequency energy also can be bounced off of the ionosphere, dramati-cally extending the distance travelled (see below).

By contrast, higher frequencies, in the spectrum known as “microwave,” generally require line-of-sight (LOS) transmission and cannot travel much past the horizon. The LOS requirement can be mitigated, as seen in some emerging 802.16 technologies, through sophisticated multi-path approaches, in which the signal reaches the receiver not necessarily in a straight line from the transmitter but instead reflecting off one or more objects, creating multiple paths from the sender to the receiver.

Given these transmission or “propagation” properties, it would seem that the lower the frequency, the easier and better the transmission. But in thinking about broadband digital data transmission, there is one major advantage of higher frequencies: the higher the frequency, the more data that can be transmitted. In fact, typically 1-4 bit/s of data can be encoded for each cycle-per-second of the radio wave. Note that this not the same as the data rate delivered to a user, because many of these bits are put to other purposes, such as signalling, error detection and correction. Broadband wireless networks are all situated in the microwave spectrum in order to take advantage of much higher signalling rates. But they all must do battle with the relative difficulties in signal loss and fading.

Figure 3.1: Comparative Prices for Mobile Data ServicesPrice per Mbit/s, for selected operators in selected countries, June 2004, in USD, at different monthly usage thresholds, compared with NTT DoCoMo’ s unlimited usage monthly price

Source: 3G Mobile, ITU research

Operator (country) Technology 1 Mbit/s 10 Mbit/s 100 Mbit/s 1 Gbit/sOrange (France) GPRS 1.46 1.22 1.22 1.22T-Mobile (Germany) GPRS 10.68 2.62 2.09 0.23TIM (Italy) GPRS 1.82 1.82 1.82 1.82Telefonica (Spain) GPRS 7.29 7.29 1.46 1.46Vodafone (UK) GPRS 13.71 3.86 2.15 2.15NTT Do CoMo (Japan) (package) W-CDMA 9.14 3.92 1.35 0.07NTT Do CoMo (fl at-rate) W-CDMA 35.00 3.50 0.35 0.04

CHAPTER 3 35


from 2.5 to 3G systems can be done incrementally and with relative ease. For instance, 3G services can be offered on a cell-by-cell basis and only in those areas with sufficient subscriber demand. In practice, however, 3G upgrades are done across an entire region in order to provide consistency in service levels for mobile customers and to allow broad marketing and sales plans.

3.3.2 The CDMA Migration Path

There is a family of 3G upgrades for CDMA networks that are called CDMA2000. This includes the CDMA2000 1x systems which support data rates up to 307 kbit/s. The CDMA Development Group (CDG) refers to all CDMA2000 subscriber numbers as “3G” users, but the real equivalent to 3G is in fact CDMA2000 1xEV (for Evolution), which is a

higher-speed version of 1x. Within this set of technologies are CDMA2000 1xEV-DO (data only) and 1xEV-DV (data/voice).

Recent versions of EV-DO and EV-DV support theoretical data rates of 3.1 Mbit/s downstream and 1.8 Mbit/s upstream. Real-world rates are about half that speed.

The performance and market adoption of 3G networks (CDMA2000 and W-CDMA) has, to date, been disappointing. This is partly because of high prices associated with network upgrades and the extraordinary amounts paid for 3G spectrum in some areas (mostly Europe). There also is a relative lack of suitable applications to motivate subscriber adoption and hand-set upgrades. But the 3G vision remains compelling: ubiqui-tous Internet access enjoyed while moving at vehicular speeds, with broadband bandwidths and latencies that are comparable to DSL or cable modem services.

Box 3.2: TD-SCDMA – A Chinese Standard

Time-Division Synchronous Code-Division Multiple-Access (TD-SCDMA) is a 3G standard created by the Chinese Academy of Telecommunications Technology, working with equipment vendors such as Siemens. Similar to the WiBRO initiative described below, TD-SCDMA offers an interesting illustration of a way that the public sector can collaborate with the private sector to create a local network standard. In fact, the Chinese have explicitly positioned TD-SCDMA as a way for the country to avoid dependence on “Western technologies.” And with China having more mobile subscribers than any other nation in the world, the country has a sufficient market size to support its own standard.

The TD-SCDMA standard is an evolution from the GSM standard in much the same way W-CDMA is. Field tests have shown the system to work at vehicular speeds and at a 21 Km distance from the base station. Data rates are published as ranging from 1.2 kbit/s up to 2 Mbit/s.

The fact that the published data rates are so broad is more than a curiosity, as technical trials have shown a disappointing delivered capacity. The government has called for an “intensive industry-wide effort” to deliver on the technology and to make it competitive.

Figure 3.2: Migration Paths from 2G to 3G

Figure 2 — From 2G to 3G

Evolution of mobile systems to 3G

TDMA

GSM GPRS

cdmaOne

2G First step into 3G 3G Phase 1 Evolved 3G

Voice and up to 14.4 kbit/s data 64 – 144 kbit/s 384 kbit/s – 2 Mbit/s 384 kbit/s – 10 Mbit/s

Actual user data rate 14 kbit/s 20 – 60 kbit/s 60 – 120 kbit/s Over 384 kbit/s

Time 2000/2001 2001/2002 2003

CDMA2000 1X CDMA2000 1XEV

WCDMA

EDGE

PDC

CHAPTER 3


36

3.3.3 The Evolution of WiMAX

A range of technologies fall under the WiMAX moniker, including those that conform to the emerging set of IEEE 802.16 standards. WiMAX systems promise to:

– Have very high capacities (up to 134.4 Mbit/s in a 28 MHz channel),

– Travel long distances (50 km or more),– Not require line-of-sight,– Work at vehicular speeds (under the 802.16e extension),– Enjoy high spectral efficiency (under the 802.16a/d exten-

sion), and– Be inexpensive (with base stations in the $10,000 range).

It sounds like a broadband wireless “dream come true,” except that not all of these extensions have worked yet in the real world. Moreover, not all of these desirable qualities can be enjoyed at the same time with the same network (for example, as of yet there is no cheap, efficient, high-capacity system that works at vehicular speeds). Still, WiMAX systems show great promise for the provision of broadband Internet access serv-ices, especially in remote areas or wherever fully ubiquitous access, at vehicular speeds with seamless handoff, is not a high priority.

The 802.16 standard and its extensions were still being finalized at the time of publication, but a number of vendors already had begun offering broadband metropolitan area net-work (MAN) technologies that will comport with the 802.16 standard as soon as it is stable. One example is the Canopy system from Motorola. Canopy transmissions can travel 50 km in a single hop, providing 10 Mbit/s of shared bandwidth.

Outdoor access points can list for USD 1,000, with customer premises equipment costing USD 500.

Samsung and LG Electronics of Korea (Rep.) have devel-oped a WiMAX-styled technology called WiBro (for “wireless broadband”), which is designed for the 2.3 GHz band. It offers 512-1,024 kbit/s per user, and allows users to travel at near-vehicular speeds (around 60 km per hour). The system has emerged with assistance from the government of Korea (Rep.), which was eager to see a locally produced technology and had promoted WiBro as the basis for the 802.16e mobile WiMAX standard. Other important stakeholders, however, were not supportive of WiBro as a standard-setting technology for tech-nical reasons. The 802.16e stand ardization process could have been weakened if two competing technologies were to emerge. But at this point, companies have agreed to converge on a shared single standard.

The European Telecommunications Standards Institute (ETSI) has also developed broadband metropolitan area net-work standards under the name HiperMAN. Like WiBro and other related technologies, these systems allow for long-dis-tance transmissions (over tens of kilometres) and high band-width (up to 280 Mbit/s at each base station). The WiMAX Forum has been working with the HiperMAN, WiBro, and IEEE 802.16 standards to try to ensure inter op erability among all of these various systems.

3.3.4 The New Kid on the Block: 802.20

WiMAX started as a fixed wireless technology and has since been evolving to support mobility under the .16e exten-sion. The IEEE’s 802.20 standard, however, originated explic-

Box 3.3: 802.16 Extensions in the Works

802.16a Works in 2-11 GHz range and supports mesh deployments.

802.16b Increases the amount of spectrum that can be used in 5-6 GHz range, and provides Quality of Service guarantees.

802.16c Works in a higher frequency range of 10 to 66 GHz.

802.16d Brings improvements to 802.16a; intended to supplant 802.16 and 802.16a.

802.16e Supports mobile devices.

Box 3.4: The Evolution of a “Southern” Solution

Broadband corDECT is an important, although incremental, improvement on the corDECT system designed at the Indian Institute of Technology, Madras. The initial DECT system, based on the European standard originally developed for cordless telephones, provided LOS connectivity at a maximum of 70 kbit/s. One significant strength of this system has been its relative low cost.

While the benefits of non-line-of-sight (NLOS) systems are palpable, there is no immediately available technique to develop an NLOS DECT-based system. This is because of the power restrictions on the spectrum used by DECT radios. Broadband cor-DECT is able to turn this “bug” into a feature by taking advantage of the line-of-sight requirement to enhance capacity. Through spatial reuse of spectrum, enhanced modulation levels, and the use of radio wave polarization techniques, Broadband corDECT delivers 256 kbit/s – and, ultimately, as much as 512 kbit/s – of dedicated bandwidth to each user. Costs are considered affordable at around USD 150 per subscriber, including base station and customer equipment costs.

Source: Midas Communications

CHAPTER 3 37


itly as a mobile broadband technology. The fact that its original design requirements included mobility should prove benefi-cial to the standard compared with others, which have had to add mobility on as an extension. Even so, the 802.16 family of WiMAX systems has a first-mover advantage and, at present, stronger industry support from major players.

Through its recent acquisition of Flarion Technologies and their Flash-OFDM technology, Qualcomm has just dem-onstrated its support of the emerging 802.20 standard, which has not yet been finalized.

3.3.5 OFDM Technology

Orthogonal frequency-division multiplexing (OFDM) is emerg-ing as a leading technology for very-high-bandwidth wireless connectivity. As the speed of wireless services increases, it increases the need for more and more radio spectrum, which is expensive and hard to acquire. In turn, this increases the importance of spectral efficiency, which refers to the number of bits that can be encoded into a single radio cycle. OFDM-based technologies, including WiMAX, enjoy spectral efficiencies of around 4 bit/s per hertz. This compares favourably with the

Box 3.5: Mesh Networks

A mesh network inter con nects communication nodes to pass messages directly between each node, without a dedicated intermediary (i.e., a server). The communication nodes used in a mesh network can be of any type – from handheld sensors to desktop computers to Wi-Fi routers.

The difference between mesh networking between clients (also known as multipoint-to-multipoint or peer-to-peer network-ing) and traditional networking is that in the latter, a dedicated router is required to pass all messages from one client to another. Mesh networks still frequently make use of dedicated routers, but they also can be configured without them. The arrangement (topology) of client communication devices (nodes) is also very flexible. The key is that each node is capable of talking directly to its neighbours (within range), whether that node is a client (computers, sensors, kiosks) or infra struc ture (routers, access points, gateways).

There are several reasons to choose a mesh network over a traditional one, including greater scalability, extensibility, resiliency, and physical infra struc ture advantages. Scalability refers to the flexibility of a mesh network to choose the path of a message being routed. When an additional node is added to a traditional network, it has to be able to communicate with a dedicated router, which either increases the load on an existing router or requires installation of a new one. On a mesh network, however, the additional node not only consumes routing capacity but also adds its own capacity to the network. This means that the routing capacity and aggregate throughput within the network grows as the network grows, rather than starting at some limit and decreasing, which is what happens in a traditional network.

Extensibility is a very important advantage of mesh networks – particularly wireless ones. Extensibility stems from the fact that any node can have its message hop from one point to another via a peer node. In a traditional (point-to-multipoint) wireless network, being out of range of a dedicated access point means a loss of connectivity. But in a mesh network, connectivity can be maintained as long as there is another client node within range that can be used as a bridge to eventually reach the access point or destination. There is no fundamental limit to the number of hops a message can make before reaching its destination, but im-ple men ta tion details mean that, as the network extends farther and farther outward, it will eventually be beneficial to add a new dedicated access point.

Another advantage of mesh networks is their tolerance of failures. Typically, in a mesh network there will be more than one path available for data to take. A traditional network only offers a single path. If, in a traditional network, a router suffers a failure, the nodes connected to it lose all connectivity. Another benefit of the multi-path character of mesh networks is that the “network diameter” – the minimum number of hops separating any two communicating nodes – can be smaller than in a traditional network. This can result in reduced latency on the network.

There are drawbacks to mesh networking. First and foremost is that each communicating node needs to be willing and able to route traffic. This activity requires computational and electrical power and can slow down the appliance or unduly drain its battery. In addition, mesh networks require additional preparation and setup of each client and may be harder to maintain than a more centralized, traditional network.

One market in which mesh networking could make significant inroads is rural access. A mesh network provider based in a metropolitan area could offer services in remote areas by “piggybacking” connectivity over a series of subscribers in the direction of the end-user. Data traffic on the outer edge of the network – in a remote village, perhaps – would only need a wireless connection strong enough to reach the next, closer subscriber to the metropolitan area. This second subscriber would then pass the traffic to another, closer subscriber and the process would continue until the traffic reached the backbone Internet connection. By using all subscribers as transit points, the mesh network can quickly reach distant areas with relative ease.1

1 ITU Internet Reports: The Portable Internet, 2004, see http://www.itu.int/publications/bookshop

CHAPTER 3


38

802.11d standard, for instance, which has a spectral efficiency of less than 2 bit/s per hertz.

OFDM works by segmenting available spectrum by fre-quency and then carrying a portion of the user’s data on each of these frequencies. Each frequency is unique and does not overlap, making it orthogonal to the next. This ensures that there is no interference between the various tones. This tech-nique, along with other sophisticated improvements in digital signal processing, has produced an efficient and speedy net-work technology.

Flash-OFDM can deliver capacity of about 1 Mbit/s down-stream and 500 kbit/s upstream while stationary. Available bandwidth is diminished, however, when moving at vehicular speeds. One significant strength of the Flash-OFDM system is its spectral efficiency, which is about 4 bit/s per hertz (a similar efficiency is planned for the 802.16a OFDM extension). This is a great advantage when spectrum is scarce or expensive. In some rural and underserved areas, where microwave radio bands are relatively underutilized, this advantage is not as com-pelling.

Another early system intended to track the emerging 802.20 standard is iBurst, which is based on radio technologies (and in particular smart antenna designs) produced by U.S.-based ArrayComm. Early deployments of this mobile broad-band system have gone up in Australia and South Africa.8

3.3.6 Integrating Heterogeneous BWA NetworksWhat this BWA roadmap makes clear is that there are a

number of complementary and often-competing standards, standard-setting institutions, proprietary offerings and vendors. Considerable effort is being put to inter op erability and merging these various systems. But it still is likely that many enterprises will end up with a heterogeneous collection of wireless net-works. Moreover, the functionality of many of these systems is converging (See Box 3.6).

Ongoing research is focusing on ways to integrate the var-ious network technologies. For instance, the Third Generation Partnership Project (3GPP) has been studying systems to inter-con nect 3G systems with WiMAX or Wi-Fi networks. Issues have included hand-off as well as authentication, authoriza-tion and accounting (AAA) and other considerations. As the capabilities and purposes of fixed wireless and 3G wireless networks converge and approach the performance of wire-line networks, perhaps it will become easier to make a strategic choice between these families of technologies.

3.3.7 Civil Engineering Costs for Broadband Wireless

As long as Moore’s law holds true, there will be a con-stant, exponential growth in the performance and affordability of computer technologies. Regrettably, however, Moore’s law has yet to apply to the performance or cost of steel, cement, or labour. As a result, the “civil engineering” costs for outdoor radio networks – the costs of radio towers, cement foundations, and related requirements – are beginning to outdistance the cost of the solid state radio equipment.

Consider the cost of radio masts or towers, which are essential for fixed wireless and 3G radio networks. Without towers, it is impossible to position signals above foliage and buildings that might obstruct them, to extend the horizon, or to meet other transmission requirements. In most major cities, the skyline is littered with these towers. They range from masts on the order of a metre in height, positioned on the roof of a customer’s premises, to 10-metre cellular network towers and 100-metre (or higher) towers populated with high capacity point-to-point microwave antennas and, perhaps, broadcast TV or radio facilities.

The cost and effort required to put up a radio tower can be influenced by several factors, including local climate (whether there are high winds or frequent icing) and the intended load-

Figure 3.3: A mesh network topology versus a traditional network

CHAPTER 3 39


ing of the tower (for example, one small antenna or many large ones). Towers are designed with these factors in mind, and a choice must be made between three principal support strate-gies:

• Free-standing,

• Bracketed against an existing structure, or

• Supported with guy wires.

By far the most important factor in determining the cost and complexity of a tower design is its height. The height requirement largely dictates the type of tower that is needed, the foundation strength and the requirements for guy wires, among other things. Grounding is important, particularly in areas that experience seasonal electric storms. Lightening strikes can spell the death of antennas and radio electronics. In order to protect the equipment, lightening attractors can be affixed at the top of the tower. The principle of a lighten-ing attractor is simple: it should be higher than the equipment being protected and it should offer less resistance to the flow

of electricity than “competing” local options (such as the radio equipment). Fixing an attractor above the radio and antenna is easy enough. But ensuring low electrical resistance requires careful attention to grounding. Especially in rural and under-served areas – where, for instance, a relay base station may be positioned – industrial grounds are not available and it is nec-essary to dig an earth pit to ground the arrestor. This activity can add substantial expenses to the costs.

Figure 3.5 shows empirical data for the tower and labour costs in the United States and Ghana, as well as grounding costs in the United States. A simple regression performed on these data points yields the linear cost models that are depicted visually below. Results indicate that:

• In Ghana a 3-metre tower with installation and grounding costs just under USD 220, and an additional USD 270 for each additional metre.

Box 3.6: Convergence of WLAN and 3G Networks

There is today no single, optimal broadband technology. Each major family of technologies features strengths and weaknesses. Figure 3.4 provides a preliminary decision matrix that can help develop strategic choices among the technology types. Notwith-standing these technology differences, there is significant convergence among the various offerings towards a unified set of design characteristics for an ideal broadband wireless network. These characteristics would include:

– High bit-rates in an all-IP environment, including IP Version 6 (IPv6) support

– End-to-end QoS

– Multimedia support

– Mobility at automobile and train speeds

– Seamless session management

– Security, security, security

– Support for flexible and dynamic spectrum and interference management (including software-defined radios)

– Advanced authentication, authorization, and accounting protocols.

These goals are more than a pipe-dream. The fixed wireless and 3G families of networks are converging on each other, and beginning to converge on this network design “wish list.”

Figure 3.4: The Convergence of WLAN and 3G Technologies – A Matrix

Current WLAN Strengths• Relatively inexpensive

• Data ready and high-bit rates

Current 2G strengths• Ubiquitous

• Mobility

• QoS guarantees

Emerging WLAN Strengths• Mobility (e.g. mobile WiMAX 802.16e)

• QoS (e.g. WiFi 802.11e)

• VoIPEmerging 3G Strengths

• Multimedia support

• High data rates (e.g. W-CDMA)

CHAPTER 3


40

Figure 3.5: Empirical Data for Tower, Installation, and Grounding Costs

Figure 3.6: Linear Model of Tower, Installation, and Grounding costs (USA/Ghana)

CHAPTER 3 41


• In the United States a 3-metre tower with installation and grounding costs nearly USD 1,360, and almost an addi-tional USD 770 for each additional metre.

Of course, these models are just estimates to help to clar-ify and guide thinking. As a point of comparison, for instance, Motorola Canopy outdoor access points can list for approxi-mately USD 1,000, while their subscriber modules can be under USD 500. Clearly any cost analysis of wireless broad-band equipment should consider such civil engineering costs, since these can dominate costs for radio equipment.

3.3.8 Stratospheric platformsStratospheric platforms operate within the stratosphere

portion of the atmosphere, under aerodynamic conditions. High- and low-altitude platform stations (HAPS or LAPS) offer the coverage benefits of satellites at costs close to fixed infra-struc ture.9 A HAPS platform may be an airship soaring in the stratosphere at an altitude about 20 km – well above normal aircraft but below orbiting satellites. There are no commercial HAPS services yet, but active research and development efforts are under way. LAPS services may employ an airship kept sta-tionary at an altitude of about 3 km. Commercial LAPS serv-ices are available, especially targeting temporary needs such as emergency response situations or sports events.

The key markets for LAPS and HAPS will likely be rural and developing areas that are underserved by traditional infra-struc ture. But they could also play a key role in conjunction with newer wireless technologies such as WiMAX. Since LAPS and HAPS are potential platforms for delivering connectivity for a range of wireless systems, their radio equipment could make use of the most current technologies to provide fast line-

of-sight connectivity. Since the line-of-sight requirement could be met for many applications, the frequencies and correspond-ing transmission speeds could be much higher.10

3.4 A Decision Framework for Broadband

Facilities-based competition is a desirable element for any nation’s broadband market, but it simply may not exist in all (or any) parts of all countries. For instance, some high-income metropolitan areas may have multiple wire-line broadband pro-viders (a cable TV system operator and a DSL operator) along with multiple mobile operators. Some areas may be served by smaller, local entrepreneurs or even municipalities offering WLAN hotspots. Each of these would compete to provide the most cutting-edge data solutions. But in some low-income areas – and especially in rural and sparsely populated areas

– this level of competition may not be commercially viable.

Initially, policy-makers and regulators may succeed in inducing only one broadband provider to enter rural or sparsely populated markets, while leaving the door open for other entrants once demand for broadband services has been established. In these environments, therefore, strategic choices may have to be made between various network families. In other words, some contexts may commercially support facili-ties-based competition (or, said another way, “complementary” networks), while other regions may face a choice of one net-work over the others.

As mentioned in chapter 2, there are currently three dom-inant technology families for the deployment of broadband Internet connectivity:

Figure 3.7: Radio Tower Types and HeightsA simple radio pole, perhaps just a few meters high (left); a free-standing radio tower used for a cellular phone network, perhaps 10 meters high (centre); a guyed radio tower holding various antenna, perhaps 100 meters high (right).

CHAPTER 3


42

1 Broadband wireline networks – DSL, cable TV and fibre solutions,

2 Broadband wireless access – 3G mobile and WiMAX, and

3 Non-terrestrial wireless options – VSAT and other sat-ellite and stratospheric platforms.

Each of these solutions is in play today, and each has niches in which it is the most appropriate. In this section we will consider various contexts – geographic, economic, demo-graphic, and public policy environments – and how they might determine the viability of each family of network solutions. The starting point of this analysis is that there is no significant environment on the planet today in which broadband Internet does not make commercial, social, and institutional sense.

These families of technologies under discussion in this chapter mostly describe the edge network, that is, the “last mile” component of the overall network that connects the base sta-tion or central office to the subscriber’s premises. The backhaul network is the facility that connects the various base stations together. And finally, this traffic is aggregated by the backhaul network and passed on to the Internet “cloud” which we will refer to as the core network.

This chapter has focused mostly on edge (or access) net-works because, in particular, they are the most expensive and difficult to deploy. But this does not hide the fact that for broadband wire-line networks, the availability of core network facilities, and in particular, fibre networks, is vital to promoting broadband deployment.

The following subsections explain the four main environ-ments in which edge networks can be deployed: (1) converged environments, (2) complementary environments, (3) compet-ing environments and (4) exclusive environments. Different policy and regulatory choices may apply to each environment, and new regulatory frameworks may have to be developed as broadband markets evolve and mature.

3.4.1 Converged EnvironmentsConverged environments are found in densely populated areas

with high-value fixed and mobile service subscribers, such as in metropolitan areas in North America. In a converged envi-ronment, fixed wire-line networks compete with or comple-ment BWA networks. And BWA networks are being integrated so that users of converged appliances (mobile handsets, per-sonal digital assistants (PDAs) or laptops) can move seamlessly between 2.5G or 3G networks and WiFi or WiMAX networks.

As these converged environments develop, two approaches to the integration of the wireless networks are being explored. Tightly coupled integration can be implemented at a relatively low level within the network. For example, a WiMAX net-work can be made to appear to a 3G mobile network facility as just another cellular access network. So there can be seamless handoffs between cellular and WiMAX networks.

By comparison, another approach, loose coupling, inte-grates networks at the Internet Protocol layer. This form of integration will still support limited handoff between networks. Finally, and more immediately, hybrid appliances can support a

limited form of network convergence, for instance today’s Wi-Fi and 2.5G-enabled handsets.

Converged environments are the most capital- and tech-nology-intensive, requiring significant vertical integration within network operators. They also place significant cost bur-dens on each user, because converged (or multi-mode) hand-sets and appliances are only now becoming available on the market and are consequently expensive.

3.4.2 Complementary EnvironmentsConverged environments are just now emerging. In

today’s reality, most broadband operators either compete or complement each other (or both). In a complementary environ-ment, broadband network providers offer their services to spe-cific niche markets. Networks are often broadly pervasive but not wholly ubiquitous. For instance, a single user may sub-scribe to three different broadband network services provided by three separate operators: 1 A cable-based home broadband network,2 A WLAN hotspot service available at coffee shop chains,

airport terminals, and other public places, and3 A 2.5G EDGE network provided by a major cell phone

operator.

None of these networks currently are integrated, as they would be in a converged environment. But they all respond to specific niche demands, often from the same end user. The same individual would use different services when working at home, enjoying a cup of coffee or tapping into email while on the go.

3.4.3 Competing EnvironmentsThere are many areas where the number of high-value

subscribers is substantial, but the organizational and capital environment has not yet been able to support much in the way of converged facilities. In these areas, multiple broadband net-works will be deployed in competition. Indeed, this is the case for most metropolitan areas in Europe, North America and parts of Asia and Latin America.

Even an area such as the complementary environment described in the previous subsection, with three complemen-tary broadband networks, there will also be competing network facilities. For example, the dominant local cable TV provider will offer home broadband (cable modem) service. Compet-ing with this service might be a DSL home broadband service provided by the incumbent local telephone operator. Similarly, multiple WLAN hotspot services might compete with each other, although not often at the same physical location (a single provider will often have exclusive rights in any given hotspot).

Indeed, we expect that in areas that enjoy a large number and high density of broadband users, instances of converged, complementary and competing broadband networks may exist in a simultaneous and overlapping manner.

3.4.4 Exclusive EnvironmentsIn marked contrast, much of the developing world has

no broadband provider at all. While regulators seek to pro-

CHAPTER 3 43


mote competitive broadband provision, a more likely scenario, at least in the short term, may be broadband provision in an

“exclusive” environment. It is not that regulators seek to create monopoly markets. The problem is that they may only succeed, at least in the short term, in attracting one broadband provider.

Sparsely populated rural areas, low-income areas, and places with challenging physical environments are all candi-dates for such exclusive broadband providers. These environ-ments are the most important and the most compelling for bridging the Digital Divide. Since only a single broadband network is likely to be commercially viable, at least initially, it is critical for potential broadband providers to pick the techno-logy family and governments to put into place the institutional and policy environment to support the network.

How can we best conceptualize and support broadband services in these exclusive environments, especially in those

“green-field” areas where there currently is no broadband net-work access? It is useful to consider the three technological platform families separately. The following subsections discuss the two “extreme” cases: fixed wire-line access and non-terres-trial wireless access, followed by the still-challenging case for fixed and mobile terrestrial wireless networks.

3.4.4.2 Wire-line Broadband for Exclusive Environments

Wire-line broadband platforms make good sense in areas with a large population of stationary users, and where opera-tors have sufficient access to capital, public policy support, and organizational integration. Indeed, fixed-wire networks are the cheapest per bit, when the goal is to transmit a very large numbers of bits. Researchers have argued that fixed wire-line solutions are the most cost-effective, compared with available wireless solutions, in environments with more than 40 broad-band subscribers per square kilometre.

The choice between the three primary wire-line technolo-gies (DSL, cable, and fibre) will probably be driven mostly by the predominant physical layer of the existing networks. If there is a high-quality copper network already in place, then broad-band DSL service might be the easiest and most cost-effective to implement. Similarly, if cable TV systems pass most homes, then those systems’ coaxial and fibre networks might be able to support broadband data.

It is important to note, however, that exclusive environ-ments are usually in areas that have neither a large popula-tion of high-value users nor the necessary capital to support the financially intensive requirements of building broadband, wired networks. In other words, few if any exclusive environ-ments would be best served by a fixed wire-line broadband network.

3.4.4.2 Non-Terrestrial Platforms for Exclusive Environments

VSATs are the clear solution for extremely remote areas and areas with extreme geographic conditions (for exam-ple, mountainous regions). Today, every part of the planet is touched by multiple satellite signals providing, for a price, broadband Internet services. But high prices tend to make satellite connectivity suitable largely for the most remote and difficult settings (where there are no alternatives). It is not

uncommon to be confronted with a monthly fee of thousands of dollars for broadband satellite service.

One possible approach that leverages economies of scale and creates opportunities for sharing expenses is to utilize VSAT technologies for backhaul networks. A local BWA net-work can then share this connectivity with multiple local sub-scribers.

3.4.4.3 Terrestrial Wireless Networks for Exclusive Environments

It is likely that the majority of exclusive environments, especially in rural and semi-urban areas of low- and middle-income countries, will receive broadband Internet access using some form of terrestrial wireless technology. This subsection considers the case for WiMAX and related networks in these contexts, exploring the business and policy environments that are needed to support these networks, then contrasting these networks with their mobile 3G counterparts.

WiMAX, Wi-Fi, 802.20 and related networks are an attrac-tive solution in many environments, especially if there are likely to be five or fewer subscribers per square kilometre. Deploy-ments can be bottom-up and potentially incremental. Wireless systems can work in environments with relatively weak institu-tional support and small capital markets.

Exclusive BWA networks can be visualized as a series of circles connected by sticks. The circles represent areas of cov-erage by a point-to-multipoint radio transceiver, representing the edge (access) network. The sticks represent the backhaul network that ultimately aggregates the traffic at a head end and passes it on to a higher-tier ISP – the core network. The

“circle” of edge connectivity need only have a radius of hun-dreds of metres, using popular Wi-Fi systems. The backhaul networks can have “sticks” that extend for many tens of kilo-metres, using WiMAX-related technologies.

Alternatively, the edge-network circles could have a radius of 10-20 kilometres, employing WiMAX base stations and omni-directional antenna facilities. They, in turn, could be backhauled with long-distance, point-to-point microwave net-works or fibre, wherever they exist.

In an exclusive environment with a very small number of anticipated subscribers, the fixed-wireless approaches of WiMAX and related networks promise to be the most cost-effective approaches. One reason is that they can connect only those specific areas requiring service, rather than spreading a ubiquitous signal over a wide area. This is a benefit in sparse and low-demand areas. Furthermore, most WiMAX technolo-gies do not support vehicular speeds (though newer 802.20 and WiMAX technologies are emerging that do handle mobility), nor do they usually support seamless handoff between cells.

As noted in Figure 3.4 above, however, distinctions in mobility, cost, ubiquity, and even capacity between WiMAX, 802.20, and the 3G family of networks, are eroding in favour of rapid convergence on a fairly similar feature set. For instance, mesh Wi-Fi and WiMAX systems are available that can provide nearly seamless and ubiquitous coverage. Similarly, WiMAX

CHAPTER 3


44

standards (and 802.20) that support vehicular motion are in the works.

3.4.4.4 3G for Exclusive Environments

Mobility and ubiquity are key elements of today’s 3G deployments. These may not, however, be critical for exclu-sive environments. In under-served and remote environments, a ubiquitous and mobile broadband connection, with seam-less handoff at vehicular speeds, may be more than the market needs or can bear. In fact, there are many examples of deployed 2G networks that are used mostly as fixed-mobile voice net-works. The Grameen Phone network is well known for provid-ing rural voice services in Bangladesh via village “phone ladies.” Providing support for vehicular travel and seamless handoff to these “phone ladies” is, according to the project’s principal organizer, an engineering solution in search of a problem. And if vehicular mobility is generally not needed for voice traffic, it is questionable whether it is necessary for broadband data.

In some areas, there may well be sufficient markets for ubiquitous and mobile 3G in exclusive environments, includ-ing relatively remote and sparsely populated areas. After all, mobile 2G networks enjoy very high penetration, even in some of the most remote and low-income areas in the world. One potentially viable approach may be to leverage existing 2G networks (or even build new, green-field mobile networks), upgrading them to 2.5G or 3G support. This may well depend on industry response and the regulatory environment.

To consumers, 3G and fixed BWA networks may appear to be converging, as fixed wireless adds mobility and ubiquity and 3G increases capacity. The two technologies do, however, come from very different communities. Fixed wireless tech-

nologies have a data networking and information technology pedigree, while 3G services are primarily tele com mu ni cation industry products. This difference is reflected in management of the two different kinds of networks. 3G networks are usu-ally deployed after top-down planning, by large, vertically inte-grated operators with access to substantial investment capital. Corporate headquarters announce new deployments, which commonly cover large areas. Fixed wireless networks, by con-trast, have often been driven by communities, small entrepre-neurs and retailers, who employ an incremental approach.

There are many tradeoffs between the various types of BWA networks that can be deployed within an exclusive envi-ronment. In some cases, upgrading 2G networks to 3G capa-bilities will be the most promising solution. In other areas, the 2G base station infra struc ture might be suitable for deploying WiMAX (or even 802.20) access points. Alternatively, in some green-field settings it may be desirable to abandon the mobile phone market and target specific areas for fixed wireless solu-tions. In other green-field environments, it may be most appropriate to leapfrog straight to a 3G mobile network that can provide broadband data to an entire region.

3.5 Power requirements for broadband

A reliable electrical power source is crucial to achieving pervasive broadband connectivity. High-capacity wireless net-works, and the multimedia appliances that broadband appli-cations call for, can heighten electric power demand. So any effort to deploy broadband networks must also consider the availability of power solutions, both on and off the electrical

Figure 3.8: Mobility v. Data Rate for Popular BWA Systems

CHAPTER 3 45


grid. Today’s modern desktop personal computers (PCs) need electricity to power:

– The central processor (now often consuming 80 watts),– The screen or monitor (consuming up to 100 watts),– The hard drive (which can consume 100 watts when spin-

ning),– RAM memory, graphics cards and other on-board devices

(up to 40 watts).

Low-power versions of these technologies do exist today. For instance, power-efficient laptops can be engineered to require less than 40 watts under normal operation.

Figure 3.9 shows the power consumption for a set of popular WLAN radio technologies. “BreezeMAX” WiMAX technologies can require 30-40 watts when in operation. A subscriber unit for Motorola Canopy, on the other hand, con-sumes less than 10 watts, while an indoor wireless router from Links consumes slightly more than 10 watts, on average.

What these figures suggest is that a standard, broadband-enabled PC can consume a fairly considerable amount of power; even low-power laptops require a significant amount. Furthermore, the power load of a PC is non-linear; it fluc-tuates over time (unlike a light bulb, for example), and a PC requires high-quality power sources.

This leads to the question of whether high-quality, reli-able power sources are available in areas that have inadequate or unreliable access to the electrical grid. Major solutions for off-grid powering of ICT systems include solar cells, small wind systems, micro-hydro power technologies, and generator sets (powered by diesel, gasoline, or bio-fuels). Each of these solutions has its benefits, as well as drawbacks. The follow-ing sections provide two case studies that illustrate innovative

applications of off-grid power for ICTs in developing coun-tries.

3.5.1 Tanzania: The Kasulu Teachers Training College

At the Kasulu Teachers Training College (KTTC) in Tan-zania, the “First Wave” of agrarian reform has mingle with the wave of new information technologies to create a broadband service that is changing the lives of some of the most impover-ished people in Africa. An Internet centre at the KTTC equips the next generation of teachers in Tanzania with Information Society tools, despite the fact that Kasulu has no electricity supply and its 15 phone lines have no data capabilities. What is unique about the project is that it is run by an unusual source of power – cow manure.11

Kasulu, population 33,668,12 is located in northwestern Tanzania. Being close to the Burundi border, the region has seen a huge influx of refugees escaping from violence in that country. In 2000, the Global Catalyst Foundation (GCF), in collaboration with the United Nations High Commissioner for Refugees (UNHCR) and Schools Online, decided to explore the feasibility of setting up a Kasulu Internet Project13 to promote cooperation and understanding between refugee Burundians and their Tanzanian neighbours in the Kasulu area. Another objective was to promote economic development and entrepreneurship in the impoverished region.

3.5.1.1 The Kasulu Programme

GCF is a private foundation established by Kamran Ela-hian, an IT entrepreneur and founder of Global Catalyst Partners, a Silicon Valley venture capital firm. GCF’s mission is “empowering people through technology,” and it supports

Figure 3.9: Power Consumption of Some Wireless Radio Products

CHAPTER 3


46

projects that “improve education, alleviate poverty, promote social tolerance and celebrate diversity” across the world.14 Schools Online, another non-profit started by Elahian, aims to help students use the Internet for learning and cross-cultural dialogue. Since 1996, the organization has assisted more than 5,700 schools in the United States and more than 400 schools in 35 other countries, helping to procure the equipment and support services to get online.15

Through the Kasulu Internet Project, GCF, Schools Online and UNHCR have collaborated in establishing Inter-net centres in three locations:

– At a UNHCR administered refugee camp, dubbed “Meat-ball,”

– At the Kasulu Folk Development College, and– At the Kasulu Teacher Training College.

GCF contributed USD 120,000 in operational funds for the three centres for three years; it is responsible for the overall management of the project. The local communities have con-tributed their labour to build computer labs. UNHCR pro-vides logistical and ad min is trative support, and Schools Online provides the hardware and software infra struc ture and funding for professional development and capacity building.16

At the KTTC facility, the 800 students enrolled in the teacher-training programme are the main users of the compu-ter labs, which are open to community members from 7-10 pm during the week (they are also open on Saturdays and Sundays), on a fee-for-use basis. The price is about USD 1 per hour (1,000 Tanzanian Shillings). Local rep re senta tives of non-

governmental organizations and UNHCR are frequent users of the labs, as well.

The teachers and students take CISCO Academy courses and study for the International Computer Driver’s License (ICDL), which has seven skill levels in word processing, file management, Internet services and other fields. The CISCO Academy has seven students at KTTC, who pay approximately USD 200 for a six-month course. Some 10 students from the town of Kasulu are also studying elements of the ICDL cur-riculum, paying approximately USD 200 per course. Another 30 students from the KTTC also attend these classes but pay lower fees. Revenues from these training programmes help KTTC pay for the cost of Internet access, maintenance and ongoing operations.

The Tanzanian Ministry of Education is now planning to make IT training compulsory for teachers. This will increase the demand for computer classes at KTTC, because all of the area’s teacher candidates will require instruction.

3.5.1.2 Kasulu’s Internet Links

The school is connected to the Internet through a VSAT service provided by I-way. The I-way connection offers 128 kbit/s downlink and 30–40 kbit/s uplink speeds. The Global Catalyst Foundation pays USD 500 per month for the connec-tion. The Internet is reasonably reliable, although it does fail occasionally.

As mentioned above, the GCF turned to an unusual source of power – cow manure – to power the KTTC compu-

Figure 3.10: UNHCR Refugee Camps in Tanzania

Source: Food and Agriculture Organisation.

CHAPTER 3 47


ter lab. The droppings of twelve cows are collected and fed into a 50-cubic-metre biogas plant that generates methane. This methane is then mixed with diesel in a 70-30 ratio and fed into a power generator producing 10 kilowatts of power – enough to run 15–16 computers for eight hours daily. Six shared UPS systems provide a 30-minute power backup.

Nothing goes to waste in this system. After the methane is extracted, the remaining sludge is removed from the biogas plant to provide fertilizer for the crops the college raises to feed its staff and students. The biogas processor, the dairy cattle, cowshed and 10-kilowatt generator cost a total of about

USD 18,120. The biogas system was built in late 2001, and the computers arrived at KTTC 2002.

Apart from powering computers, the college also wants to use methane for cooking. Currently, the college consumes several tons of timber every year to cook food for its students. Replacing timber with methane would be more ecologically friendly, and the cooks would also be spared the harmful effects of wood smoke. To do this, the college plans to increase the capacity of the biogas plant from 50 cubic metres to 200 cubic metres. To provide manure for the increased capacity, the col-lege would have to maintain 40–50 cows. The additional meth-

Figure 3.11: The KTTC Computer Laboratory

Source: Michael L. Best

Figure 3.12: KTTC’s Eco-Friendly Power Plant

The biogas digester (top left), methane/diesel generator (bottom left), and cow corral (right).Source: Michael L. Best

CHAPTER 3


48

ane generated would fuel a 30-kilowatt generator and allow for a 100-per cent methane-powered system.

The Swedish International Development Agency (SIDA) is considering funding the expansion. It would like to repli-cate the biogas system at 30 teacher training schools in Tan-zania and perhaps 3,000 secondary schools. The organization is also funding a micro-hydro power project in Kasulu where, along with a power micro-grid, fibre optic cables will also be strung. The fibre optic cables will be connected to the Internet through VSATs and used for networking NGOs, schools and clinics.

3.5.2 Tanzania: The Mtabila Refugee Camp

More than 50,000 Burundian refugees stay at the Mtabila refugee camp, where conditions at the camp are bleak and there is very little work to occupy the residents. The refugees often live in simple houses and survive mostly on food provided by the World Food Programme (WFP). Most of the adults at the camp are women,17 although the camp also has a significant adolescent population that faces an uncertain future.

With no telephone service in the camp, the 10 computers installed at the Mtabila Internet Centre are the refugees’ main connection to the outside world. When the centre is open to the general public there is often a long wait for Internet access. Every day, a clientele of about 30 refugees use the centre to send email, at a charge of approximately USD 0.20 (20 cents) per session.18 The Internet has reduced the refugees’ sense of isolation and enabled them to communicate with the outside world at a cost that is cheaper than the postal service. News and information downloaded from the web also help the refu-

gees stay abreast of world events. The refugees also learn to access websites, including those in the Kirundi and Kiswahili languages.

The Internet Centre is equipping refugees with skills they can use to help reconstruct Burundi. The Kasulu Online19 website reports that some 2,000 refugees have received higher education, directly benefiting from Internet access. Other refu-gees benefit through intermediated access, as others pass on current events to them and help them establish contact with friends and relatives who have email.

Ten teachers, selected by the refugee community, are undergoing training for the International Computer Driv-ing License (ICDL) at the camp. These same individuals also supervise the centre. The big surprise for these teachers, as well as camp ad min is trators, has been the demand for email. Many of the refugees have friends and family scattered throughout Africa, Europe and North America, and email is an affordable way to keep in touch with their community.20

Broadband Internet access is provided through a VSAT terminal, with the electrical power generated using photo-voltaic cells. The system’s setup cost approximately USD 37,152. It consists of 48 solar panels, generating 75 watts each. Solar power was chosen over biogas in this case, because solar eliminated dependence on fuel supplies and spare parts that a generator-powered system would entail.

Setting up the centres was a demanding task because of the poor quality of the telephone lines, which could barely sup-port email transmissions. Also, there were delays in importing essential equipment and challenges in dealing with relations among various governmental, international and non-govern-

Source: Michael L. Best

Figure 3.13: The Mtabila Camp VSAT and PV Systems Mtabila camp VSAT and PV systems (top); refugees waiting outside the computer lab (bottom left); and inside the computer lab (bottom right).

CHAPTER 3 49


mental agencies involved in the project. Despite these hurdles, the project was beginning to deliver the intended benefits, as can be seen from a refugee’s email from the Mtabila camp:

“I’m not able to tell you how happy we are to get connection to the Internet! Before I was connected to the Internet I felt lost. But now that I am connected, I feel saved. The world will not forget us now, because we, the refugees, can speak to the outside.”21

3.6 Conclusion

This chapter has surveyed the development paths of three important families of broadband access technologies: wire-line services, broadband wireless access, and non-terrestrial broad-band wireless systems. All three of these families of systems can provide optimal choices in various circumstances. In densely populated, high-income areas, a variety of these systems will converge, compete, or complement each other. In other set-tings – those with limited numbers of potential broadband subscribers – single network systems are likely to prevail in isolation. And in cases where the density of potential subscrib-ers is lowest, various wireless broadband solutions are likely to be the most cost-effective.

These sets of technologies can be placed on a continuum, based on the density of high-value subscribers. With densities of 40 people per square kilometre and above, wire-line solu-tions are likely to be cost-effective. As density declines to just a few people per square kilometre, fixed wireless solutions, such as WiMAX, are more optimal.

Broadband deployment does not depend solely on tech-nological choices. Capacity building to develop local human

resources is vital. Education and training – including practical showcases and pilot programmes – are equally important in the quest to promote broadband access. Governments can further organize open regional, national and local forums dedicated to identifying and meeting the broadband needs of all key stake-holders, including:

• End users (both current and potential);

• National and local development programmes on health-care, education and e-government;

• Public sector institutions such as universities, libraries and local and national government offices;

• Local entrepreneurs;

• Traditional tele com mu ni cation operators;

• System manufacturers; and

• Owners of fibre infra struc ture from other sectors, such as transport and utility companies.

Together, these stakeholders can strive to meet their broad-band requirements and achieve the UN Millennium Develop-ment goals. Once broadband needs are identified, they can be met through the development of open market-based networks, fostered through effective regulatory practices.

Broadband deployment in rural areas of developing coun-tries is likely to be fuelled both by civil society, including a range of public service actors, and the private sector, including local entrepreneurs that are building sustainable businesses. What is needed are regulatory frameworks designed to lower business risks and open markets to a full range of potential broadband providers. These key issues are addressed in the next chapter, which discusses the role of regulators in providing an environ-ment that promotes broadband deployment.

CHAPTER 3


50

1 See http://standards.ieee.org/getieee802/portfolio.html for more information about the IEEE802 standard.2 In Sweden, apartment house companies own their own fibre networks to serve the tenants in their buildings. These networks connect to a point of presence of all

operators at a metropolitan area hub. In a departure from the fibre to the home model, they outsource the link level of their fibre from the home networks to service-providers serving their tenants.

3 See: http://www.itu.int/ITU-D/treg/publications/AfricaIXPRep.pdf 4 See http://grouper.ieee.org/groups/bop/ for further information about P1675 from the IEEE).5 See http://www.iec.ch/tctools/dashbd-e.htm for further information about the working groups with the CISPR.6 For further information about this process, please see http://www.iec.org/online/tutorials/acrobat/opt_net.pdf.7 See http://www.iec.org/online/tutorials/sdh/ for more information about SONET.8 Commercial acceptance has been disappointing on these systems and it is thought that ArrayComm intends to abandon this technology in order to focus on WiMAX

based systems.9 ITU Internet Reports: The Portable Internet, 2004, see http://www.itu.int/publications/bookshop10 Portable Internet.11 “Cow pats fuel computers” BBC Online. See http://news.bbc.co.uk/2/hi/technology/2957488.stm12 See http://www.tanzania.go.tz/census/districts/kasulu.htm13 See http://www.kasuluonline.or.tz/14 See http://www.global-catalyst.org/15 See http://www.schoolsonline.org/whoweare/index.htm16 See http://www.schoolsonline.org/whatwedo/update_sep_2002.htm17 “Rape at the end of the world.” See http://www.unfpa.org/focus/tanzania/rape.htm18 “Cow pats fuel computers” BBC Online. See http://news.bbc.co.uk/2/hi/technology/2957488.stm19 See http://www.kasuluonline.or.tz/20 “Tanzanian refugee camp gets wired for Internet” See http://hrea.org/lists/huridocs-tech/markup/msg00932.html21 “Start. Succeed (or Not). Repeat.” Los Angeles Times. See http://www.latimes.com

51


CHAPTER 4

4.1 Introduction

Previous chapters have explored the promising landscape of new broadband opportunities that is emerging as waves of innovation reshape the ICT sector. These innovations are occurring across all aspects of the industry, from technologi-cal developments and business models to regulatory and policy frameworks, creating broadband opportunities for end users, large-scale network operators, small entrepreneurs, local com-munities and governments alike.1 New broadband opportuni-ties require a new vision by potential broadband providers, and a new paradigm for policy-makers and regulators. Broadband is completely transforming the ICT sector. Put simply, broad-band cannot be treated as “business as usual.”

Network operators that fail to join the broadband world risk being left behind. The reality is, however, that deployment of broadband access technologies in developing countries is often constrained by a lack of tele com mu ni cation infra struc-ture – especially backbone networks – and by large-scale net-work operators’ concerns about potential revenue generation. Larger commercial operators are often discouraged from pro-viding broadband access in marginal areas because of the high costs of deploying networks and the fear that retail charges may be too high, relative to disposable incomes, to result in large-scale take-up.2 Meanwhile, potential broadband market entrants are often kept out of the market by regulatory frame-works designed for another era.

A new, pro-broadband regulatory paradigm will harness the power of all potential broadband providers, tailoring the regulatory framework as needed. This pro-broadband regula-tory paradigm will expand on the existing regulatory practices developed after the revolutionary advent of the internet. Broad-band technologies now allow discussions about the internet to be taken to a higher level, because they accelerate download speeds and provide extended mobility. In response, regulators will seek to spur competition at all levels of the broadband value chain, from the link and transport infra struc ture layers to the content required to fuel demand and the computers needed to access broadband services and applications. This will require regulators to take a comprehensive and coordinated approach, as identified by the world community of regulators participating in the 2004 ITU Global Symposium for Regula-tors (See Box 4.1).

The new broadband regulatory paradigm will paradoxically require regulators to do both more and less than “regulation as usual.” Regulators will do more in making potential broadband providers – such as local communities and non-governmental organizations – aware of the technologies and applications they could provide. Regulators will also coordinate more actively with other government agencies and public institutions (such as universities) to drive the demand for health, education and government services that employ broadband technologies and applications. Meanwhile, regulators will do less by dismantling outdated regulatory frameworks that restrict market participa-tion.

A successful, comprehensive policy framework in a devel-oping country with major challenges in rural connectivity can not only foster greater investment by large scale-operators, it can encourage public institutions and smaller market players to deploy broadband networks to suit their own operational or commercial objectives. As highlighted in Chapters 2 and 3, many broadband technologies can be deployed incrementally, without large-scale, nationwide deployment plans. The abil-ity to deploy broadband on an incremental basis allows local communities and development project managers to include broadband capabilities in multiple projects. Similarly, small and micro entrepreneurs can launch new businesses based on broadband access.

The key is for regulators to determine how to minimize any obstacles to incremental deployment, which may include:• Prohibitive pricing for inter con nection with incumbent

operators;• High costs to access existing infra struc ture or resources

from parallel utility sectors (i.e., energy or transport utili-ties); and

• Extensive and onerous licensing processes.

Once small-scale market entrants can establish demand for broadband services, larger operators are more likely to see the benefits of market entry, themselves. If need be, their interest can be further awakened through appropriate rewards schemes, such as tax exemptions or targeted subsidies from universal service funds.

It is also important to highlight programmes and content that may drive initial broadband take-up in areas where it has

Authors: Yang-Soon Lee, Dr. William Bratton & Wu Wei Shi, Spectrum Strategy Consultants and Susan Schorr, ITU/BDT

4 THE ROLE OF THE REGULATOR IN BROADBAND DEVELOPMENT

CHAPTER 4


52

Box 4.1: GSR 2004 Best Practice Guidelines

Regulators meeting at the 2004 Global Symposium for Regulators, held 8-10 December, 2004 in Geneva, approved a set of “best practice” guidelines for regulators to promote the development of low-cost, easily accessible broadband networks and ap-plications. The full text of the document is as follows:

“We, the regulators participating in the 2004 Global Symposium for Regulators, have identified and proposed best practice guidelines to achieve low cost broadband and internet connectivity. Our goal is the creation of national regulatory frameworks that are flexible and enable competition between various service providers using multiple transport and technology options. We believe the best practices outlined below will help bring social and economic benefits to the world’s citizens.

An enabling regulatory regime that encourages broadband deployment and internet connectivity

1) We encourage political support at the highest government levels with such support expressed in national or regional policy goals. These include an effective regulator separated from the operator and insulated from political interference, a transpar-ent regulatory process, and adoption and enforcement of clear rules.

2) We believe that competition in as many areas of the value chain as possible provides the strongest basis for ensuring maxi-mum innovation in products and prices and for driving efficiency.

3) We encourage regulators to set policies to stimulate competition among various technologies and industry segments that will lead to the development and deployment of broadband capacity. This includes addressing barriers or bottlenecks that may exist with regard to access to essential facilities on a non-discriminatory basis.

4) We believe that the primary objective of regulation should be to secure fair and reasonable access for competitive broadband services, including internet connectivity.

5) We encourage the maintenance of transparent, non-discriminatory market policies in order to attract investment.

6) We encourage regulators to adopt policies that are technology neutral and do not favor one technology over another.

7) We encourage regulators to take into consideration the convergence of platforms and services and that they regularly reassess regulatory regimes to ensure consistency and to eliminate unfair market advantages or unnecessary regulatory burdens.

8) We encourage regulators to allocate adequate spectrum to facilitate the use of modern, cost effective broadband radiocom-munications technologies. We further encourage innovative approaches to managing the spectrum resource such as the ability to share spectrum or allocating on a license-exempt non-interference basis.

9) We urge regulators to conduct periodic public consultations with stakeholders to inform the regulatory decision-making process.

10) We recommend that regulators carefully consider how to minimize licensing hurdles.

11) We encourage the development of a regulatory framework that permits ISPs and broadband providers to set up their own last mile.

12) We encourage regulators to provide a clear regulatory strategy for the private sector in order to reduce uncertainty and risk, and remove any disincentives to investment.

Innovative Regulatory Policies Must Be Developed To Promote Universal Access

1) We recommend that the promotion of access to low cost broadband interconnectivity should be integrated from “grass-roots” efforts to identify local needs all the way through the “tree-tops” of international law. Governments, business and non-governmental organizations should be involved.

2) We recommend that regulators adopt regulatory frameworks that support applications such as e-education and e-govern-ment.

3) We encourage each country to adopt policies to increase access to the internet and broadband services based on their own market structure and that such policies reflect diversity in culture, language and social interests.

4) We encourage regulators to work with stakeholders to expand coverage and use of broadband through multi-stakeholder partnerships. In addition, complementary government initiatives that promote financially sustainable programmes may also be appropriate, especially in filling in the market gap that may exist in some countries.

5) We encourage regulators to adopt regulatory regimes that facilitate the use of all transport mechanisms, whether wireline, power line, cable, wireless, including wi-fi, or satellite.

6) We encourage regulators to explore programmes that encourage public access to broadband and internet services to schools, libraries and other community centres.

7) We encourage regulators to implement harmonized spectrum allocations consistent with the outcome of ITU Radiocom-munication Conference process and each country’s national interest. Participation in this well-established framework will facilitate low-cost deployment of equipment internationally and promote low-cost broadband and internet connectivity through economies of scale and competition among broadband vendors and service providers.

CHAPTER 4 53


been absent. In developing countries, governments are often the largest ICT users. Governments can launch top-down e-government initiatives to induce citizens to get online, where they can avoid bureaucratic appointments, queues and paper-work and save on travel costs. E-government services also help the government itself by developing ICT capacity to interface with the citizenry en masse. But e-government applications will only drive broadband take-up if the public accepts them as useful.

Once demand is proven, commercial broadband providers will respond with entertainment content designed and priced for developing-country users. In short, promoting broadband access and services in urban and rural areas of developing countries requires a new vision characterized by reduced regu-latory burdens and an environment of innovation, collabora-tion and creativity.

4.1.1 A Look Back

Ten years ago, the predominant model for providing tele-com mu ni cations involved a state-owned incumbent opera-tor deploying a fixed-line public switched telephone network (PSTN). Since then, that model has given way to alternative competitive network platforms such as cable television, fibre optic, satellite, second generation (2G) mobile cellular, and more recently, 3G and broadband wireless – all enabled by new technologies. Most tele com mu ni cation networks have been provided on a nationwide, or at least regional, basis.3 Today, most large-scale networks are in private hands, mean-ing they must be profitable to operate. Even the majority of PSTNs today have been at least partially privatized.4 The busi-ness model adopted by large-scale network operators requires significant infra struc ture investment and high subscriber rev-enues to remain financially viable. Most national regulatory frameworks were designed for such large-scale network opera-tors and service providers.

What has this first wave of sector reforms achieved? The short answer is: a lot. Telephone access has quadrupled since 1990, from 10 per cent to 40 per cent of the world’s popula-tion. By the end of 2004, there were an estimated 1.19 billion fixed telephone lines in operation around the world and 1.8 billion mobile subscribers. The rise in mobile subscribers is phenomenal, and more than 50 per cent of mobile subscribers today are in developing countries.5 Still, many of these users

are located in urban areas. The challenge for regulators today is to build on this urban mobile growth to provide an enabling environment to bring both voice and broadband internet serv-ices to rural areas.

The stage is now set for such a rural renaissance. An entirely new player has appeared on the scene: the community broadband provider, largely enabled by low-cost broadband wireless technologies. Local community initiatives are start-ing to provide broadband services to users in remote areas not served by large-scale networks. These small-scale broadband providers range from public institutions such as libraries, edu-cational institutions, health facilities and local governments, to non-governmental organizations and small and micro entre-preneurs that can be profitable at margins of only a few dol-lars a day. Whether private or public, all community broadband providers are concerned with costs. The lower the costs of providing broadband services and applications, the greater the opportunity for community initiatives to succeed in bringing ICTs to rural users. Keeping regulatory costs down will give these local initiatives a better chance of success.

4.1.2 A Look Ahead

A new regulatory framework, tailored to the unique cir-cumstances of local community initiatives, may be needed to help small-scale broadband providers foster growth in rural areas. This chapter explores options for such a new regulatory framework.

Meanwhile, the regulatory framework designed for large-scale network operators is also evolving to provide greater incentives to deploy broadband access networks in rural areas. This chapter takes a hard-nosed look at the commercial realities of major operators. Commercial operators are profit-driven and cannot be expected to provide services that do not yield profits. Where regulators are unable to induce large operators to build broadband networks in rural areas, either through financial incentives or build-out requirements, they can focus efforts on developing new pools of potential broadband users. This new source of demand will then give major operators incentives to extend their networks out to these new markets, offering inter-net backhaul and inter con nection with urban networks.

This chapter first identifies the importance of broadband and the potential positive impacts that access to broadband technologies can provide, before discussing the key issues

Box 4.1: GSR 2004 Best Practice Guidelines

Broadband is an Enabler

1) Regulation should be directed at improving the long term interests of citizens. Broadband can contribute to this by improv-ing and enabling education, information, and increased efficiency. It can reduce costs, overcome distance, open up markets, enhance understanding and create employment.

2) We encourage regulators to educate and inform consumers about the services that are available to them and how to utilize them so that the entire population benefits.

3) We urge regulators to work with other government entities, industry, consumer groups, and other stakeholders to ensure consumers have access to the information they need about broadband and internet services.”

See: http://www.itu.int/ITU-D/treg

CHAPTER 4


54

associated with such deployment in developing countries. The chapter then highlights the role of competition in accelerating the deployment of supporting networks and reaches some con-clusions about the appropriate regulatory framework to create a conducive environment for deploying broadband networks. The chapter will conclude by identifying other potential policy measures regulators can undertake to support the deployment and take-up of broadband access technologies.

4.2 The Importance of Broadband in Developing Countries

The deployment of broadband access technologies deliv-ers several positive impacts in developing countries. These include:• Eroding information differentials resulting from geo-

graphical constraints that prevent marginal communities from participating in regional, national or international processes.

These processes may be social (for example, relating to education or health), political, economic or financial. Resi-dents of rural areas, for example, need access to financial infor-mation and advice, as well as information about markets (both as buyers and sellers). Broadband technologies, therefore, can facilitate the integration of marginal communities into wider processes beyond the geographical limitations of their specific areas.• Access to regional, national and international

resources through broadband access technologies can substantially improve the living standards of marginal communities.

For example, improving accessibility to e-health systems that allow remote diagnoses and treatment is particularly useful in marginal areas, where access to medical equipment and expertise would otherwise be limited. Similarly, broadband access technologies can be used to provide remote education and training services (for example, the African Virtual Univer-sity6 and, at a more advanced level, Universitas 21).7 The pro-vision of such access can have a significant positive impact on living standards.• Broadband access technologies can enhance the sustaina-

bility of marginal communities by supporting the trans-fer of knowledge and expertise to marginal commu-nities.

Rural doctors can receive regular training via e-health sys-tems, allowing them to stay in their communities rather than leaving, either temporarily (for training) or permanently (for professional growth and improved fortunes). Online training also brings to the community the latest medical techniques and treatment options. Similarly, broadband access technologies can enable the transfer of agricultural knowledge, improving productivity and limiting soil exhaustion and desiccation from unsustainable farming techniques.

• Improved information flows from broadband access tech-nologies can increase the range of market opportuni-ties available to marginal communities.

Farmers need not be limited to local buyers. They can use broadband technologies to access geographically remote markets, including auctions. This benefit is particularly pro-nounced for perishable agricultural products where the use of online markets lets producers establish contacts with a wider range of potential buyers across larger geographies, boosting prices and maximizing rural incomes;• Broadband technologies can improve business produc-

tivity in developing countries.

Access to greater information sources, e-mail and other supported services (VoIP, for example) allow businesses to lower their business costs and improve their revenue-gener-ating potential. In India, for example, widespread broadband deployment is expected to increase labour productivity by 11 per cent, leading to direct employment of 1.8 million and total employment of 62 million by 2020.8 • The deployment of broadband access technologies may

support the growth of regional and national IT industries.

This growth can impact positively on GDP growth. For instance, in Korea (Rep.), broadband deployment has sig-nificantly underpinned the IT sector, which accounted for approximately 50 per cent of the GDP growth rate in 2002.9 The Confederation of Indian Industry’s National Broadband Economy Committee estimates that broadband will contrib-ute USD 90 billion to the Indian economy between 2010 and 2020.10 Specific examples of how the deployment and take-up of broadband access technologies has benefited marginal com-munities and developing countries are illustrated in Box 4.2.

There is no doubt that broadband access technologies can provide substantial benefits to end users in developing coun-tries and marginal communities. But the services provided over the networks must be sufficiently targeted to offer end users real benefits. In addition to regulations that aim to create a competitive broadband market providing universal and equita-ble broadband access, governments are also focusing on devel-oping policies to include local communities in the design and im ple men ta tion of broadband initiatives.

4.3 Key Issues in Promoting Broadband in Developing Countries

Given the potential economic and social benefits of broad-band access technologies, there is a clear need to develop ways to boost the relatively low levels of broadband take-up in many developing countries. The fundamental problem is that there is an array of constraints to take-up in these countries, and they affect all the components of the value chain (See Figure 4.1). There are impediments in the supply of online content, as well as the supply of broadband services and products. There are also barriers to widespread connectivity. Finally, there currently is limited demand for broadband in many developing countries.

CHAPTER 4 55


Limited consumer demand, however, is likely related to a lack of consumer awareness of the potential benefits of broadband. Once more broadband networks are deployed, it is quite possi-ble that consumer demand (in terms of content and volume) in the developing world will not be so different from demand in the developed world. The main difference may remain in the rates that consumers are able to pay, and the portion of their disposable incomes they are willing to allocate to broadband services and content.

In terms of supply, the key constraints to deployment of broadband services and products in developing countries are:• Insufficient compelling content, especially in local lan-

guages and with specific reference to local circumstances or issues;

• Lack of understanding about the benefits of broadband;• Little or no incentive for fixed-line incumbents to offer

broadband access technologies (particularly if they risk

cannibalizing PSTN and ISDN revenue streams) in the absence of market competition;

• Competing demands for investment of operators’ capital (in developing countries operators often believe they can generate a better return on capital by deploying mobile networks than additional fixed-line or broadband net-works);

• Lack of market competition to encourage operators to develop and commercially deploy broadband services; and

• Lack of a regulatory framework designed to encourage broadband deployment by large-scale incumbents, or to sponsor market entry by potential broadband competitors, which might include public institutions (e.g., universities) or local, community-based providers.

Box 4.2: The Impact of Broadband Access – Examples in Developing Countries

In the Reserva Ecologica do Xixuanú of Brazil’s Amazon region, a telemedicine project has been launched to transmit medi-cal information from local communities, via satellite, to the United States for remote diagnosis.

Rural villages in Bhutan that were previously not connected by traditional telephone service are now provided with inex-pensive basic voice telephone access using wireless broadband technologies.

In China, students in rural villages are able to participate in distance-learning courses set up by major Beijing universities, using VSAT broadband satellite access.

In a small, remote town in the mountainous region of northeastern Ecuador, Wi-Fi technology has enabled the mayor to access online government databases. In addition, broadband access is used to promote ecotourism in the area and to help local businesses (SMEs or “small and medium enterprises”) become more competitive.

In Laos, WLAN networks have been rolled out to villages, allowing people to make local and international VoIP calls, signifi-cantly improving their connectivity. Other activities made possible by the WLANs include accounting, letter writing and e-mails, as well as support of local business activities.

In South Africa, institutions are connected to international institutions using broadband technologies, to advance coopera-tion in various research and development initiatives.

In Uganda, rural schools can gain access to educational tools via broadband access.

Recognizing the benefits of broadband for the agricultural sector, the government in India has announced plans to set up a network of computer kiosks in 25,000 villages. The goal is to help farmers sell their produce to the highest-paying customers. The national rollout is due to be completed in 2007.

Figure 4.1: A Simplifi ed Broadband Value Chain

CHAPTER 4


56

In terms of connectivity, the key constraints to take-up are:• Lack of hosting/storage facilities within many developing

countries, a situation that requires much content to be stored overseas, straining international connectivity;

• Limited international connectivity, which impacts on the data rates available, the quality of the service and the cost of bandwidth;

• Lack of backbone connectivity in many areas – and where backbone networks do exist, they are owned by incum-bents that control the costs and quality of leased lines available to competitors; and

• Concern on the part of large-scale network operators over the commercial viability of deploying broadband networks in rural or remote areas, where the costs of network oper-ation and service provision may threaten to make services unaffordable for consumers.

In terms of demand, the key constraints in developing countries are:• Lack of consumer demand, resulting from limited con-

sumer purchasing power;• Lack of consumer awareness (which is linked to pricing

and low purchasing power), coupled with a lack of coor-dination by key stakeholders (i.e., universities, public institutions and local communities) that could drive fur-ther awareness and demand;

• Excessive pricing of broadband products and services, especially when compared with average incomes;

• Greater priority placed on mobile voice communications than data services; and

• Limited availability of affordable end-user terminals. Deploying broadband access networks only makes sense if potential users have the computers or handsets with which to access them.

Many of these factors, of course, are not unique to devel-oping countries. But they are more daunting than in devel-oped countries, where consumers are more likely to be able to afford broadband services and products at price points that more accurately reflect the underlying costs of providing them. Similarly, the relatively low penetration rate of personal com-puters (PCs) in developing countries significantly weakens the demand for broadband access. PC penetration is an issue of affordability, but also a factor of electricity provision, since it is impractical to use a personal computer where there is no electricity to power it.

Innovation and competition in developed countries, how-ever, are pushing manufacturers and engineers to develop very low-cost, simple devices that can be tailored for applications and content in developing country contexts. This may reflect the development of a niche market for low-priced, narrowly functioning equipment, as manufacturers consider how to cater to untapped markets. Considering the supply, connectiv-ity and demand factors that limit broadband take-up in many areas of developing countries, regulators have an important role in minimizing the impact of these factors or developing appropriate solutions.

4.4 Providing Incentives for Network Investment

In developed countries, relatively high levels of teleden-sity allow for a greater focus on the promotion of service-based competition. The primary role of most regulators in develop-ing countries, however, is to create an environment conducive for network investment where little or no tele com mu ni cation infra struc ture exists. But before confidently committing them-selves to network investment, operators look for government regulatory approaches that are consistent and not arbitrary. This is especially true for broadband access technologies, which are often capital-intensive and – in developing countries – consid-ered a more risky investment.

Therefore, any regulatory framework (and this includes the track record on enforcement) that creates uncertainty or the risk of financial loss will deter market entry and subsequent network investment. Regulatory agencies that fail to provide consistency risk deterring investment, thereby constraining the development of their tele com mu ni cation networks, including broadband ones. Within this context, there are several steps a regulator can take to promote the deployment and take-up of broadband networks. These are detailed in the following sub-sections.

4.4.1 New Entry and Market Liberalization

Market liberalization remains the most effective mecha-nism to encourage greater investment in tele com mu ni cation networks. Experience shows that liberalization through the licensing or authorization of new operators will yield greater benefits than incentive- or obligation-driven approaches tar-geted at only a monopoly or duopoly. The absence of liber-alization, meanwhile, removes a significant incentive for an incumbent to invest in networks, new services or quality of service (QoS). Historical precedent in countries such as Brazil, Hungary, and India demonstrates that both incumbents and new entrants invest more when faced with market competi-tion.

It is clear, however, that after market liberalization, the incumbents in many developing countries continue to be in a position to invest significantly more in network deploy-ment than are new entrants, especially in marginal areas. So the incumbent can often be the most important (and often the largest) source of funds for tele com mu ni cations investment in the longer term.

Given this, a number of commentators have argued that excessive market competition may reduce the incentive for incumbents to continue large-scale investment. This may be particularly the case where tariff re-balancing has not been fully undertaken and new entry results in rapid margin ero-sion for incumbents in the more lucrative long distance and international markets. This may result in a deterioration in the incumbent’s financial performance (and stock valuation), reducing its capability and willingness to undertake new net-work investment or diversify into new lines of business. Regu-latory frameworks, then, should not unduly impact network investment and the diversity of new lines of potential business.

CHAPTER 4 57


Because of the costs of network deployment – especially of broadband access technologies – there have been some sug-gestions that network “over-building” (duplication) could be avoided by creating a “super” network operator that would provide wholesale network access to retail service providers. In effect, this would amount to a transition to service-based competition rather than network-based competition. This idea often appeals to policy-makers in developing countries, because of their concerns about the availability of investment funds. But im ple men ta tion raises significant competitive concerns.

Incumbent operators tend to advance the argument for “super operator” status in an attempt to limit competition. “Super” network status allows incumbents to retain complete control over infra struc ture and, by extension, over the devel-opment of competition. In both developing and developed countries, incumbents have proven remarkably effective at controlling access to their infra struc ture even when the regula-tor is relatively strong. It is unlikely, therefore, that a “super” network scheme would be effective in the absence of a very strong regulator that could ensure that service providers have equitable access to the network.

One alternative is for governments to fund the construc-tion of such comprehensive networks directly, and to oversee access to them by both incumbents and new service provid-ers. Such an approach has been used successfully in Singapore, where the government in 1997 funded the rollout of SINGA-PORE ONE, a broadband backbone network accessible by the incumbent and new operators, including the cable TV opera-tor. Of course, not all governments have the resources to build such broadband backbone networks – particularly nationwide.

A more practical alternative, therefore, may be offered by infra struc ture sharing. For example, allowing mobile operators to roam onto each other’s 2G and 3G networks in rural areas would save significant network costs while enabling greater network coverage. In fact, competitors have even started shar-ing the bulk of their wireless access network facilities in non-rural areas; one example is Telstra’s and Hutchison’s shared 3G network in Australia. Similarly, France has allowed infra struc-ture sharing among 2G operators in order to reach unserved rural areas known as zones blanches. Such roaming and infra-struc ture-sharing arrangements could also apply to new broad-band wireless networks.

Fibre backbone networks are scarce in most developing countries, making broadband deployment more challeng-ing. Fibre backbones can boost the capacity of DSL networks. Extending fibre closer to rural areas can also facilitate internet backhaul for wireless broadband technologies. Again, rather than resorting to a “super” fibre backbone operator, regulators can promote synergies between different kinds of utilities or projects that employ internal communications links. Energy and transport infra struc ture projects, such as electrical plants, highways, railways and pipelines, could be encouraged to deploy fibre as part of their projects. Telecommunication oper-ators could then access these fibre facilities to augment their networks.

Regulators could also provide incentives, such as tax breaks, for 2G mobile operators to build their own backbones using fibre, instead of the more commonly used microwave links. The regulatory framework could allow owners of such communication resources to lease unused capacity to others for commercial deployment.

In addition, countries are introducing new regulatory tools to encourage network investment by smaller market play-ers. Ireland, for example, has found that rather than impos-ing national broadband rollout and coverage obligations on large-scale operators, it can achieve greater success by allowing wireless broadband providers to enter small local service areas. Ireland’s practice of licensing small local service areas--defined as a 15-kilometre radius around a base station—has led to a significant rise in new broadband subscribers in non-urban areas.11

Encouraging competitive market entry is part of a larger package to promote broadband deployment by a full range of potential broadband providers. Additional elements of this larger package are explored below.

4.4.2 The Role of Foreign Ownership

New market entry and subsequent investment, including in broadband access networks, is likely to be supported if there are no restrictions on foreign ownership of licensees. This is particularly true in developing countries, where capital avail-ability may be limited. Foreign ownership brings the possibility of incremental capital funding, as well as managerial expertise and international best practices. Increasingly, governments and

Box 4.3: Regulatory Methods To Boost Deployment

Regulatory frameworks, tailored to the unique circumstances of local community initiatives, can encourage small-scale broadband providers to provide broadband access in rural areas.

An incentive-based regulatory framework using targeted grants or tax exemptions can encourage both large and small-scale network operators to deploy broadband networks in rural areas.

Where financially viable, broadband deployment requirements can be made part of the licensing commitments required from new market entrants.

Governments can also promote the development of sufficient supporting network infrastructure to enable the provision of broadband services (e.g. backbone connectivity). They can also drive initiatives (e.g., e-government programmes) that provide an important source of demand for broadband facilities and services.

CHAPTER 4


58

regulators seek to attract foreign ownership, rather than restrict it on the premises of national security, cultural protection and domestic economic development (although these often remain concerns for policy-makers).

There are persuasive reasons to facilitate and support par-tial and even full foreign ownership of new entrants in many circumstances. These reasons include:

• There is an international trend to ease foreign-ownership restrictions, partially encouraged by World Trade Organi-zation (“WTO”) agreements, but also driven by increasing recognition of the substantial benefits that foreign owner-ship provides;

• It is possible to address investment requirements through licence commitments, ensuring that these are not so high as to lead to excess capacity; and

• Foreign investors are more likely to see greater investment risks in situations where they lack managerial and opera-tional control. Greater financial ownership is likely to be associated with greater managerial control and, by exten-sion, enhanced company performance.

In addition, there are increasing examples of foreign own-ership of new entrants that result in substantially improved tele com mu ni cation infra struc ture. Foreign ownership will not, per se, result in greater broadband access network deployment. But it may support such deployment through greater access to capital, more managerial experience and, potentially, lower unit costs.

Besides allowing foreign ownership, governments can also tap into global capital markets themselves and tap international lending agencies for funds to improve and upgrade their tele-com mu ni cation networks. This is much less difficult than ever before, because it has been proven decisively that operating tele-com mu ni cation networks – when they are properly deployed and managed – is a commercially lucrative business.

4.5 Broadband Licensing

Regulators can often affect the rate and prevalence of broadband network build-outs through their licensing pro-visions. In general, regulators are experimenting with new licensing frameworks, often eliminating operator-specific con-ditions in favour of generally applied regulatory codes. Licens-ing frameworks can also be employed to provide incentives or direct mandates for network deployment.

4.5.1 Consistent Licensing and Authorization Frameworks

The deployment of broadband can be expedited through relaxing the licensing conditions for large-scale broadband access providers and by establishing a consistent licensing framework that is clearly targeted to achieve a set of defined policy objectives. In addition, regulators are increasingly using general authorizations in lieu of onerous licensing regulations to ease market entry.12 Where licences are distributed, it is important to:

• Have an efficient ad min is trative process that is transparent and consistently applied, together with minimal ad min is-trative requirements;

• Ensure that any terms and conditions included in licences are not financially punitive and allow operators to achieve sufficient financial return over the life cycle of their investments;

• Ensure that licensing fees are commensurate with the required activities of the licensee – that is, if the licensee is expected to deploy substantial broadband access infra struc-ture, fees should be reduced to reflect this high level of investment, rather than treating operators as “cash cows” to be milked for government revenues;

• Require regulators to establish and enforce appropriate monitoring mechanisms to ensure that licensees meet their commitments or specific conditions of license agree-ments;

• Recognize that re-negotiating licence commitments increases the risk associated with network investment to promote broadband access deployment and take-up.

While licences, or at least general authorizations, are usu-ally required for large-scale broadband infra struc ture operators, regulators are increasingly lightening such requirements for operators and service providers in small, rural areas. Facilitat-ing broadband market entry in these areas allows broadband providers to test their broadband business cases on a small scale. Some small-scale broadband providers may later decide to commit to more large-scale deployment. Thus, regulators can replace licensing requirements for commercial community broadband providers by a general authorization or registration framework, just as some countries have already established

“open entry” policies for internet service providers (ISPs).

Where broadband access will be used exclusively for public services, such as in health facilities or libraries, regula-tors may question whether licensing should apply at all. It is also particularly important that licence fees for very small broadband providers be kept as low as possible, if not elimi-nated altogether. Licensing obligations that may apply to large-scale operators, such as rollout and coverage obligations or contributions to universal access funds, can be minimized or eliminated in a regulatory framework targeted to community broadband providers.

A case can also be made for allowing resale of broadband services without any licensing requirements in rural areas. For example, broadband subscribers in a rural area could be allowed to use their broadband connections to set up a public kiosk and resell the service. The customers of these kiosk services might not otherwise be able to afford service at all, at subscription or monthly rates. In this way, additional economic activity would be generated through increasing broadband access.

It is important to note that reducing or eliminating licens-ing requirements is not synonymous with ceasing to regulate service providers. In some countries, for example, tele com mu-ni cation licensing is not widely used as a regulatory instru-ment. Instead, regulatory rules are enacted through universally applicable regulatory codes, decisions or orders. Even with

CHAPTER 4 59


open-entry or simple notification policies, local commercial broadband providers could still be subject to government over-sight in areas such as consumer protection and spam. Again, they could be treated like ISPs, which often come under gen-eral business regulation that applies to all commercial entities

— or at least a certain group or “class” of companies.

4.5.2 Using Licensing To Encourage Broadband Deployment

Regulators can use licensing frameworks to provide incen-tives for network deployment by large-scale operators, espe-cially in early stages of market liberalization. This often works particularly well with respect to the deployment of broadband access technologies.13

The intent of this approach should be to encourage opera-tors to deploy networks that may otherwise not be considered commercially viable or may create less value than other options. The incentives can either be rewards for meeting licence com-mitments or (where they can be enforced) financial penalties for failing to meet agreed commitments.

Such licensing incentives could take several forms:

• Extension of licence periods;

• Access to other operators’ infra struc ture;

• Allowing the provision of other, more lucrative services under the same licence;

• Access to universal access/service funds;

• Reduced licence fees;

• Tax incentives, including reduction of taxes and duties for both operators and end users; and

• Financial penalties for failing to meet licence commit-ments.

For example, the right to serve rural areas could be bun-dled with more lucrative locations in licence tenders.14 In addi-tion, multiple services can be bundled under one licence or authorization. In 2005, for example, Brazil announced that it would issue licences enabling operators to offer the “triple play” of voice, internet and broadcast services. Such an incentive framework has been used to encourage network deployment in other countries, as well. For example, in Hungary, the incen-tive was possible extension of a licence period (See Box 4.4).

Along with incentives, regulators can also threaten to revoke licences if commitments are not met. There is, how-ever, a natural tendency on the part of regulators not to revoke licences, especially if doing so would cause disruption to sig-nificant numbers of subscribers.

Also, in addition to providing tax incentives such as tax “holidays” or tax concessions to operators, reducing the burden of taxes and duties on equipment can lower costs for end users, encouraging wider adoption and usage. A recent study commis-sioned by the GSM Association indicated that in many devel-oping countries, up to 20 per cent of the total cost of mobile telephony stems from taxes and duties. Reducing or abolishing these would allow more people to afford services.

To date, most incentive mechanisms have been applied only to traditional local telephony (PSTN) services, but there is no reason why they could not also be applied to encourage deployment of broadband access networks. For example, new entrants (or even existing operators) could be offered appropri-ate incentives to deploy all types of broadband access technolo-gies, especially in rural areas. Given the relative success of such incentives for PSTN deployment, they could have a positive impact on broadband deployment, especially if they were avail-able to all industry players.

Box 4.4: Licensing Incentives for Network Deployment, selected examples

Hungary

In February 1994, Hungary was divided into 54 franchise areas for local tele com mu ni cation access. The incumbent and new entrants were invited to submit bids for each franchise area to act as the monopoly local access service provider until January 2002.

The licence conditions required each operator to achieve annual growth in local access lines of 15.5 per cent, per year. By January 1997, licensees also had to fulfil 90 per cent of customer demand for new local access lines within six months.

In 1993, teledensity was 14.5 per cent (14.5 per 100 inhabitants). This had increased to 26 per cent by year-end 1996 and to 36.1 per cent by year-end 2002. After 2002, all service providers with significant market power were required to provide local loop unbundling to other service providers, including new entrants (including a special Reference Interconnection Offer or RIO for unbundled local loops).

Brazil

Following the deregulation of the tele com mu ni cation sector in 1997, operators that met their universal service obligations were allowed to acquire additional licences, including mobile and long distance service authorizations.

In early 2004, the regulator Anatel certified that Brasil Telecom had met its universal service targets. This allowed the opera-tor to roll out mobile and long-distance services in addition to the local service it was already offering in the southeastern part of Brazil.

Meanwhile, Brazil’s other two major landline operators, Tele Norte Leste and Spain’s Telefonica Internacional, had already met their targets and were offering wireless phone and long distance services in other regions by 2004.

CHAPTER 4


60

On the other hand, using exclusivity periods as an incen-tive mechanism poses the danger of “crowding out” potentially more efficient new entrants and new investment sources. As long as licensing frameworks are appropriate, market liberali-zation generally will yield greater benefits than the exclusivity approach (often coupled with build-out obligations) adopted in many markets – both developing and developed.

4.5.3 Technology NeutralityIn principle, broadband regulation should be technologi-

cally neutral, and licensing and authorization regimes should reflect this. Increasingly, licensing focuses on generally encour-aging the construction of – and investment in – broadband access networks, rather than defining the specific method of delivering broadband services. This principle is particularly relevant to spectrum licensing for broadband wireless services, but it also can apply to wire-line deployment, giving licensees the flexibility to use copper, fibre or hybrid networks.

Technology- and service-neutral licences and authoriza-tions also enable broadband providers to offer a full range of services (including the “triple play” of voice, internet and video) in rural areas, increasing revenue stream options. In Venezuela, for example, rural licences allow operators to offer mobile and multimedia services in addition to fixed access, long-distance and international services. India and Uganda have allowed operators to provide both fixed and mobile services under the same licence, leading to increased competition and subscrib-ers as well as lower prices for consumers. Hong Kong, China is issuing unified licences for broadband wireless access pro-viders, allowing them to adapt to technological developments. Today, the licences allow broadband wireless access providers to offer fixed services, but as BWA technology develops, the same licences can be used to offer mobile broadband wireless services, as well.

Pure technology neutrality in licensing may be quite dif-ficult to achieve. For example, some wireless technologies and services are specifically reliant on, and stand ardized for, certain radio frequencies. But regulators are increasingly providing licensees with the maximum flexibility possible to select which technologies they wish to adopt, within approved standards and international frameworks.

Such flexibility may encourage broadband access deploy-ment in marginal areas and developing countries by allowing licensees to select the delivery technology that most minimizes costs and accelerates financial return. Operators can custom-ize the components of their network infra struc ture to suit the particular service offerings and technical requirements of their business plans. This will allow them to leverage whatever existing economies of scale they have been able to achieve in other, possibly adjacent, markets. Such flexibility is a hallmark of India’s unified access service licensing framework, which gives operators a choice of using either GSM or CDMA within their assigned spectrum blocks (See Box 4.5).

Some countries specify the use of particular technolo-gies (often made by specific companies) as a tool for industry development or trade policy. But mandating use of a specific technology within a defined spectrum block – or for particular

kind of service – may not result in the most efficient alloca-tion of spectrum or the most rapid deployment of that serv-ice. For example, if the required technology turns out not to be the optimal one, in terms of cost or supported functional-ity, market take-up is likely to be disappointing and spectrum usage will be limited.

This does not imply, however, that technical standards or international radio frequency allocations are unimpor-tant. Employing a common technical standard provides many advantages, including:• Economies of scale (both in terms of network and end-

user equipment);• Commitments and support from large-scale vendors;• More consistent road maps for product evolution and

development;• More inter-operability;• Reduced consumer switching costs; and• International roaming capabilities.

Standards also provide agreed-upon best practices that can drive more efficient usage of spectrum and energy/power

– two particularly important elements in resource-scarce areas. These advantages have often accelerated deployment of tele-com mu ni cation services above and beyond what could have been expected if technologies had remained fragmented.

Nevertheless, accepting the benefits of standards does not mean that regulators should specify which standards should be used in any specific spectrum allocation or service. Nor does it necessarily mean that regulators should limit operators’ tech-nology selections to internationally stand ardized technologies. The regulatory framework could provide licensees with the flexibility to select the appropriate technology for their circum-stances in order to encourage the deployment of broadband access infra struc ture. There is, of course, the risk that licensees may choose to deploy non-standard, unique and proprietary technologies, but given the substantial benefits of internation-ally recognized technologies, this is unlikely in most cases.

4.6 Alternative Approaches to Broadband Deploy-ment Regulators should not limit themselves to considering

only current wire-line telephony and mobile network operators as potential broadband service providers. New market entrants can be allowed, and even encouraged, to enter the broadband sector, either by building or converting their existing infra-struc tures into broadband platforms, or by taking advantage of unbundling and infra struc ture sharing.

4.6.1 Alternative Broadband Platforms Aside from broadband wireless and fixed-line networks

of major telephony providers, there are other platforms that can be used to provide broadband voice and data transmission and internet access. As discussed in Chapters 2 and 3, cable TV systems, satellites and electrical power grids all offer cur-rent and potential alternatives. Cable modem technology has already demonstrated its role as a broadband access technology

CHAPTER 4 61


– in fact, it remains a major market leader in North America. Satellite broadband access remains limited, to date, but it nev-ertheless is considered a viable option for developing countries. Broadband over power line (BPL) is often discussed, although the extent of its viability has yet to be proven in developing countries, where electrical grid infra struc ture may be as scarce as wire-line networks..

In more developed countries, a primary driver of broad-band take-up is competition between telephony (PSTN) oper-ators and cable system operators (for example, in Hong Kong, China; Korea (Rep.); the United Kingdom and the United States). In many of these countries, in fact, cable systems were the first movers in the broadband consumer market, prompt-ing tardy incumbent telephone operators to respond with xDSL. Furthermore, in the absence of large-scale local loop unbundling access, the only substantial competitive threat to telephone operators still comes from operators of alternative access networks – primarily cable TV systems.

Cable infra struc tures are more limited in developing countries although they have been deployed in a number of upper middle-income economies (e.g. Hungary), some lower middle-income countries (e.g. Thailand) and even in some low-income economies (e.g. India). In these markets, however, cable is present mostly in urban areas and even there, mostly in higher-income neighbourhoods. Fundamentally, cable TV is perceived as a high-income market. It is also a relatively expen-sive technology to deploy, given the infra struc ture require-ments. Options for lowering costs (for example, using poles rather than burying lines) often expose systems to risks from theft or damage. In some countries, cable systems suffer from

poor QoS and are not robust enough to offer reliable broad-band access. But this has not deterred some operators, where they have infra struc ture in place, from offering cable modem services.

Considering its lower cost structure, satellite broadband delivery has attracted attention, although market take-up remains limited. Satellite services are increasingly offered, how-ever, in many developed countries, including Canada, Ireland, the United Kingdom and United States. Satellite operators such as Eutelsat, Hughes Network Systems and Shin Satellite have been aggressively promoting satellite broadband solutions. Box 4.6 provides examples of satellite broadband deployments in developing countries.

The primary advantage of satellite service provision is that it avoids the high cost of terrestrial backbone infra struc ture. Furthermore, the VSAT terminals required to access the sat-ellite broadband service can use battery or solar power, elimi-nating the need for connection to the power grid. This allows rapid provision of broadband access to all areas of a country, assuming that the satellite footprint is sufficient and there is sufficient transponder capacity.

The key drawback of satellite platforms, however, remains limited space-segment capacity. New satellites are expensive to design, build and launch, and transponder lease prices can be similarly expensive as a result. By contrast, small VSAT dishes have declined in cost substantially in recent years.

Similarly, although power-line broadband technol-ogy promises to support the delivery of tele com mu ni cations services over electrical infra struc ture (with only minor modi-fications), BPL remains largely in the trial stage. The use of

Box 4.5: Unifi ed Licensing Frameworks, selected examples

India

In 2004, The Telecom Regulatory Authority of India (TRAI) established a unified licensing regime for local access networks, in response to an increasing overlap between GSM mobile services and CDMA “limited mobility” offerings. Now in its first phase, the Unified Access Service Licensing (UASL) framework covers all basic and cellular services. In the next phase, the government plans to implement a fully unified licensing regime covering all tele com mu ni cation services.

This single-licence approach reduces the financial burden on operators, in terms of licence fees and registration charges, for offering multiple services. The UASL represents a significant lowering of entry barriers for new and smaller market players, potentially translating into lower prices for end-users.1

Nigeria

In February 2005, the Nigerian Communications Commission proposed reviewing Nigeria’s tele com mu ni cations regulatory framework, with the goal of establishing a unified licensing regime. Under the new regime, new and existing fixed wireless and mobile licensees will be allowed to provide both fixed and mobile services, subject to geographical/regional limitations stated in their licences.

With the removal of fixed-mobile differentiation, licensees will be free to offer voice, data or multimedia services as they deem fit, once spectrum is allocated.2

1 See ITU Trends 2004, Chapter 5. TRAI press release August 2004. http://www.trai.gov.in/Newpressrelease.pdf

2 The Nigerian Communications Commission, 2005. http://www.ncc.gov.ng/Headlines/REPORT%20ON%20POST%20EXCLUSIVITY.doc

CHAPTER 4


62

existing electrical power infra struc ture, however, could reduce the substantial investment requirements to build or upgrade tele com mu ni cation networks. This potential has attracted a lot of attention as a mechanism for providing an alternative fixed local loop with broadband capabilities. Of course, as already mentioned, BPL is intrinsically dependent on existing elec-trical infra struc ture, which is notably absent in large areas of developing countries.

4.6.2 Wholesale Provision and UnbundlingThe deployment of broadband access service has, to a

certain extent, been encouraged in more developed coun-tries through various forms of local loop wholesale products, including full local loop unbundling and line resale. These products allow new entrants to access end-users without sink-ing the investments required to build local loop infra struc ture. The various forms of local loop wholesale access are discussed in Box 4.7.

Success in providing wholesale local loop products has varied in different countries, but in Hong Kong, China, the United States and Germany, unbundled local loops now account for a significant proportion of direct exchange (or access) lines. The establishment of appropriate unbundled loop access prices, at levels that permit incumbents to retain a mod-erate margin, has encouraged alternative operators to employ unbundling to provide broadband services to end-users. This has played a significant role in creating demand for broadband access.

In developed countries, the local loop has commonly been constructed before market liberalization. Moreover, the incum-bent usually has had a significant period of time as a monopo-list to recover the upfront infra struc ture costs. So in developed countries, unbundling has been less a tool to promote further network build-out or teledensity than a technique to encour-age service competition. So developed countries have been able to price unbundled local loop access based on incremental costs of offering a wholesale service, rather than historical costs of building the loop infra struc ture. The situation in developing countries is fundamentally different for the following reasons:1 Many developing countries are liberalizing primarily to

achieve network expansion and higher teledensity – not to increase service competition;

2 Liberalization is occurring much earlier in the develop-ment chain – without landline networks being fully con-structed;

3 Many incumbent operators may not have had sufficient time to recover the costs of local loop network deploy-ment that has incurred, and therefore, pricing based on incremental cost methodologies may not be appropriate;

4 Use of fully allocated cost approaches may have greater justification and validity for loop access pricing in devel-oping countries, but these will result in higher wholesale prices (not least because they may reflect the inefficien-cies and historic network structures of the incumbent), reducing the attractiveness of local loop wholesale access to alternative operators.

In addition, regulators in developing countries may lack ad min is trative expertise and resources to implement and regu-late a wholesale market structure. Even in developed countries, regulators have struggled to establish a local loop wholesale framework that encourages the development of market com-petition. Many incumbents have successfully impeded whole-sale access by setting up technical or limitations to co-location, slowing down processes for provisioning wholesale loops or pricing loops at levels that make a mockery of alternative oper-ators’ business cases.

Not surprisingly, few regulatory authorities in developing countries have opted to build broadband deployment strate-gies around the provision of wholesale local loop products by incumbent operators. Those that have opted to implement local loop unbundling recognize that its success rests on the ability to enforce the associated requirements. Although the provision of wholesale local loop products remains very limited in developing countries, a few developing markets are planning and implementing local loop access mandates. In September 2005, the members of the West African Telecommunication Regulators Assembly (WATRA) agreed to a set of regulatory guidelines that includes support for bit stream access.15

The provision of wholesale local loop products, including local loop unbundling, will meet with greater success in pro-moting the deployment of broadband access networks where certain regulatory and commercial conditions are in place. These success factors include:• The existence of an extensive and well-developed incum-

bent network;• Clear and complete regulations that spell out all unbun-

dling requirements to ensure that strong operators do not impede access to their exchanges; and

Box 4.6: Satellite Broadband in Developing Countries, selected examples

Algeria: Provision of satellite broadband to enterprises and public organizations in rural areas.

Ethiopia: Schoolnet is a programme funded by the United Nations Development Programme (UNDP) to provide satellite broadband to 400 schools.

Thailand: Commercial deployment of satellite broadband services and delivery to schools.

Uganda: Trial provision of satellite broadband services to number of rural schools.

CHAPTER 4 63


• Conditions that encourage continuing investment by both incumbents and new operators in new infra struc ture roll-outs.

Regulators may decide to end wholesale local loop requirements once new operators achieve an appropriate level of commercial scale. Regulators can then place more emphasis on frameworks that encourage network deployment.

4.7 Competition and Industry Regulation

This section explores how regulators can address the vari-ous issues that arise as broadband markets evolve and mature. As with telephony markets, broadband markets may undergo transitions from monopoly or near-monopoly structures to greater competition. Moreover, they evolve in the context of existing and allied tele com mu ni cation markets, such as voice telephony.

4.7.1 Backbone and International Connectivity Issues

Multiple network bottlenecks can occur in the provision of broadband services, both by incumbents and new market

entrants. These bottlenecks often stem from either a lack of needed infra struc ture, or from monopoly ownership of all infra struc ture by a single operator – typically, the incumbent. For example, in some markets, incumbents have used their ownership of backbone infra struc ture to impose constraints on the ability of new entrants to compete by:

• Imposing excessive access/leasing costs – In some markets, for example, incumbents levy higher leasing charges on com-peting tele com mu ni cation operators than on their corpo-rate customers;

• Imposing service limitations – Some network operations try to limit the amount of bandwidth or QoS they will sup-port; and

• Imposing restrictions on points of inter con nection – Some incum-bents set technical restrictions or limit access to interna-tional switches.

In a number of markets, control of backbone and interna-tional connectivity (whether by incumbents or other monop-oly operators) has been used to manipulate and constrain the development of competition. This can lead to artificial short-ages of bandwidth and inflated prices, thereby hampering the provision of robust global tele com mu ni cation services. Due

Box 4.7: Local Loop Wholesale Options

There are three main types of local loop wholesale access:

Local loop unbundling allows access seekers to have management control over the copper pairs connecting a subscriber to the incumbent’s main distribution frame (MDF). The access seeker can provide both voice and data services on the incumbent’s network.

Shared access refers to an arrangement where competitive service providers have access to either voice or data transmission over the incumbent’s network. The access seeker leases part of the copper pair spectrum while the incumbent maintains control of the copper pair.

Bit stream (or wholesale) access involves the incumbent installing high-speed access links to its customers and opening these links to competitors. In this case, the access seeker has no management control over the physical line and is not allowed to add any equipment to the network.

Total Service Resale allows an alternative provider to purchase the network operator’s service at a wholesale discount, rebrand the service, and resell it to the consumer, allowing the alternative provider to build a customer base and obtain a retail sales margin over the wholesale rate.

Box 4.8: Local Loop Unbundling in Poland

In February 2005, the Polish Office of Telecommunications and Post Regulation issued a directive requiring the incumbent tele-com mu ni cation operator, Telekomunikacja Polska (TP), to give other operators access to its local loops. The move was part of a plan to further liberalize the Polish tele com mu ni cation market.

In the first phase of local loop unbundling, TP will provide both full (voice and data) and shared (data) access to competing operators. There also are plans to extend the local loop unbundling offer to include bit stream access. The regulator is still reviewing the cost model submitted by TP, but local loop prices are expected to be based partly on benchmarking against other European providers operating in competitive markets.1

1 See http://europa.eu.int/information_society/topics/telecoms/implementation/annual_report/7report/slides281101

CHAPTER 4


64

to such concerns, the regulator in Singapore, for example, has moved to open up access to submarine cable landing stations. There are two main options available to regulators to head off the adverse consequences of monopoly control of backbone infra struc ture:

1 Impose a tighter regulatory framework on owners of bot-tleneck facilities to ensure that other operators can access such backbone infra struc ture at an appropriate pricing point; and/or

2 Encourage existing or new licensees to deploy alternative infra struc ture.

Regulatory intervention is often required where the costs and timelines associated with duplicating bottleneck facilities would be so excessive as to be commercially nonviable. Some countries now permit new entrants to install VSATs with inter-national access. Some also allow local broadband providers to connect directly with international backbone networks, rather than terminating traffic through an incumbent’s international gateway. But in many developing countries the incumbent still controls access to international network infra struc ture and is able to use this control to impose excessive prices on other operators, undercutting competition. So regulators are increasingly stepping in to ensure that new entrants gain fair and competitive access to existing backbone infra struc ture. In addition, regulators are finding that reliance on international connectivity can be reduced by developing internet Exchange Points and local caching, which the government can actively encourage or establish.16

It is also necessary to encourage the construction of sup-porting backbone networks, especially if the intent is to deploy infra struc ture into rural areas. However, such requirements are more likely to be effective if they are not so prohibitive as to deter market entry. With respect to encouraging licen-sees to deploy alternative infra struc ture and, furthermore, to deploy infra struc ture into areas previously not accessed by tele-com mu ni cation networks, there are a number of options and approaches for regulators to consider:

• Facilitating access to existing tele com mu ni cations infra struc ture used for alternative activities—For example, railway signalling or pipeline monitoring activities involve communications links, which can be made available to licensed tele com mu-ni cation operators.

• Ensuring and facilitating access to government land, including railways, electrical grids and road networks – Governments can streamline and stand ardize the application process for access to rights of way and ensure just and reasonable fee structures.

• Ensuring that tele com mu ni cation networks are incorporated into new infra struc ture developments – Governments can include broadband network conduits in road projects or incor-porate cable arrays in new electrical grids (for example, Chile’s ICT project side-stepped difficulties of geographi-cal isolation and infra struc ture shortage in rural areas by taking advantage of electrical plant).

• Creating broadband alliances to pool financial and other resources and to enhance negotiating power with network vendors – The

Wireless Broadband Alliance gives operators in the United States, the United Kingdom and the Asia-Pacific region benefits of scale in areas such as testing of products and services, influencing development and adoption of tech-nology standards and negotiating international roaming agreements.

Experience from some more developed markets (for example, Australia and the United Kingdom) shows that back-bone networks can be more quickly established by leveraging existing infra struc ture. For example, cables can be strung across electrical pylons that can also act as the location of radio anten-nas for wireless technologies. Other alternative infra struc tures include main roads, gas and oil pipelines, and water channels (especially maintained canals).

The key remaining issue, however, is what to do where such supporting infra struc ture does not exist. The lack of basic support infra struc ture (electrical grids, railways and pipelines) in rural areas is often heightened by large distances or rugged terrain. In such circumstances, the regulatory framework can promote the use of wireless technologies – satellite trunking

– especially where traffic is not expected to be substantial. This can be achieved through the expeditious allocation of unused or little-used spectrum and by lowering licence fees for trunk-ing in remote areas.

Developing countries may also explore the pooling of resources to collectively launch satellites that provide broader regional service, or to back a commercial satellite operator. Shin Satellite of Thailand, for example, has just launched its Ipstar satellite, which has a wide footprint across the Asia-Pacific region. Although it is a commercial operator, Shin has extended broadband access to rural areas of Thailand, Laos, Cambodia and Myanmar.

4.7.2 Funding Broadband DeploymentThe need to deploy infra struc ture into more marginal

geographies is based on the recognition that without such access, the Digital Divide will continue and grow – both between developing and developed countries and between urban and rural areas. Given the benefits of broadband deploy-ment, including cheap voice communication, regulators are playing a critical role in seeking to reduce this divide through the promotion of broadband access deployment.

The first step regulators are taking is to develop regula-tory regimes that are conducive to investment and supportive of commercially viable broadband access network deployment. In other words, they are specifically addressing the market effi-ciency gap in broadband provision. The need to first address the market efficiency gap in promoting ICT development was explored fully in the 2003 edition of this report. Efforts to close the market efficiency gap in basic tele com mu ni cations can also apply to the broadband service market. Regulators should remove unnecessary regulatory burdens and encourage market demand for broadband deployment, allowing market forces to promote broadband access wherever possible. It also can include facilitating market entry by small enterprises and micro-entrepreneurs, as well NGOs, libraries and local gov-ernments. In addition, regulators can encourage large-scale

CHAPTER 4 65


operators to deploy broadband networks in areas viewed as not commercially viable in return for access to potentially more value-creating business activities or other incentives.

It may also be necessary, however, for regulators to estab-lish mechanisms to fund network deployments, especially if there is evidence that regulatory incentives and lower-cost

network alternatives will not be enough in certain areas. This shortfall between market-based measures and universal access is termed the true access gap. In any given area, this true access gap (which represents, in effect, the failure of the market to deliver needed services) can only be determined after all attempts to address the market efficiency gap have fallen short.

Box 4.9: Removing Network Bottlenecks in India

The Indian tele com mu ni cation regulator, TRAI, is planning to introduce more competition into the market for international private leased circuits (IPLCs), which are currently “bottleneck” facilities. This will have implications for international telecom services like International long distance (ILD), as well as internet access and broadband data offerings.

There is a significant lack of competition in the IPLC market at present, forcing smaller operators out of the market for several services relying on leased line capacity. The reasons for lack of competition include:

• A limited number of landing stations, which are owned and controlled by a small number of operators;

• The ability of IPLC providers – who are also Internet Service Providers (ISP) – to charge other ISPs prices that are significantly above costs, harming competition the ISP market;

• The ability of IPLC providers also to charge substantial prices to ILD resale operators, with whom they also compete in the ILD market.

In response to this, TRAI has undertaken, or is planning, the following initiatives:

• Setting a tariff ceiling for various transmission capacities, based on each incumbent’s costs, and removing differences in IPLC prices based on usage volumes by ILD operators and ISPs.

• Reviewing the need to permit IPLC resale, which had been banned because it was believed that resale would retard construction of ILD infra struc ture in the country.

• Recognizing the need to facilitate access to cable landing stations by new service providers, as well as by new international cable carriers.

• Planning to facilitate mutual sharing of landing station infra struc ture international cable capacities among the carri-ers.1

1 TRAI, 2004, Fixation of Ceiling Tariff for International Leased Line Circuit (Half Circuit). http://www.trai.gov.in/consultation%20paper-30th%20april%202004.pdf

Box 4.10: Using the Indian Rail and Gas Facilities for Backbone Connectivity

Railtel Corporation of India was set up in 2000 to exploit communication assets lying idle along India’s rail network. Since then, the company has laid 25,000 km of fibre optic cables along rail lines. Railtel provides leased lines to tele com mu ni cation service providers, along with other infra struc ture like tower space and co-location services. In addition, it is also an ISP, operating a network of internet kiosks set up at railway stations.

The company plans to open an additional 300 cybercafes at railway stations, providing services such as VoIP and video con-ferencing for local people with no other access to computer equipment or broadband service. Owing to India’s extensive railway network, this allows broadband access to be rapidly extended into many marginal areas.1

Similarly, Gailtel, the tele com mu ni cation services arm of the largest gas transmission company in India, operates as an inte-grated tele com mu ni cation infra struc ture provider. The company started leasing bandwidth to tele com mu ni cation operators like Bharti and Tata in mid-2001. It also operates as an ISP, serving corporate and residential customers.

The company has laid an optic fibre/co-axial (OFC) hybrid network along about 8,000 km of natural gas and LPG pipelines, and it plans to extend the network to 18,000 km around the country. The network currently serves 73 cities across eight states. Due to infrastructure cost savings derived from overlaying the OFC network on existing pipelines, Gailtel is able to offer broadband services to its customers at substantially lower costs, compared with its competitors.2

1 http://www.railtelindia.com/

2 http://www.convergenceplus.com/apr03%20expert%20view%2002.html

CHAPTER 4


66

Box 4.11: GSR 2003 Universal Access Best Practice Guidelines

During the 2203 Global Symposium for Regulators, delegates adopted a broad manifesto for enabling universal access to telecommunication networks and services, including broadband ones. The best practice guidelines are as follows:

“We, the regulators participating in the 2003 Global Symposium for Regulators, have identified and propose the following best practice guidelines to achieving universal access to information and communication technology (ICT) services.

A An enabling regulatory environment: the role of governments and regulators

The success of any universal access/service policy is dependent upon political support at the highest level that recognizes the role of ICTs as a tool for development.

1) It is essential that Regulators exist or be established where they do not yet exist, and that their key role in implementing universal access policies and promoting competition be recognized and reinforced.

2) A series of policy and regulatory reform measures can be taken to achieve universal access to ICTs. These include:

a) Formulating a national policy that identifies appropriate and realistic universal access/service objectives that take into account the differences between universal access-public access to ICTs-and universal service-household or private access to ICTs.

b) Including all citizens, regardless of gender, ethnicity, socio-economic level or geographic location, in national universal access/service objectives.

c) Reviewing universal access/service policies, regulations and practices periodically to adapt to the evolving nature of ICT services and the needs of end users.

d) Conducting periodic public consultations to the extent possible with stakeholders to identify their needs and modify accordingly universal access policies, regulation and practices.

e) Designing universal access policies, regulations and practices in order to create incentives for the private sector to extend universal access to communications services.

f) Establishing a fair and transparent tele com mu ni cation regulatory framework that promotes universal access to ICTs.

g) Adopting technologically neutral licensing practices enabling service providers to use the most cost-effective technology to provide services for end users.

h) Adopting a framework of inter con nection rates linked to costs.

i) Reducing regulatory burdens to lower the costs of providing services to end users.

j) Developing an effective regulatory body responsible for implementing policies directed towards assuring the best quality reliable services at the most affordable prices that meet the needs of consumers-existing and future.

k) Promoting competition in the provision of a full range of ICT services to increase access, affordability, availability and use of ICTs.

3) Countries can use regulatory reform as the first step in achieving universal access, recognizing that further steps may be necessary to achieve ubiquitous access to ICTs, e.g., in rural areas or to users with special needs.

4) Appropriate licensing schemes for rural service providers could be granted to meet the needs of un-served and under-served areas.

B Access to information and communication infra struc tures

1) The lessons learned from the initial experiences developing countries have achieved with mobile cellular services can be applied to a broader range of ICT services to foster universal access. These lessons include providing services in a competi-tive framework, using new technologies that offer both innovative services and affordable pricing options (e.g., pay as you go options such as pre paid cards) to a wide range of end users.

2) Other measures to promote affordable ICT equipment could include national manufacturing of ICT equipment, reduced customs tariffs and duties, and end-user loans to foster affordability of ICT equipment.

3) A full range of public access options can be developed, including the creation of public telecentres.

4) Local input (including the content useful for local populations) into projects increases their long-term financial sustain-ability.

5) Educating local people on the benefits of ICTs and their use increases their long-term financial sustainability.

CHAPTER 4 67


When all inefficiencies have been leached out of the market through sector reforms, and the true access gap remains – only then may government intervention be necessary. This interven-tion can take the form of targeted and limited “smart” subsidies to spur the deployment of broadband access networks to areas and populations that would otherwise not be reached.

Regulators have several ways to address the true access gap, including:• Licensing special rural operators to deploy broadband access net-

works in defined geographies – Licensees can be selected based on bidding for the minimal subsidy required to achieve specified targets. By licensing such rural operators within specific areas, regulators can “leapfrog” the gradual diffu-sion of new technologies from urban to rural geographies;

• Providing funding for local community initiatives to provide broadband access. Many universal access fund programmes take a top-down approach, directing fund ad min is trators (often regulators) to identify the communities for which targeted subsidies will be made available. But a bottom-up approach could also be used, allowing communities them-selves to apply for funds to deploy their own broadband access networks. This would help to ensure local commu-nity involvement in, and demand for, broadband access.

• Giving Direct and indirect financial support in return for the deployment of broadband access networks. Governments could provide tax exemptions to operators that roll out tele com-mu ni cation infra struc ture in rural areas. Where this is insufficient to attract commercial operators, governments could offer full or partial subsidies. Alternatively, the gov-ernment could provide preferential loans to operators for building broadband access networks.

• Requiring operators to deploy broadband access networks. Again, broadband deployment mandates could be tied to funded mechanisms, drawn from government revenues or con-tributions made by all operators. Contributions could be either as a flat-rated or set as a percentage of revenues. The operator(s) responsible for the provision of universal broadband access in any given area would receive finan-cial incentives or payments for each new broadband access

line installed. But it could also be given the freedom to determine in which specific locations it would deploy such infra struc ture. Governments may prefer to provide funds only where the costs of providing broadband serv-ice would exceed revenues from that service.

Any government involvement in financially supporting the deployment of broadband access networks has potential drawbacks. For one thing, regulators should ensure that the allocated funds are used for the expressed purpose of deploy-ing infra struc ture in marginal areas. This requires the regula-tor to have the institutional capability to manage and oversee the allocation process and to provide maximum transparency. There are examples of such regulatory frameworks, including Korea (Rep.) and a number of South American countries (See Box 4.12).

Deployment of broadband access networks by large-scale network operators – even those employing wireless archi-tectures – may only be commercially viable in many areas of developing countries (and even a number of areas in developed countries) if an incentive framework is incorporated into the licensing approach, coupled with government funding and initiatives that generate customer demand. One approach is to use regulatory incentives to support and promote broadband access network deployment by large-scale operators. Another approach is to encourage small-scale players to serve local communities by removing regulatory burdens that often apply to large-scale operators, and allowing small players to test their business cases for broadband access and build demand. The key issue is deciding which approach has the best fit with the underlying regulatory and institutional capabilities, and which approach minimizes the extent of government involvement while maximizing the commitments of private investors.

4.7.3 Enforcing Broadband Market CompetitionTo ensure a reasonable level of competition in an emerg-

ing broadband market, it is necessary to establish a regulatory framework that prevents anti-competitive behaviour by opera-tors. Regulators need to monitor dominant operators, which can be defined as those that are capable of acting unilaterally in

Box 4.11: GSR 2003 Universal Access Best Practice Guidelines

C Guidelines in regard to finance and management of universal access policy

1) Universal service funds can be viewed as an option that complements regulatory reform and developed as a mechanism within a broader market-oriented approach to achieving universal access.

2) Universal service funds can be financed by a broad range of market players, managed by neutral bodies such as regulators, and be used to kick-start public access projects that meet the needs of the local community.

3) Governments may consider a full range of other financing mechanisms, including tax incentives for ICT providers and end users.

4) Competitive minimum subsidy auctions could be used, as an option, to reduce the amount of financing necessary for public access projects financed by a universal service fund.

5) Public access projects can be designed to achieve long-term financial self-sustainability, especially where consideration is given to innovative low-cost technologies

See: http://www.itu.int/ITU-D/treg

CHAPTER 4


68

the market (such as on pricing or provisioning terms), without regard to discipline from competitors, buyers or sellers. Such regulation need not be onerous or excessively restrictive, it just needs to be effective in prohibiting and punishing conduct that has the potential of preventing or constraining the develop-ment of market competition.

Examples of potential business conduct by dominant operators that should be prohibited include:

• Predatory pricing (providing services at less than cost);

• Mandatory product bundling (requiring end users to take products in which the operator is not dominant in order to access products in which the operator is dominant);

• Price discrimination (applying different prices and terms and conditions to favour or disadvantage particular cus-tomers); and

• Cross-subsidization (using profits generated in one serv-ice market, in which an operator is dominant, to subsidize its operations in a competitive market).

In response, obligations typically applied to dominant operators include:

• Requirements to provide inter con nection to competing operators on appropriate terms and conditions; and

• Obligations to ensure that tariff structures comply with regulatory requirements, including appropriate price con-trols.

All of these issues are relevant to the broadband access market. A dominant operator – usually the incumbent tel-ephone operator – can distort the development of the broad-band access market by undertaking behaviour to stifle or disrupt competition. Common techniques are predatory pricing and cross-subsidization, both of which can undercut smaller broadband access providers’ revenues and drive them from the market.

In general, developing broadband markets should be sub-ject to ex-ante rather than ex-post regulation, given the potential for anti-competitive conduct by dominant operators. Ex-post regulation tends to be work better when there is no dominant carrier, and where there is sufficient evidence that competitive market forces will function properly without extensive ex-ante rules. Given the current structure of most tele com mu ni cation markets – including broadband access markets – such competi-tive circumstances do not commonly exist.

Even within a regulatory framework that applies more rigorous scrutiny to dominant operators, it remains impor-tant to avoid excessive regulatory burdens. Regulations can be carefully tailored and minimized to address key concerns:

Box 4.12: Korea’s KII Project

The Korea Information Infrastructure (KII) Project was established in Korea (Rep.) in early 1995 as an avenue to promote nationwide broadband usage. The ultimate aim of the project was to provide broadband networks to 13.5 million subscribers with the average transmission rate of 20 Mbit/s by 2005.

To achieve this, the government supplied public funding to facilities providers to partially ease the burden of investment in access networks. Greater funding was reserved for remote areas. In addition, some enterprises were offered a tax exemption on deployment of broadband infra struc ture.

The government provided loans, at preferential rates, worth USD 77 million to facilities service providers in 1999. It com-mitted an additional USD 77 million in 2000 for the purpose of deploying infrastructure in less densely populated areas. Subse-quently, public funding was extended to cover infrastructure build-outs in rural areas, with additional investments amounting to approximately USD 900 million.1

1 Ministry of Information and Communication, Korea http://www.mic.go.kr

Box 4.13: Use of Universal Access Funds, selected examples

In 1995, the Chilean government established a Universal Access Fund to promote installation of public telephone systems in isolated rural areas. The USF employed competitive bidding to determine which provider should receive the subsidy in any area. The fund managed to secure approximately USD 60 million in investments, of which about 85 per cent came from private companies. With support from the fund, public telephones were provided to about 6,000 rural localities, reaching 2.2 million inhabitants within seven years. Reliance on market forces and minimal regulation, among other things, has been credited for the success of the scheme.

In Peru, universal access funds for the Fondo de Inversión en Telecommunicaciones (FITEL) are collected through a 1 per cent tax on the gross revenues of all public telecommunication companies. The funds are then allocated through public competi-tive bidding to operators deploying infrastructure in locations of priority social interest.

CHAPTER 4 69


(a) preventing anti-competitive business conduct, (b) ensur-ing compliance with licence commitments and (c) protecting end users. Strict price regulation practices, for example, can be replaced by a general price cap regime, with the goal of ending price regulation when the market is genuinely competitive. At the same time, market development of new services also may hasten the end of price regulation. The rise of new offerings for VoIP, which is usually not price-regulated, may force domi-nant operators to lower prices to retain market share.

Section 4.6 of this chapter has already detailed the role of competition between cable TV and tele com mu ni cations companies in driving competition and take-up in the broad-band access market. Therefore, cross-ownership of tele com-mu ni cations and cable network companies can serve as an impediment to broadband development. Integrated providers that own both telephone and cable TV networks are unlikely to deploy both DSL and cable modem services, because the two would cannibalize each other. While good for the opera-tor, deploying only one of the technologies would bring higher prices for consumers.

4.7.4 Other Regulatory ConcernsThis chapter has identified the need to regard broadband

access as an information delivery platform. As broadband take-up and usage widens, regulation of broadband services and applications will become an increasing concern for regulators. In view of the potential and growing demand for IP telephony, regulators in several countries have begun developing a regula-tory framework to address issues associated with VoIP. These key concerns are examined in detail in Chapter 6.

In addition, regulators will also need to consider regula-tory measures pertinent to any content delivery platform as broadband deployment takes off. Given the superior func tion-alities of broadband access, the need for regulating broadband may be even more pressing than for other tele com mu ni cation services. Content-related concerns include:• Having an appropriate framework to effectively guard

against intellectual property infringement;• Setting content regulation guidelines to protect consum-

ers, especially minors; and• Setting up appropriate and suitably non-invasive regula-

tory mechanisms to ensure that national security is not compromised.

4.8 Increasing Broadband AwarenessSo far, this chapter has focused on the role of regulators

in stimulating the supply side of the broadband access market – that is, encouraging network build-out. There are, however, significant opportunities for regulators to promote demand for broadband through the expansion and improvement of appli-cations and services available to end users.

4.8.1 The Government’s RoleGiven scarce capital resources in many developing coun-

tries, regulators (and appropriate ministries) often must demon-

strate the relevance and necessity of promoting the deployment of broadband access networks, especially if funding is required in some form. This requires a clear demonstration of the ben-efits of broadband access and the preparation of a holistic ICT strategy. Some countries have used e-government projects to initiate and support ICT projects, including broadband devel-opment. Such projects have several objectives, including:• To enhance the efficiency of government by converting

paper and manual transactions to online ones, simplifying and speeding up government processes;

• To improve government linkages with the population and give citizens more ready access to government services, officials, and information;

• To expand the reach, awareness and understanding of broadband access among the population – including the government’s own employees (the public sector is often the largest employer in developing countries); and

• To provide a framework to attract local and foreign invest-ments.

The ITU is working on a Global E-Government Project to enhance government services through the use of secure and trusted internet infra struc tures and applications in selected developing countries.17 The Vietnamese government, mean-while, is cooperating with the World Bank to implement an e-government plan, with the idea of promoting sustained take-up of broadband.

Within such projects, the provision of broadband access networks is just one component of a broader strategy to improve ICT positioning and provide a positive impact across the economy.

Telecommunication regulators are in a good position to initiate greater collaboration and cooperation with other gov-ernment agencies to promote broadband take-up. In addi-tion to e-government projects, which invite participation by multiple agencies, inter-governmental working groups can be established to facilitate the development of broadband infra-struc ture, particularly in commercially less viable areas where resources can be shared. For instance, the Singapore ONE initiative, which was officially launched in 1997, was designed to be a collaborative effort between government and indus-try to implement a nationwide broadband network. It was jointly supported and driven by multiple government agencies, including the tele com mu ni cations regulator, the Economic Development Board, the Media Development Authority and a research agency.

Regulators and government agencies can take the lead by actually building a broadband backbone network. The suc-cess of the Singapore ONE initiative was largely driven by the government’s construction and operation of a core Asynchro-nous Transfer Mode (ATM) backbone network, which enabled broadband access to be provided extensively to public librar-ies, schools and training centres across the island nation. This served to raise awareness of broadband and drove broadband take-up across various communities.

In order to really act as advocates for broadband access, however, regulators must be able to clearly and credibly dem-

CHAPTER 4


70

Box 4.14: Limiting Cross-Ownership in the EU

In 1999, the European Union issued a directive requiring the separation of telecommunications and cable TV operations into distinct legal entities. It was believed that cross-ownership of telecommunications and cable operations would prevent cable companies from providing low-priced voice telephony services in competition with telephone companies. In addition, without cable competition in the internet access market, there would be little incentive for telephone companies to upgrade their existing networks to full-scale broadband capability via xDSL.

Hence, under the directive, dominant public telecoms network operators were required to run their cable operations as sepa-rate legal entities. It was thought that this would prevent emergence of new, anti-competitive bottlenecks and would encourage competition and innovation in both the telephony and cable TV markets.

Box 4.15: E-Government in Vietnam

The Ministry of Posts and Telematics (MPT) in Vietnam has recently drafted a National ICT Master Plan for 2006-2010 and an E-Government Master Plan covering the same period. These initiatives will be implemented in cooperation with the World Bank. The ICT Master Plan aims to achieve:

(1) A wide diffusion of ICTs throughout Vietnam’s economy and society, making it a larger contributor to GDP;

(2) Establishment of a nationwide information and communications network; and

(3) Comprehensive ICT skills development.

As part of the process to improve e-readiness in Vietnam, the ICT project will address the following issues:

• Strengthening the technical and managerial capacity of the MPT in implementing the ICT initiatives;

• Facilitate increased access to tele com mu ni cations, in the context of a gradual move towards a more competitive market environment and private-sector participation;

• Promote greater awareness of ICTs and e-applications in the business community, with the view to encourage businesses to adopt e-commerce;

• Support enhanced government online presence and content at the national and municipal levels, through interactive and dynamic portals;

• Roll out e-government services to businesses in areas such as e-procurement and business or land registration; and

• Support extensive training and awareness-raising efforts to encourage diffusion of ICTs in the private sector.

onstrate its advantages. Indeed, the regulator can even take the lead in coordinating the provision of broadband access to various government ministries and departments. This is the case in Singapore, where the regulator, the Info-Communi-cations Development Authority (IDA) coordinates all ICT procurement for the government. As a further example, the Turkish Ministry of Health is partnering with ITU to develop

“e-healthcare” systems. Specific initiatives include giving health-care providers and citizens access to health-related information via broadband access, and development of a Primary Health-care Information Systems and Electronic Health Records ini-tiative.

Providing content that is relevant to local communities, particularly in local languages, can also increase the relevance and potential impact of broadband access networks. Morocco, for example, is in the early stages of developing Arabic content for both mobile and internet services. In Laos, a Lao-language version of Linux-based graphical desktop, along with Lao-lan-guage office tools, has been developed and provided to villages as part of a plan to promote WLANs.

Fundamentally, in order to be able to increase awareness of broadband access networks and to successfully position broadband at the forefront of a country’s tele com mu ni cation strategy, regulators need to be able to articulate the potential benefits, both direct and indirect, within a cost-justified frame-work. Furthermore, the rationale for broadband is likely to be stronger when contained within a broader and holistic ICT strategy (including policies to improve the rate of PC penetra-tion). This is the challenge for regulators, especially when there are so many other demands on a country’s tele com mu ni cation infra struc ture and scarce resources.

4.8.2 Promoting Broadband Equipment Take-Up

Promoting broadband access requires more than just access to broadband networks. People in developing countries also require access to personal computers or other end-user terminals – and this inevitably raises cost and affordability issues. Put simply, the question is whether there are low-cost terminals that are affordable to users in developing countries. What role can regulators play in improving the penetration of PCs and portable devices in the developing world? Without a

CHAPTER 4 71


solution to the problem of low PC penetration rates, the provi-sion of broadband access networks will have little impact.

One relevant option for most developing countries is to concentrate PC usage at single locations – whether these are cybercafés, community centres, schools or government offices. Often, cybercafés are the primary mode of accessing the inter-net in many developing countries (See Box 4.17).

There are, however, many alternative options for encourag-ing private PC take-up, some of which are detailed in Box 4.18.

4.9 Conclusion

Although basic telephone (PSTN) teledensity is extremely low in many countries, the rapid development of technologies

– especially wireless technologies – means that policy-makers and regulators should not lose the opportunity to plan beyond increasing basic teledensity. They should plan for broadband access, as well.

Regulators clearly face numerous challenges in the broad-band context. In particular, they face a perceived lack of local demand and available revenue streams for broadband in many countries. This could prevent the commercial deployment of broadband access networks in many areas, especially rural areas

– at least by large-scale network operators. Furthermore, con-tinued low PC penetration rates in many developing countries could effectively negate any potential positive impacts that may arise from broadband network deployment.

In response to these challenges, regulators also have many options. They can:

• Seek to maximize investment flows by liberalizing markets and permitting foreign ownership. This includes allowing broad-band providers to offer a full range of services and applica-tions, such as the “triple play” of voice, internet access and video/multimedia programming.

• Encourage the deployment of wireless broadband access networks by freeing up the requisite spectrum. This strategy can be aug-mented by a technology neutral approach to spectrum assignments.

• Create a regulatory framework that encourages a full range of potential broadband providers. Moving beyond large-scale national network operators, regulators can empower uni-versities and government offices, local communities and smaller entrepreneurs to deploy broadband access net-works. This may include tailoring regulatory frameworks to each group of potential broadband providers.

– A regulatory framework tailored to small broadband providers will enable and encourage local commu-nity providers to harness the potential of broadband technologies and enable greater broadband access in rural areas;

– Competitive large-scale operators can be encour-aged to extend their networks to rural areas through

Box 4.17: Installing Internet Centres in Southern Brazil

Sud Mennucci is a town with 7,500 inhabitants, located 700 km southeast of Sao Paulo. It will soon benefit from a plan to install two internet access centres, with 10 computers each, to provide internet access to low-income inhabitants.

Three solutions have been proposed to realize this plan:

(1) Have the federal government donate the computers for the centre;

(2) Incorporate the plan into the state government’s Acessa Sao Paulo programme, which aims to provide free internet access to dozens of municipalities in Sao Paulo state, or

(3) Engage private companies to set up the centre and provide training in the use of technology.1

1 LatinCom, 2005.

Box 4.16: The Rural-Enlaces Project in Chile

The Rural-Enlaces project is part of an ICT policy to improve education in 3,600 rural schools around Chile, reaching an estimated 130,000 students. The project involves providing learning aids, such as computers, as well as access to broadband tel-ecommunications.

The proponents of Rural-Enlaces are convinced that technology should be seen as a means to support existing pedagogical approaches to rural education, rather than as a cultural invasion of the dominant group. Hence, the project aims to involve various segments of local communities in the im ple men ta tion process.

For instance, the project is positioned as a professional development opportunity for local teachers. Apart from technological knowledge transfer, teachers benefit from being directly involved in the selection of educational software and content. In addition, they are consulted on the design of learning practices that are most appropriate and relevant to everyday reality in their communi-ties. Meanwhile, parents are also involved in helping out with various aspects of project im ple men ta tion at the local schools.

CHAPTER 4


72

infra struc ture-sharing arrangements that guarantee open access to all competitive operators;

– Competitive large-scale operators can be given incentives to deploy networks in return for appro-priate rewards;

– Regulators could seek to encourage the deployment of broadband access networks by providing direct, targeted subsidies from universal access funds or indirect financial benefits (such as tax exemptions) to a full range of broadband providers;

• Create an asymmetric regulatory regime to prevent the domi-nant operator (often the incumbent) from constraining the development of competition in the broadband access market;

• Work with other government agencies or ministries to develop ini-tiatives, such as e-government programmes, that generate demand for broadband services;

• Encourage the build-out of fibre backbone networks to boost the capability of both wire-line and wireless broadband technologies. These steps include forging synergies with transport and energy infra struc ture projects and providing

incentives for 2G mobile operators to replace their micro-wave links with fibre networks. It also means making it possible for all owners of such communications resources to lease unused capacity to others for commercial deploy-ment.

• Link broadband access development strategies to efforts to sup-port and promote PC take-up. Build government-sponsored PC kiosks and other access terminals, especially in areas where broadband networks are to be deployed.

Promoting broadband access in developing countries requires a new vision of reduced regulatory burdens, inno-vative incentives, and coordinated efforts by all links in the broadband value chain. It also requires concerted political will to achieve. It is an end to the “business as usual” approach, and governments should treat ICTs and broadband networks as tools for development. There are many challenges in the road ahead. Strategic and creative thinking to overcome these challenges is required, but with concerted efforts on all fronts, regulators can enable developing countries to join the broad-band world.

Box 4.18: Encouraging PC penetration, selected examples

In India, the government has launched the “Indian PC Programme,” which aims to improve PC penetration from the current 14 per 1,000 to 65 per 1,000 by 2008. Some of the initiatives include: (a) launching INR 9,999 (USD 230) “no compromise” PCs, subsidized by software vendors and chipmakers; (b) encouraging all incumbent operators to move towards a subscription model for offering broadband services and PCs as a package; (3) setting up loan schemes, employee provident funds and other saving funds to encourage PC adoption among government employees; and (4) amending the Income Tax Act to allow deduction of home PC purchases. Also, cybercafe kiosks have been set up along railway tracks around the country to provide computer access to rural villages. As a result, rural communities now have access to e-Government, tele-education and telemedicine services.

In Mexico, public internet kiosks were set up in community plazas under the e-Mexico project, which benefited 3,200 municipalities around the country. Each community plaza has an average of 10 computers and has internet access via satellite technologies.

In Sri Lanka, in cooperation with the World Bank, the Sri Lankan government is planning to set up tele-centres in rural areas around the country to improve community access to ICTs. The target groups include farmers, students and SMEs.

In Thailand, a “Computer ICT Programme” was launched in 2003 to provide low-cost computers. In addition, there is plan to establish a nationwide network of 751 tele-centres, located at various post offices throughout the country.

In Tunisia, the World Bank is working with the Tunisian government to subsidize 6,000 “Publinets” and 10,000 PCs under the “PC Familial” programme. There is also an investment plan to increase the number of PCs in schools.

In Uganda, Uconnect, an NGO based in Uganda, imports used computers from Europe and the United States, revamps them and supplies them to schools and organizations. About 100 mostly rural-based schools have benefited from this project.

CHAPTER 4 73


1 Large-scale operators include main national operators, new entrants and mobile operators.2 Marginal areas include remote, rural and under served areas.3 While some private companies have also deployed private networks, these remain largely in urban areas.4 By year-end 2004, some 112 countries had at least partially privatized their State-owned incumbent operator. Source: ITU World Telecommunication Regulatory

Database.5 ITU World Telecommunication Indicators Database.6 The African Virtual University was established in 1996 as part of a World Bank project. It was founded as a technology-based distance education network to bridge

the digital divide in Africa. http://www.avu.org/7 Universitas 21 is an international network of leading research-intensive universities, with the objective of facilitating collaboration and cooperation between member

countries, as well as to create entrepreneurial opportunities for them. Established in 1997, it now has 17 member universities in 9 countries. http://www.universitas21.com/

8 TRAI, 2004, Recommendations on Internet and Broadband. http://www.trai.gov.in/Recommendations on Internet and Broadband 2004-04-29 FINAL.pdf9 Yun et al., 2002, The Growth of Broadband Internet Connections in South Korea: contributing factors. http://www.ciaonet.org/wps/yuk01/10 TRAI, 2004, Recommendations on Internet and Broadband. http://www.trai.gov.in/Recommendations on Internet and Broadband 2004-04-29 FINAL.pdf11 Contribution of Ireland to GSR Consultation to identify best practice guidelines for spectrum management to promote broadband access (http://www.itu.int/ITU-D/

treg/Events/Seminars/2005/GSR05/consultation.html).12 The trend away from onerous licensing practices toward more flexible market-entry authorization approaches, including technology and service-neutral licensing and

authorizations, was explored fully in the 2004/05 edition of Trends in Telecommunication Reform: Licensing in an Era of Convergence (http://www.itu.int/publica-tions/).

13 See ITU Trends in Telecommunication Reform 2004: Licensing in an era of convergence.14 For a detailed discussion on this issue, See ITU Trends in Telecommunication Reform 2003: promoting universal access to ICTs.15 The full set of regulatory guidelines agreed by WATRA are available at http://www.itu.int/ITU-D/treg/Events/Seminars/ITU-EC-Project/Ghana/modules/Compil-

Guidelines_final_E.pdf16 The role of regulators in promoting IXPs is explored in the ITU-IDRC report Via Africa: Creating local and regional IXPs to save money and bandwidth, available

at http://www.itu.int/ITU-D/treg/17 For more information on this project see: http://web.itu.int/ITU-D/e-strategy/projects/E-Government/Executive-summary11.pdf

75


CHAPTER 5

5 BROADBAND SPECTRUM MANAGEMENT

This chapter discusses the challenges spectrum regula-tors face in allocating and assigning spectrum for broadband wireless access (BWA) and other wireless services. The main challenge is to provide for flexible, market-oriented spectrum licence rights, which can create a positive investment climate for BWA services. At the same time, regulators want to discour-age uneconomic hoarding and speculation in spectrum, which could delay the rollout of services to consumers.

Flexible spectrum rights should be granted as long as the spectrum licensees meet two absolute preconditions criti-cal to the development of communications markets. First, the licensee (or, in the case of unlicensed spectrum, the service provider) must increase competition, benefiting consumers. Second, licensees and service providers should experience the opportunity cost1 of using their spectrum allocations. This is the best way to ensure effective and efficient use of the spec-trum.

This chapter aims to explain these concepts and stimu-late thinking about how to make effective and pragmatic spec-trum management decisions, without falling into the kinds of dogmatic approaches that have often characterized spectrum management discussions.2 It is important to understand the primary goals of spectrum regulation, as well as the fundamen-tal economics of wireless access systems. It will also be useful to review some of the technological advances that are making new spectrum resources available for BWA services and applica-tions. The traditional regulatory spectrum management models can then be examined to determine whether they adequately address the challenges presented by these new technologies. Then, best practices can be identified to improve spectrum management in ways that encourage the rapid deployment of BWA systems.

5.1 Introduction

Broadband wireless services are poised to bring significant benefits to all parts of the world. These new services will pro-vide access to the internet and to IP-enabled services. Use of the term Broadband Wireless Access or BWA really describes the arrival of an era in which it is possible to enjoy the “internet everywhere, all the time.” Because many BWA applications will be ideal for consumers, this chapter will focus on frequencies

below 6 GHz, where the physical characteristics of the spec-trum are more conducive to consumer applications.

In this new era, the power of the Information Age to affect our lives will be exponentially multiplied by the freedom brought about by BWA networks. Wireless broadband technol-ogies will fuel the engines of the global economy by enabling consumers to: • Freely access the internet from the farm, the city, kiosks,

cybercafés, coffee shops, on moving trains, and in their own communities and backyards, in developed and devel-oping countries alike;

• Connect to the internet seamlessly, using a single device to make phone calls, access information and government services, vote, pay taxes and bills, and enjoy entertain-ment; and

• Live in enlightened communities that are connected to broadband, spectrum-based services that offer access to resources and opportunities never before available.

BWA networks may have even more impact on consum-ers when they are combined with other broadband platforms. Ultimately, the broadband world will feature BWA networks, with their mobility and portability, at the core of a variety of useful hybrid broadband architectures that will provide a rich, multimedia consumer experience at virtually any time or place.

Of course, the potential for BWA services to improve people’s lives ultimately depends on the amount of spectrum regulators make available. But for the first time, it appears that technological advances can increase spectrum capabilities and resources, allowing licensees to do more with the same spec-trum and enabling entirely new spectrum uses. Regulators must now consider whether traditional approaches to spec-trum management are sufficient to address the resulting chal-lenges and opportunities.

Spectrum regulators must also consider a growing number of spectrum management “best practices” that have become apparent over the last two or three decades. The best practices explored in this chapter have fostered the widespread deploy-ment of wireless services, including cellular mobile radio, broadcast television, paging, and satellite services. Following these practices has led to significant reductions in the cost of

Author: John Muleta, Venable LLP

CHAPTER 5


76

providing services and has created opportunities for entrepre-neurs to develop innovative applications for consumers.

5.2 The Economics of Broadband Wireless Access

It is important for regulators to understand something about the economics of wireless access, which are now – after two decades of wireless mobility – becoming much clearer. The principal economic drivers of wireless systems are the avail-ability and cost of spectrum, the cost of the end-user device, and the acquisition and maintenance of the subscriber (rang-ing from network management to billing and customer serv-ice operations). For a BWA system to be successful, it must be competitive across each of these categories. This demonstrates to operators and end users alike that it represents a viable value in the broadband marketplace, particularly when compared with wire-line or satellite alternatives that require significant capital expenditures to increase capacity or coverage area.

5.2.1 Spectrum as An Input

Spectrum use is defined by four parameters, of which only two – power and bandwidth – are usually set by regu-lators. These parameters determine the capacity and coverage that a particular spectrum band can deliver to an operator for the provision of services to end users. System capacity and cov-erage essentially determine the number, the size and the cost of the transmitters (including the supporting backhaul network) that an operator needs to deploy in order to deliver the desired set of services. These factors, in turn, establish the financial return threshold of the operator wishing to deploy a competi-tive BWA network.

Spectrum in the range below 3GHz has propagation characteristics that enable wide coverage areas and can more easily overcome interference from foliage, buildings and other obstructions using non-line-of-sight technologies such as beam forming. The ability to provide non-line-of-slight serv-ices reduces the number of base stations required to provide coverage in these bands. But experience drawn from mobile services indicates that an increasing customer density requires

more base stations to meet the growing demand for network capacity.

Ultimately, any viable BWA service will need to enhance both coverage and capacity, but initially it is more important to have greater coverage. It is unlikely that new BWA services will be in such immediate demand that networks will have to be designed to maximize capacity over coverage. Over time, as consumer adoption increases, BWA systems will begin to emphasize throughput and capacity by reducing cell sizes and increasing the number of base stations.

There are a number of spectrum bands operating between 0.4 GHz and 5.5 GHz that could help foster the growth of BWA, although each band represents necessary tradeoffs between capacity and coverage. In identifying these bands for new uses such as BWA, the general trend across the globe has been to reclaim the bands from incumbent licensees that have, or will have, the ability to deploy more spectrally efficient equipment that reduces their need for spectrum.

One category of reclaimed spectrum comes from govern-ment and military operations (1.5-2.4 GHz as well as 5.1-5.8 GHz). In the commercial context, satellite systems (2.0-2.3 GHz) and fixed microwave systems (3.1-3.7 GHz and 2.1-2.2 GHz) provide an avenue for reclaiming spectrum. The increasing spectral efficiency of the latest technologies, plus the widespread availability of substitute technologies and services (for example, submarine cables and fiber-optic networks) has reduced the demand for spectrum needed to deploy these serv-ices. Similarly, broadcast television bands – including the UHF (400-700 MHz) band – and Multi-channel Multipoint Distri-bution Service (MMDS) bands (2.5 -2.7 GHz) have also been the source for new BWA spectrum in various markets, because newer, more spectrally efficient digital television broadcast standards have been developed.

It is relatively easy to identify bands for new uses such as BWA, especially in the context of global harmonization trends that can provide clues to regulators about the most likely future use of a spectrum band. The difficulty lies in transferring these bands from an incumbent licensee to a new one. This paper seeks to help address these transitional issues in a pragmatic

Figure 5.1: From Silos to LayersThe ultimate consumer broadband experience will be based on a multi- platform IP network with BWA as its core.

CHAPTER 5 77


fashion by creating marketplace incentives that encourage licensees to transfer spectrum to its best and highest use, while at the same time offering services that are competitive for the particular dynamics and context of each country (whether rural or urban, developed or developing markets).

5.2.2 End User Costs and DevicesEnd-user take-up of wireless services is largely a function

of the cost of the end-user device and of the applications that work on that device. Common sense says that the lower the costs of the handset, the more likely consumers are to adopt the underlying service.

Over the last 20 years – but increasingly over the last five years – the mobile marketplace has demonstrated that con-sumer adoption in developed markets has accelerated when device costs have been below the USD 200 barrier. Although

this threshold amount might not be practical for developing markets, it is important to understand that mass adoption of end-user devices in developed countries has the positive scale effects that can easily translate to lower costs for end-user devices distributed in developing economies. Additionally, wireless equipment manufacturers have also learned from the mobile market and are now offering end-user devices with limited functionality to developing country markets in order to further accelerate the manufacturing scale effect that reduces the overall cost of making these devices.

5.2.3 Service Delivery and Management

Reducing the cost and complexity of applications, and the associated service delivery mechanisms, that work on wire-less devices has a significant effect on consumer adoption and leads to accelerating the scale economies of manufacturing the

Box 5.1: Defi ning Broadband Wireless Access

Internationally, the ITU considers broadband wireless access (BWA) to encompass mobile or fixed access technologies that provide connections at speeds higher than the primary rate (e.g. 2 Mbit/s). This encompasses technologies within the IMT-2000 family, as well as newer technologies such as WiMax and WiBro. Nonetheless, it is likely that each country will continue to make its own decision about the definition of BWA given its technological and economic development stage.

Singapore’s regulatory authority describes wireless broadband as “an access technology that offers high-speed data access over the air. A wireless broadband network, typically operating at frequency bands less than 6 GHz, provides broadband speeds ranging from 256 kbit/s to tens of Mbit/s. Each base station generally serves an area of up to several square kilometres. Wireless broadband networks can deliver network connectivity to fixed locations using standards like IEEE 802.16d, and in the near future, to mobile users using standards like IEEE 802.16e and IEEE 802.20.”

Kenya’s Communications Commission defines broadband fixed wireless access “as intentional radiators that use wideband digital modulation techniques and provide a wide array of high data-rate fixed communications for individuals, businesses, and institutions.”

Mauritius uses a three-part definition for BWA, in accordance with ITU-R Recommendations:

• Wireless Access systems are broadband radio systems that may be deployed either indoors or outdoors. These systems in-clude:

– Fixed wireless access which may be defined as “Wireless access application[s] in which the location of the end-user termination and the network access point to be connected to the end-user are fixed.”

– Mobile Wireless Access which may be defined as “Wireless access application[s] in which the location of the end-user termination is mobile.”

– Nomadic Wireless Access which may be defined as “Wireless access application[s] in which the location of the end-user termination may be in different places but it must be stationary while in use.”

Box 5.2: Spectrum Bands for BWA

450 – 500 MHz (Regions 2 and 3) and 600-1000 MHz (Region 1) – These could be used for broadband mobile access services, given the significantly enhanced propagation characteristics of the band. These bands are under discussion as candidates to be included as IMT-2000 bands during the 2007 ITU World Radiocommunication Conference (WRC).

1.5-2.5 GHz – Many of the bands in this range are currently identified as IMT-2000 bands. They are also under discussion, across all regions, as part of the 2007 WRC process. They could be used for both fixed BWA and broadband mobile services some-times referred to as “IMT-2000 and beyond” (that is, “3.5G” and “4G” technologies). These bands are also identified for hybrid satellite and terrestrial wireless systems that permit seamless continental roaming for BWA operators.

3.4-3.7 GHz – Allocated across all regions for licensed BWA services.

5.1-5.7 GHz – These are allocated across all regions for unlicensed BWA uses.

CHAPTER 5


78

devices. In all geographic markets, the advent of flat-rated voice services and the use of prepaid pricing plans have increased the rate of adoption of wireless mobile devices and led to sig-nificant economies of scale. Additionally, the drive to add IP services to wireless devices will increase the rate of consumer adoption by making wireless access at the service layer in dis tin-guish able from wired alternatives.

The impact of the IP on BWA service delivery can be best understood by looking at the IEEE 802.16 WiMAX standards development process currently under way. The key innova-tion of the 802.16 process has been to simplify the commu-nications protocol stack, so that economies of scale can take place in developing radios and the associated chipsets – the most expensive elements – while providing greater freedom for developing applications at the service layer.

5.3 The Technology Revolution

In addition to understanding the economics of spectrum-based services, regulators should also have a working under-standing of the revolutionary changes in technology – many of them driven by micro-electronics – that are permanently alter-ing radio systems and spectrum engineering.

5.3.1 Applied Information Theory in Radio System Design

Spectrum represents the temporal and spatial oppor-tunities to transmit information using the electromagnetic spectrum. The range of frequencies in the electromagnetic spectrum is typically divided into eight bands, spanning from 3 Hertz to 300 GHz.

The characteristics of signal propagation depend on the frequency band on which the signal is transmitted. These sig-nals are typically transmitted by an antenna device that trans-mits energy in one or multiple directions. Shannon’s capacity theorem, the fundamental theorem of radio communications design, states that the rate of information transfer by a radio is limited by the available bandwidth and the ratio of signal to noise within the band. Within this physical limit, the spectrum resources available to a radio are determined by four factors: (1) specified bandwidth, (2) the allowable power or energy emission within the band, (3) the bit error rate acceptable to the end user and (4) the throughput desired by the consumer.

Most regulators have, up until now, only defined the power limits and the bandwidth and left the other two fac-tors to be determined by the marketplace. Traditional radio systems were designed using analogue equipment and had limited computational power available to them. The results were radios that operated using very narrow throughput and bit error rate (that is, quality-of-service demands) parameters and that were highly sensitive to the operating environment. Traditional radios were not flexible and could not be used to take on new tasks or operate in new environments. This made them unsuitable for broadband networks.

5.3.2 Advances in Microprocessor Technology

The advent of miniaturized and powerful comput-ing resources available through digital signal processors (DSPs), non-programmable hardware computing components and field programmable gate arrays (FPGAs) has made it possible to create radio systems that can dynamically change along all four vectors (bandwidth, power, throughput, and bit error rate) that define spectrum. By using powerful microprocessors to dynamically change the four variables, radio system designers can now create new spectrum capabilities where none had pre-viously existed (see Figure 5.3). By incorporating digital and microprocessor technologies into the design fabric of radio sys-tems, engineers are now creating radios that can dynamically operate outside the constraints of a particular intersection of bandwidth and power limits normally set by regulators.

For example, a radio designed to optimally perform at design PB (in the middle of the circle in Figure 5.3) can now be redesigned, on the fly, to operate within any of the five possible regions surrounding the optimal point. Each of the five regions represents a trade-off between power, bandwidth, throughput and bit error rate. The light areas (those below the design line) represent increased performance in the form of either throughput or bit error rate, in exchange for reduced power and bandwidth. In contrast, the darker areas above the design line provide for increased power and bandwidth utiliza-tion but poorer performance in terms of throughput and bit error rate. Computing resources enable radios to make these impromptu tradeoffs, increasing the flexibility of these sys-tems to handle different types of market environments without having to change radios.

The net result is that consumer radios are becoming more flexible and highly adaptable. From a spectrum regula-tor’s perspective, the additional flexibility and adaptability of the technologies means that new spectrum resources must be accounted for and usage rights must be assigned.

5.4 Adapting Spectrum Regulatory Models for BWA

Keeping in mind the issues of spectrum economics and technological advances, spectrum managers in the new broad-band era face the challenge of achieving three separate but interrelated goals, simultaneously:(1) To provide the proper incentives for spectrum licensees,

both existing and new, to invest in broadband services;(2) To expand consumer choices by enabling sustainable com-

petition for similar services across multiple technological platforms; and

(3) Implementing policies that discourage wasteful and anti-competitive behaviour resulting from uneconomic specu-lation and hoarding of spectrum.

5.4.1 Existing Regulatory Models

There are essentially three conventional methods of man-aging spectrum to attempt to achieve these goals. First, there is the command and control model, in which strict operating parameters and service rules define licensees’ spectrum rights.

CHAPTER 5 79


An alternative licensing model is the exclusive rights model, in which a licensee is given rights – which may be (within limits) transferable and flexible – to use a specified spectrum band within a defined geographic area and during a fixed period of time. In the current understanding of the exclusive rights model, spectrum use rules are primarily technical (as opposed to service-based), because they are designed only to protect the spectrum licensee and adjacent spectrum users from generat-ing or receiving harmful interference – not to mold or develop a certain service or market structure.

The third regulatory model is the commons model, or unli-censed model, which allows unlimited numbers of users to share a block of frequencies without giving any one user or group of users priority or individualized rights of use. Uses are limited only by technical criteria that specify bandwidth and

emitted power but provide no enforceable rights to protect against interference. A well-known form of commons approach has been the deployment of WLANs using Wi-Fi technology.

Having defined these models, it must be said that none of them is adequate by themselves to address today’s rap-idly evolving world of broadband spectrum. The command and control approach grants spectrum rights on such narrow grounds that they are often of limited utility for broadband opportunities. Plus, the model requires constant government intervention, with attendant delays and hassles, to change the operating rules of the licences. The exclusive use model has compelling arguments for providing market incentives to new entrants. But it also creates perverse incentives for incumbent licensees to engage in speculative or anti-competitive hoard-ing of spectrum, as a way to thwart real or perceived competi-

Figure 5.2: From VLF to EHFThe ultimate consumer broadband experience will be based on a multi- platform IP network with BWA as its core.

Frequency Range Description3 to 30 Khz Very low frequency band (VLF)30-300 KHz Low frequency band (LF)

300 KHz-3 MHz Medium frequency band (MF)3 MHz-30 MHz High frequency band (HF)

30 MHz-300 MHz Very high frequency band (VHF)300 MHz-3 GHz Ultra high frequency band (UHF)3 GHz-30 GHz Super high frequency band (SHF)

> 30 GHz Extremely high frequency band (EHF)

Figure 5.3: Changing the Technology ParadigmNew technologies permit new spectrum uses to be created by trading off power or bandwidth with throughput and bit error rate dynamically on the same radio.

CHAPTER 5


80

tion. Finally, the commons approach – which today provides low entry barriers and can result in significant scaling in terms of consumer adoption – can become self-limiting through the lack of enforcement mechanisms to manage and prevent over-crowding and overuse.

Rather than adopting any of these models in isolation, regulators can adapt and combine them in ways that are appro-priate to the circumstances. Above all, regulators can condition the granting of new spectrum resources or rights upon demon-strations that they will be used to increase overall competition for broadband services across all platforms. Moreover, spec-trum rights or access can be granted to entities that prove they are capable and willing to use them economically, because they are willing to experience the real and quantifiable opportunity costs of these spectrum rights. This pragmatic approach insures that market forces are harnessed to achieve the goal of deploy-ing BWA systems as rapidly and as efficiently as possible.

5.4.2 Re-Examining the Models

At this juncture, regulators may wish to review the exist-ing models to see how they might be updated or adapted to address the newer technological and market realities of the BWA era. The following subsections consider each model in turn.

5.4.2.1 The Command-and-Control Model

Some experts view the command-and-control model as simply a means to exercise tight government control over spectrum use. A more honest assessment, however, is that the approach grew out of the realities of conventional radio design during the last century. Essentially, those realities called for four steps, all of them bureaucratic: (1) allocation, (2) enactment of service rules, (3) assignment (licensing), and (4) enforcement of the rules and license requirements.

The regulator’s task is never done under a command-and-control regime. Regulators must continually revisit and referee the spectrum environment as new radios are introduced into the marketplace. In order for a new radio system to enter the market, the regulator must address everything from system configuration, co-channel and adjacent channel effects, power flux density, coding, out-of-band emissions, and innumerable other technical criteria. Of course, the cost of the regulator’s involvement is the time it takes to transfer spectrum rights to their best possible uses. Given the rapid rate of technological change, the time lags associated with government allocation processes amount to a significant drag on the ability of licen-sees to rapidly deploy new BWA systems.

5.4.2.2 The Exclusive Use Model

In essence, this model gives licensees the right to use their spectrum however they see fit, as long as they follow technical rules that are designed to protect the licensee from causing or receiving interference. In its purest form, the model confers something like a fee-simple “ownership” of spectrum (with technical covenants) for a set period of time. Licensees face few or no restrictions on marketing their spectrum rights, includ-ing through secondary market trading of the usage rights.

Proponents of the exclusive licensing approach tout the economic incentives it provides for licensees to seek the most productive and profitable use of spectrum. Detractors, however, fear that the model could lead to hoarding or under-investment by incumbent providers that seek to raise costs for competitors who need spectrum as an input to their own offerings. With-out competition, there may be few incentives for incumbents to develop new services or more spectrum-efficient systems. Under the exclusive use model, incumbents left to their own devices can simply buy up spectrum rights, with no guaran-tee they will use those resources to advance innovative and

Box 5.3: Software Defi ned Radios, Adaptive Array Systems and Mesh Networks

The availability of computing resources for radio signal processing has enabled three exciting advances in radio technology: software defined radios (SDRs), Adaptive Array Systems (AASs) and Mesh Networks.

SDRs are essentially radios that can be re-configured and adapted at the point of use and for different applications. This results in a multi-band, multipurpose radio. In the ideal software radio scenario, the radio signal is directly converted to digital signals at the antenna. All other radio functions are performed in the digital domain by software on the host platform, which might be a flexible digital signal processing (DSP) chip, a computer or even a mobile telephone.

Unlike conventional antennas, where the energy is diffuse, AASs use computational algorithms to direct energy to parallel and simultaneous channels within the same frequency bands. This exciting technology uses computing to combine an array of antennas and radio frequency energy in order to detect and calculate radio signals on a highly refined basis. This allows the system to suppress interfering signals and automatically track desired signals. The result is a significant increase in overall system capacity by enabling greater re-use of the same radio frequencies.

Mesh networks were developed through military communications technology research. Also known as ad hoc or “infra struc-ture-free” networks, they are designed to maintain high quality-of-service in unstructured or harsh spectrum environments. Mesh networks eliminate the need for a spoke and hub radio network and allow one subscriber to communicate with another without requiring synchronization from the base station (See Box 3.5 in Chapter 3). Because mesh networks lack any common infra struc-ture, they represent new spectrum capabilities that are ungovernable from a central point, making them the latest challenge to regulators’ ability to manage and allocate spectrum use rights.

CHAPTER 5 81


competitive wireless systems (including BWA networks) in the marketplace.

5.4.2.3 The Spectrum Commons Model

The benefit of the commons model is that it removes the input costs of obtaining spectrum from the economic equation of market entry. Service providers can begin service at lower costs, then scale up rapidly. End users face lower service prices and can more easily afford handsets and terminal devices. Reg-ulators wanting to jumpstart BWA deployment might easily select the commons approach, especially in rural and under-developed areas where factors such as low population den-sity and lack of backhaul infra struc ture might otherwise limit broadband service deployment and take-up.

Of course, the commons model has risks. Regulators must consider the potential long-term effects of reducing entry barriers and eliminating licensing. Commons models hold the keys to their own undoing, because the very success in rapidly deploying systems and proliferating users can lead to interfer-ence, crowding and an unstable spectrum environment. This overcrowding phenomenon is more commonly known as the

“tragedy of the commons.” Regulators can mitigate the risk by implementing power limits, modulation requirements, back-off schemes, and other measures. Of course, imposing new technical rules by regulatory fiat might effectively convert the commons approach into something more closely resembling the classical command-and-control model, defeating the very basis for establishing a commons model in the first place.

Some regulators are using a mix of licensed and unli-censed spectrum to address the need for low-cost broadband services in rural areas. Ireland, for example, allows small opera-tors to launch services in rural areas using unlicensed spectrum. When the operators establish a successful business case, they can migrate to licensed spectrum (See Box 5.4).

5.4.3 A Flexible Approach for New Times In the broadband context, the fundamental future chal-

lenge for spectrum regulators is how to efficiently and effec-tively distribute new spectrum resources that technology is making available. A pragmatic model, unencumbered by any particular spectrum theology but able to draw from them all, seems the optimal way to meet this complex challenge.

This pragmatic approach takes into account the fact that technology is creating new spectrum capabilities and resources faster than most regulators can redistribute spectrum rights using traditional means. The modern spectrum regulator needs a practical, outcome-oriented policy framework. The desired result, in this case, is rapid deployment of broadband serv-ices, not a grand theory about the rights involved in spectrum use. The problem with the traditional spectrum management models is that they focus on defining usage rights of spectrum licensees without articulating how to help achieve the underly-ing policy goals of the regulator.

Today’s broadband markets feature both high levels of spectrum licence incumbency and low levels of inter-modal competition. A pragmatic approach that rewards economic risk-taking by spectrum holders will reduce the likelihood that they

will “warehouse” spectrum simply for its perceived scarcity value.

So what constitutes a pragmatic approach? Regulators can start by offering to grant spectrum holders maximum flexibil-ity for their spectrum rights on the condition they meet two threshold obligations:

(1) They must demonstrate, even before gaining any new spectrum rights, their commitment to increasing inter-modal broadband competition; and

(2) They must agree to license conditions that positively enforce the opportunity cost of their newly allocated spec-trum rights.

5.4.3.1 Defining Flexible Spectrum Rights

In this pragmatic model, flexibility is defined as giving licensees enough freedom to respond to market signals and demand, within technical parameters designed to avoid harm-ful interference with adjacent licensees. As the universe of wireless services expand, licensees must use these newfound capabilities to provide new capabilities that will be demanded from broadband networks. These capabilities include:

• Providing either portable or mobile services to increase the personalization of communication services and enhance societal and individual productivity;

• Achieving spectral efficiency and overall network effi-ciency to create economic returns on investments;

• Reducing the cost of the customer equipment (handsets and terminals) in order to encourage rapid and widespread consumer acceptance; and

• Enabling integration and convergence with other plat-forms to provide seamless connectivity.

With the advent of new technologies, licensees can make tradeoffs between power, bandwidth, throughput and bit error rate, in order to provide viable offerings in the broadband marketplace. Furthermore, flexible spectrum rights will allow licensees to apportion and share spectrum access with others, facilitating the availability of BWA services and increasing com-petition.

5.4.3.2 Creating Competitive Broadband Markets

A practical approach to BWA spectrum licensing calls for granting spectrum licensees not only technical flexibility to create more spectrum capabilities and resources, but also opera-tional autonomy to enter new lines of business whenever tech-nology allows them to. Regulators should grant autonomy to enter allied or new markets as long as doing so would enhance competition and would enable the licensee to make the fullest, most efficient economic use of the spectrum.

If the spectrum regulator simply grants technical and operational flexibility, there is no guarantee that the licensee will choose to deploy services that will add to competition. But granting flexibility with a concomitant obligation to provide competitive inter-modal broadband services would encourage licensees to enter and aggressively compete in emerging mar-kets, such as the BWA market.

CHAPTER 5


82

5.4.3.3 Enforcing Opportunity Costs

In addition to promoting competition, flexibility should be harnessed to enforce on the licensees the opportunity cost of using spectrum. Generally, the goal here is to give licensees price signals about the value of their spectrum holdings that discourage them from engaging in uneconomic hoarding of spectrum. Although regulators can combat hoarding by simply recapturing spectrum, there are other, more productive mecha-nisms for enforcing the opportunity cost of spectrum.

One of the more popular methods in the last decade has been the use of auctions. Unfortunately, the effectiveness of auctions has diminished as a result of the growth of the wire-less industry and the availability of large amounts of capital. This has desensitized the industry to price signals coming from standard auctions. Furthermore, in some cases auctions have been used to increase national revenues rather than as a tool for enforcing market discipline. This has bred some unsustainable auction results, creating uncertainty in the markets. Nonethe-less, adjusting for these two factors, transparent auction proc-esses are a viable method of enforcing the opportunity cost of using spectrum.

In May 2005, for example, the Info-Communications Development Authority (IDA) of Singapore successfully auc-tioned spectrum in the 2.3 GHz and 2.5 GHz frequency bands for broadband wireless access services. The starting price for each of the spectrum blocks put up for auction was SGD 1,000, and the highest closing price bid was SGD 550,000. IDA decided to grant successful bidders a 10-year licence, in order to provide investment certainty.

Singapore’s distribution of spectrum for BWA services was conducted in open and transparent fashion. IDA had earmarked the 2.3 and 2.5 GHz bands for wireless broadband services in February 2004. The following April, IDA launched a public consultation proceeding on spectrum allocation and the licens-ing framework for wireless broadband services. IDA released licensing details for broadband wireless services in February 2005, notifying interested parties that it would hold an auction if demand exceeded the supply of available spectrum.

Enabling secondary markets for trading spectrum rights also has the effect of enforcing the opportunity cost of spec-trum. Allowing the rapid transfer of spectrum rights between private parties that value these rights differently creates price signals that encourage licensees to use the spectrum to provide competitive BWA services, because these uses will be the most valued in the current marketplace. It is important to remember that BWA services are the core component of a general set of broadband services that consumers would utilize. Providing the marketplace the flexibility to combine BWA services with other platforms is more likely to increase consumer welfare than by restricting such combinations.

Additionally, with secondary markets for spectrum rights gaining greater acceptance among licensees and regulators, experts in the field are now considering how to combine auc-tions and secondary markets to create new mechanisms that rapidly drive spectrum to its highest and best use. Two-sided auctions, for example, let regulators and spectrum incumbents combine their spectrum resources into a simultaneous auction that transparently recalibrates both the geographic and tech-

CHAPTER 5


82

Box 5.4: Eire’s Response to BWA

Ireland contributed several “principles” to the 2005 GSR Best Practice Guidelines on Spectrum Management to Promote Broadband Access. Here are excerpts, discussed as “Principle Two” and “Principle Three.”

Principle Two: Balancing the Use of Licensed and Licence-Exempt Spectrum.

A number of local initiatives have taken effect to provide broadband access using licence-exempt spectrum. In Ireland, from July 2002, wideband data transmission systems for the provision of fixed wireless access networks/metropolitan area networks (FWA/MAN) have been permitted in the 5.8 GHz (5725 – 5875 MHz) band on a licence-exempt basis, provided that the maximum radiated power does not exceed 2W eirp. This higher power level, over and above the current European harmonized standard, has increased the coverage achievable and hence the utility of the 5.8 GHz band.

This initiative provided some impetus for small market players to enter the market at very low cost, to gain some experience of broadband provision and to test-market demand for various broadband services. A number of successful operations using the licence-exempt spectrum, having proved their business case, have now moved to licensed spectrum. [Irish regulator] ComReg has committed itself to continue to identify appropriate spectrum allocations, both licensed and licence-exempt, for Wireless Access Services that are supported by choice and availability of equipment.

Principle Three: Access to Cost Effective Backhaul Infrastructure.

Just as consumers in semi-rural or rural areas may not have access to ADSL, the providers of wireless broadband are hampered by the lack of cost-effective backhaul infra struc ture, e.g., fibre. The alternatives such as satellite or point-to-point wireless fixed links are significantly more expensive compared to the costs of providing a wireless base station for broadband access.

In Ireland, consideration is being given to permitting the use of point-to-point links within the broadband access spectrum to provide a cost effective backhaul operation. While this is difficult to accomplish from a spectrum management viewpoint, it is seen as a viable alternative to the traditional and more expensive alternatives.

Ireland’s contribution to the 2005 GSR Best Practice Guidelines is available at http://www.itu.int/ITU-D/treg/Events/Seminars/2005/GSR05/consultation.html.

CHAPTER 5 83


nical limits on spectrum rights. Two-sided auctions are being developed as a way of smoothly restructuring bands to allow new, innovative services like BWA.

Another more traditional, but equally efficient, method of enforcing opportunity cost is to impose build-out or construc-tion obligations on licensees. Although these build-out obli-gations are effective in imposing costs, they tend to be blunt regulatory instruments because they are conditioned on prior assumptions about marketplace conditions. But combined with secondary spectrum markets, these obligations can serve a valuable role in dissuading licensees from hoarding spectrum.

In shared license bands, the sharing rules developed by regulators are the best method of enforcing the opportunity cost of using the spectrum. These rules generally determine the level of barriers to entering the shared bands, the amount of interference permitted between and among users of the band, and the power levels permitted (and therefore the cover-age range). These key parameters all help to define the oppor-tunity cost of using the spectrum. In the case of unlicensed bands, low power limits combined with the lack of interference management circumscribe the use of the bands and encourage a high degree of efficiency.

5.5 Defining Best Practices

Spectrum regulators need also to look at a number of spectrum management best practices developed over the last two or three decades as tools that can be used to encourage BWA deployment. This section explores several versions of best practices, beginning with those specifically endorsed by the world’s regulators, as members of the Global Symposium for Regulators (GSR).

5.5.1 The GSR Guidelines

Recognizing that spectrum is a scarce resource that needs to be managed effectively and efficiently, the delegates to the 2005 GSR, held in Yasmine Hammamet, Tunisia, drafted a set of best practice guidelines for spectrum management to promote broadband access. The 2005 guidelines continue the tradition of best practices agreed to at the GSR conferences in 2003 and 2004 on promotion of universal access, and low-cost broadband services, respectively.2 The 2005 guidelines are reprinted here in full:

“We, the regulators participating in the 2005 Global Sym-posium for Regulators, have identified the following:

1. Facilitate deployment of innovative broadband technologies: Regulators are encouraged to adopt poli-cies to promote innovative services and technologies. Such polices may include:

• Managing spectrum in the public interest.

• Promoting innovation and the introduction of new radio applications and technologies.

• Reducing or removing unnecessary restrictions on spectrum use.

• Adopting harmonized frequency plans defined by ITU-R recommendation in order to facilitate the im ple-men ta tion of competition.

• Embracing the principle of minimum necessary regula-tion, where possible, to reduce or eliminate regulatory barriers to spectrum access, including simplified licence and authorization procedures for the use of spectrum resources.

• Allocating frequencies in a manner to facilitate entry into the market of new competitors.

• Ensuring that broadband wireless operators have as wide a choice as possible of the spectrum they may access, and releasing spectrum to the market as soon as possible.

2. Promote transparency: Regulators are encouraged to adopt transparent and non-discriminatory spectrum man-agement policies to ensure adequate availability of spec-trum, provide regulatory certainty and to promote invest-ment. These policies may include:

• Carrying out public consultations on spectrum man-agement policies and procedures to allow interested parties to participate in the decision-making process, such as:

public consultations before changing national fre-quency allocation plans; and

public consultations on spectrum management deci-sions likely to affect service providers.

• Implementing a stable decision-making process that provides certainty that the grant of radio spectrum is done in accordance with principles of openness, trans-parency, objectivity – based on a clear and publicly available set of criterion which is published on the regulator’s website –and non-discrimination and that such grants will not be changed by the regulator with-out good cause.

• Publication of forecasts of spectrum usage and alloca-tion needs, in particular on the regulator’s website.

• Publication of frequency allocation plans, including frequencies available for wireless broadband access, in particular on the regulator’s website.

• Publication of a web-based register that gives an over-view of assigned spectrum rights, vacant spectrum, and licence-free spectrum, balancing any concerns for con-fidential business information or public security.

• Clearly defining and publishing radio frequency spec-trum users’ rights and obligations, including on the regulator’s website.

• Clearly defining and publishing licensing and authori-zation rules and procedures, including on the regula-tor’s website.

• Publication of legal requirements for imported equip-ment and foreign investment, in particular on the rel-evant government agency website.

CHAPTER 5 83


CHAPTER 5


84

3. Embrace technology neutrality. To maximize innova-tion, create conditions for the development of broadband services, reduce investment risks and stimulate competi-tion among different technologies, regulators can give industry the freedom and flexibility to deploy their choice of technologies and decide on the most appropriate tech-nology in their commercial interest rather than regula-tors specifying the types of technologies to be deployed, or making spectrum available for a preferred broadband application, taking into consideration the need for and cost of inter op erable platforms. • Regulators can take into consideration technological

convergence, facilitating spectrum use for both fixed and mobile services, ensuring that similar services are not subject to disparate regulatory treatment.

• Regulators can provide technical guidelines on ways to mitigate inter-operator interference.

• Regulators can ensure that bands are not allocated for the exclusive use of particular services and that spec-trum allocations are free of technology and service con-straints as far as possible.

4. Adopt flexible use measures: Regulators are encour-aged to adopt flexible measures for the use of spectrum for wireless broadband services. Such measures may include:• Minimizing barriers to entry and providing incentives

for small market players by allowing broadband suppli-ers to begin operations on a small scale at very low cost, without imposing onerous rollout and coverage condi-tions, to enable small market players to gain experience in broadband provision and to test market demand for various broadband services.

• Recognizing that wireless broadband services may be used for both commercial and non-commercial uses (e.g., for community initiatives or public and social purposes) and that broadband wireless spectrum can be allocated for non-commercial uses with lower regula-tory burdens, such as reduced, minimal or no spectrum fees; regulators can also allocate and assign spectrum for community or non-commercial use of broadband wireless services.

• Recognizing through flexible licensing mechanisms that wireless broadband technologies can provide a full range of converged services.

• Adopting lighter regulatory approaches in rural and less congested areas, such as flexible regulation of power levels, the use of specialized antennas, the use of simple authorizations, the use of geographic licensing areas, lower spectrum fees and secondary markets in rural areas.

• Recognizing that in markets where spectrum scarcity is an issue, the introduction of mechanisms such as sec-ondary markets can in some cases foster innovation and free up spectrum for broadband use.

• Recognizing the role that both non-licensed (or licence-exempt) and licensed spectrum can play in the pro-motion of broadband services, balancing the desire to

foster innovation with the need to control congestion and interference. One measure that could be envisaged is, for example, to allow small operators to start opera-tions using licence-exempt spectrum, and then move to licensed spectrum when the business case is proved.

• The promotion of shared-use bands, as long as inter-ference is controlled. Spectrum sharing can be imple-mented on the basis of geography, time or frequency separation.

• Developing strategies and implement mechanisms for clearing bands for new services as appropriate.

• Recognizing the need for cost-effective backhaul infra-struc ture from rural and semi-rural areas, regulators can consider the use of point-to-point links within other bands, in line with national frequency plans, including any bands for broadband wireless access.

5. Ensure affordability. Regulators can apply reasonable spectrum fees for wireless broadband technologies to foster the provision of innovative broadband services at affordable prices, and minimize unreasonable costs that are barriers to entry. Higher costs of access to spectrum further reduce the economic viability in rural and under-served areas. Auctions and tender processes can also be managed to meet these goals.

6. Optimize spectrum availability on a timely basis. Regulators are encouraged to provide effective and timely spectrum use and equipment authorizations to facilitate the deployment and inter op erability of infra struc ture for wireless broadband networks. Regulators are also encour-aged to make all available spectrum bands for offer, sub-ject to overall national ICT master-plans, in order that prices are not pushed up due to restrictive supply and limited amount of spectrum made available and so that opportunities to use new and emerging technologies can be accommodated in a timely manner. In addition, special research or test authorizations could be issued to promote the development of innovative wireless technologies.

7. Manage spectrum efficiently. Spectrum planning is necessary to achieve efficient and effective spectrum management on both a short-term and long-term basis. Spectrum can be allocated in an economic and efficient manner, and by relying on market forces, economic incen-tives and technical innovations. Regulators can promote advanced, spectrum-efficient technologies that allow co-existence with other radio communications services, using interference mitigation techniques (for example, dynamic frequency selection). Regulators can provide swift and effective enforcement of spectrum management policies and regulations.

8. Ensure a level playing field. To prevent spectrum hoarding, especially by incumbents, regulators can set a limit on the maximum amount of spectrum that each operator can obtain.

9. Harmonize international and regional practices and standards. Regulators can, as far as practicable, harmonize effective domestic and international spectrum practices and utilize regional and international standards whenever

CHAPTER 5


84

CHAPTER 5 85


CHAPTER 5 85


possible, and where appropriate, reflect them in national standards, balancing harmonization goals with flexibility measures. This could include harmonization of spectrum for broadband wireless access that could generate econo-mies of scale in the production and manufacture of equip-ment and network infra struc ture. Likewise, global harmo-nization of standards to ensure inter op erability between different vendor’s user terminals and network equipment can be promoted. The use of open, inter op erable, non-discriminatory and demand-driven standards meets the needs of users and consumers. Coordination agreements with neighbours, both on a bilateral or multilateral basis, can hasten licensing and facilitate network planning.

10. Adopt a broad approach to promote broadband access. Spectrum management alone is inadequate to promote wireless broadband access. A broad approach, including other regulatory instruments; such as effective competitive safeguards, open access to infra struc ture, uni-versal access/service measures, the promotion of supply and demand, licensing, roll-out and market entry meas-ures; the introduction of data security and users’ rights, where appropriate; encouraging the lowering or removal of import duties on wireless broadband equipment; as well as development of backbone and distribution net-works is necessary.”

5.5.2 Additional Best Practices

In addition to the GSR guidelines, it is useful to more fully explore several best practices, to identify their benefits, as well as their potential limitations, in advancing the cause of consumer broadband services. This subsection will analyse the following best practices:• Harmonizing spectrum allocations on a global basis to

increase economies of scale at product and the service layers and to reduce end-user costs.

• Fostering the use of standards-based technologies to increase economies of scale.

• Allocating spectrum and developing technical rules that encourage adjacent spectrum users to have compatible tech-nical characteristics, as a way of limiting interference and maximizing use of spectrum (“good neighbour” policies).

• In shared spectrum bands, encouraging or mandating tech-nical standards that foster cooperative systems designed to reduce harmful interference.

• Develop efficient and transparent licensing rules and proc-esses that allow for restructuring of incumbent spectrum bands in order to implement harmonization goals.

5.5.2.1 Global Harmonization of Allocations

Broadband spectrum regulators can significantly drive down the costs of broadband wireless services and boost sub-scriber numbers by harmonizing spectrum allocations with global practices. Harmonization allows equipment manufac-turers to benefit from economies of scale by manufacturing equipment for large pools of customers in multiple countries.

Of course, the ideal of global harmonization, as depicted in Figure 5.4 (IMT-2000 bands) can lead to a significant amount of incumbent dislocation – with the regulator feeling the attendant political backlash. This is especially true in devel-oped countries with a lot of entrenched incumbents, most of which feel they have rights under the exclusive spectrum use model. Harmonization can usually be achieved only through transparent licensing systems and significant political will-power.

Needless to say, participation in the ITU and regional organizations can help provide the economic and political support to require relocation of incumbents from newly har-monized bands. For example, ITU-R Study Group 8 and its Working Party 8F are currently exploring potential allocations for BWA technologies, including WiMAX. Governments need to participate in these ITU groups, as well as in the industry-led groups that pioneer standards. This will give spectrum managers advanced warning about the global direction of these technological developments, so they can establish appropri-ate spectrum regulations before the standards become widely deployed.

5.5.2.2 Fostering Standards-Based Technologies

In order to maximize their spectrum allocation decisions, spectrum regulators must closely follow and support the norms and recommendations of multiple standards-setting organiza-tions, including: (a) the Institute of Electrical and Electron-ics Engineers (IEEE); (b) the European Telecommunications Standards Institute (ETSI); (c) the Wi-Fi Alliance (for 802.11 products); and (d) the WiMAX Forum (for 802.16 products).

Similar to global harmonization, the effect of embracing standards-based technology development is to significantly reduce the cost of the devices by reducing the number of pro-

Figure 5.4: Globally Harmonized Spectrum: IMT-2000

prietary components. This leads to faster adoption of the equip-ment and the associated services. There are many benefits to a standards-based approach, but it is also important to remember that it is a consensus-driven process. If not managed carefully, the standards process can essentially commoditize innovation and could even become a de facto form of economic regulation, especially if regulators embed the standards into their regula-tory systems.

Broadband regulators also must be mindful that a stand-ards-based approach, with its natural tendency towards com-promise, could potentially lead to suboptimal results. These unintended “costs” must be part of a broader policy tradeoff in embracing any particular standard. Generally, regulators should avoid embedding standards in their policies and rules without an extremely compelling policy rationale.

5.5.2.3 ‘Good Neighbour’ Allocations

This best practice generally involves grouping spectrum allocations based on interference and other technical compat-ibility characteristics. This good neighbour “zoning” practice can enhance the compatibility of spectrum uses, based on power or bandwidth characteristics, maximizing overall capac-ity and reducing transaction costs. In a fashion similar to the global harmonization of spectrum allocations, this practice also raises the potential for significant relocation costs, which regulators have to consider in making any decision to realign spectrum allocations.

5.5.2.4 Voluntary Sharing Guidelines for Unlicensed Bands

For users of shared spectrum – such as those in unlicensed bands – regulators can encourage voluntary coordination to better manage any interference or capacity issues. Users of these bands should understand that the alternative to self-regu-lation might be reverting to a command-and-control model that would deny them service and technology flexibility.

5.5.2.5 Infrastructure Sharing

Regulators can also enhance wireless broadband adoption by inducing licensees to share infra struc ture such as towers and backhaul facilities. Infrastructure sharing between wire-less systems promotes efficiency, reduces deployment costs and reduces environmental impacts by avoiding the construction of duplicative facilities.

5.5.2.6 Setting Different Power Limits for Rural Areas

Regulators should adopt different rules for spectrum used in different geographic markets. In lower-density environ-ments, such as rural areas or under-served communities, there is less opportunity for interference. So it may make sense to allow transmitters to operate at higher power levels. Similarly, regulators can increase the size of the bandwidth allocated to broadband services in these areas in order to increase capacity. With limited competition for spectrum in these areas, grant-ing flexibility might provide better economic incentives for the licensee to deploy a BWA network. Ireland’s efforts to grant rural BWA operators greater flexibility with regard to spec-tral power limits has enabled the country to meet the broad-

band demands of its rural population. In addition, Ireland has endeavoured to keep licensing obligations as low as possible to reduce barriers to entry (See Box 5.5).

5.5.2.7 Transparent Licensing Systems and Processes

Another best practice is for regulators to establish transpar-ent and automated licensing procedures and records in order to reduce transaction costs associated with facilitating the high-est and best use of spectrum rights. As part of their licensing efforts, regulators should periodically test, analyse, and audit spectrum resources to measure how efficiently incumbents are using it and whether there are interference issues. Regulators can use this information to form improved sharing and inter-ference rules and regulations, as well as to expand spectrum capabilities and permit new forms of uses within the band for either the incumbents or new licensees.

5.5.3 Technology Neutrality

Technology neutrality is usually understood to be when regulators apply rules and regulations in a way that does not favour one type of technology over another. Clearly, this con-cept could be at odds with some of the other best practices, such as global harmonization, adopting standards-based tech-nologies and applying good-neighbour allocation rules, which by their very nature tend to favour particular technological choices.

From the perspective of the regulator as a resource man-ager, full technological neutrality is an impossible goal, because the desire to achieve efficiency and rapid utilization of the spectrum ultimately requires decisions that point to particular technology paths. Promoting a standard or a particular spec-trum band or its configuration, directly or indirectly, obliterates the notion of neutrality. Similarly, creating a harmonized spec-trum band and associated service rules will also tend to favour particular technologies. So there is an inherent contradiction between the goal of technological neutrality and the function of the spectrum regulator as a resource manager.

Fortunately, pragmatism once again comes to the rescue. A spectrum regulator has to play different roles that lead to dif-ferent conclusions about technological neutrality. On the one hand, the spectrum regulator is a resource manager concerned with optimizing the efficient use of a scarce resource. On the other hand, the spectrum regulator is a policy advocate trying to achieve social policy goals such as universal access/service and reducing the Digital Divide.

A practical solution to the conundrum would apply tech-nological neutrality only to the means used to achieve macro policy goals, rather than the means for managing the spectrum resource. Technological neutrality is paramount only for the means applied to achieve broad social policies such as attaining universal broadband access. To achieve that goal, any combi-nation of available technologies and resources can be enlisted. One example of effective balancing between the goal of tech-nology neutrality and pragmatism can be found in the public consultation process of the Office of the Telecommunications Authority (OFTA) in Hong Kong, China (See Box 5.6).

CHAPTER 5


86

CHAPTER 5 87


5.6 Case Study: BWA Spectrum Allocation in MauritiusIn order to gather all of the threads together in a real-world

example, this chapter concludes with a comprehensive review of a major BWA allocation decision by the Mauritius Informa-tion and Communication Technologies Authority (ICTA).

During the first half of 2005, ICTA made a series of deci-sions to establish the future course of BWA services in this fast growing island economy. Although it is too early to know the results of ICTA’s allocation decisions, a brief review of the Authority’s processes and methodology reveals a regulator that has adopted key best practices and made pragmatic tradeoffs that should enable BWA to take root rapidly.

Unlike many regulatory bodies, ICTA has a broadly defined mandate that, pursuant to its 2001 charter, calls for it not just to “manage” spectrum but to increase the reach of information and communication services throughout the country. Turning its attention to BWA offerings, ICTA under-took a transparent public consultation process and arrived at its final decision within a commendable 180 days after its initial report.

In starting the process, ICTA first identified the demand for BWA offerings, while also recognizing the need to harmo-nize its allocation decisions with global allocation trends, in order to take advantage of scale economies. ICTA also iden-

Box 5.5: Eire’s Response to BWA, Part Two

Ireland’s response in the process of compiling the 2005 GSR Best Practice Guidelines on Spectrum Management to Promote Broadband Access also addressed the issue of reducing licensing obligations as well to lower market-entry barriers.

Principle One: Barriers to entry should be as low as possible

Our [Ireland’s] experience indicates that regulators should minimize barriers to entry in this area by allowing broadband suppliers to begin operations on a small scale, and not imposing onerous rollout and coverage conditions. Ireland has awarded national licences in the past for broadband wireless access that incorporated rollout and coverage obligations. But none of the licensees were able to make a viable business case and, consequently, rollout of services was less than satisfactory.

In 2004 ComReg announced a new scheme for the licensing of broadband fixed wireless access services in local areas. Each local service area was defined by a 15 km radius circle from a base station, with an interference zone extending to a 30 km radius, at the perimeter of which a certain field strength should not be exceeded in order to limit interference into adjacent areas. Since its inception, 110 licences have been granted on a first-come, first-served basis. The success of this approach, compared with the earlier attempt at national licences, is reflected in an increase of 43 per cent of customers in the last six months.

One of the key reasons for the success is that operators only take out licences for areas in which they are able to develop a vi-able business case and, as there is no national network rollout obligation, all attention is focused on the local area. Initial concerns that rollout would only occur in urban areas (due to high population) have proved to be unfounded as small entrepreneurs and local community groups have taken up the challenge to supply broadband access to many rural areas where ADSL is not available. Current rollout is shown in Figure 5.5.

Figure 5.5: BWA Coverage Areas (Circles) in Ireland

CHAPTER 5


88

tified key factors that have negatively affected past attempts in Mauritius to develop BWA services, such as congestion in the unlicensed 2.4 GHz band. ICTA found that operators in that band were exceeding designated power limits, using the licence-exempt systems for long-range transmissions – con-trary to their design and purpose. It also noted the continu-ing and significant demand from ISPs for wireless spectrum to deploy their services.

Additionally, ICTA took note of other countries’ decisions to define certain bands for BWA uses, including 2.5-2.7 GHz, 3.3-3.5 GHz, 5.1-5.3 GHz and higher-powered unlicensed use of 5.4-5.8 GHz. ICTA took into consideration the potential of new standards such as WiMAX for deploying BWA services integrated with computers and other ICT devices. The ICTA’s ultimate decisions demonstrated a practical approach to resolv-ing the various tradeoffs needed to deploy BWA networks in the face of competing demands from operators and other incumbent spectrum users. The following subsections describe how ICTA resolved issues in each band.

5.6.1 The 5.4-5.8 GHz Band

ICTA determined that the presence of radar incumbents in the 5.4-5.8 GHz band required postponing any BWA allo-cation decision in the band, even though this band had been globally harmonized for higher-powered unlicensed opera-tions through the WRC 2003 negotiations. The complexity of the radar operations, coupled with their national defence purpose, will require a more deliberate transition for this band. This issue is similar to the difficulty that the United States has experienced in implementing the necessary dynamic frequency selection (DFS) systems needed to protect sensitive military operations in that band in the United States.

5.6.2 The 2.4 GHz BWA License-Exempt Band

In reviewing the status of the 2.4 GHz unlicensed band, ICTA came to the conclusion that the previous “commons” model had led to overuse and overcrowding. It found that exist-ing operators in the band tended to exceed the power limits, which were set at 23 dam EIRP, in order to extend their ranges and overpower interference. ICTA decided to set a distance limitation for this band to discourage operators from using it for longer distance applications than the WLAN networking it was originally intended for. In its decision, ICTA took some pragmatic steps to improve the functionality and longevity of the band:• Mandated use of the bands for applications not to exceed

500 meters;• Mandated limiting the emitted power to 20 dam while

giving some transitory leeway for incumbent operators to stay at the 23 dam until 2010; and

• Required new systems to register with ICTA, allowing it to track the level of usage in the band and identify poten-tial interference problems.

These provisions may well extend the useful life licence-exempt operations in the 2.4 GHz band. Moreover, the prob-lems encountered in Mauritius underline the pitfalls associated with the commons approach.

5.6.3 The 2.5-2.7 GHz BWA Band

This band previously had been allocated for MMDS. ICTA’s decision harmonized the band with the IMT-2000 allocation, enabling three distinct types of BWA systems to eventually operate in the band. Implicit in the decision was the determination that the existing MMDS use was not as relevant or promising as the potential BWA systems that could use the

Box 5.6: OFTA’s Consultation on Broadband Wireless Access Licensing

In August 2005, the Office of the Telecommunications Authority (OFTA) in Hong Kong, China, issued its “analysis of com-ments received, preliminary conclusions and further consultation on a licensing framework for deployment of broadband wireless access.” Comments on the further consultation were invited through 31 October 2005. The August consultation followed an initial BWA consultation launched in December 2004 on whether BWA should be licensed in Hong Kong, and if so, when.

In the August consultation, OFTA expressed its view that BWA spectrum should be assigned in 2006, on a technology-neu-tral basis. Consistent with the technology neutrality principle – and having considered the respondents’ views – OFTA said it was prepared to allow the deployment of any technology that conformed to recognized open standards for the delivery of BWA services. Because BWA devices and equipment will be supplied competitively, OFTA considered it unlikely that end users would have insufficient choices in the selection of BWA devices.

OFTA also expressed the view that, although BWA is currently being deployed as a fixed service, it should also allow mobile services, once the technology is developed and cost-effective. OFTA therefore proposed that the scope of permitted services of the BWA licences be restricted to fixed tele com mu ni cation services initially, but expanded to include full mobility services after 1 January 2008. Fixed tele com mu ni cation service will include the conventional wireless local loop services, plus “limited mobility” offerings, which will not be able to hand off calls between cell sites until after 1 January 2008.

OFTA also said it would issue unified carrier licences in order to accommodate the trend toward fixed-mobile convergence, since BWA can offer both. The new Unified Carrier Licence would be valid for 15 years – the same duration as for existing fixed/ mobile carrier licences. OFTA has made clear that BWA licensees will be expected to invest in, and roll out, infra struc ture to pro-vide public services. They will not be allowed to enter the market solely as services-based operators. The consultation document may be accessed at http://www.ofta.gov.hk/en/report-paper-guide/paper/consultation/20050831.pdf.

CHAPTER 5 89


band. ICTA did not, however, make an explicit determination about the value of MMDS relative to BWA. Rather, ICTA took its cue from the public consultation process, in which only one respondent advocated the continued use of the band for MMDS.

Other aspects of the ICTA’s decision about this band were significant. First, it determined that only licensed opera-tors could provide BWA services in this band. Second, it made sure to create a channelization plan for the band that would group similar systems together, instead of using the interleav-ing employed with broadcast bands. As a result, both TDD and FDD wideband systems could be deployed to deliver BWA. The channel sizes were also changed to 5 MHz from 8 MHz segments, which are better suited for wideband systems such as those being developed by the IEEE WiMAX groups.

ICTA also allocated 40 MHz (20 MHz for uplink opera-tions and 20 for downlinks) for hybrid satellite and terrestrial services. Finally, ICTA created a transition period for incum-bent operators, making the band available for BWA services only in January 2010. ICTA left the door open for voluntary relocation, but with language suggesting that the timeline for transition might be accelerated based on “market conditions.”

5.6.4 The 3.4-3.6 GHz BWA BandThe ICTA reallocated this band from primarily a fixed sat-

ellite service band to a band in which those operations will be co-primary with terrestrial BWA operations. Fixed-link serv-ices were favoured in this band in order to provide protection for incumbent VSAT operators. Despite this limitation, BWA

advocates such as wireless internet service providers (WISPs) were able to obtain the benefit of higher-powered use, since ICTA allowed for 15 W EIRP systems to operate in the band. In assigning channels in these bands, ICTA also decided it would give priority to public operators (although not conclusively); that licenses were required and that the permitted point-to-point and point-to-multipoint links must be registered. Finally, ICTA allowed aggregation of multiple 25 kilohertz channels, while requiring a minimum of 100 MHz separation to avoid interference of duplex operations in the band.

5.6.5 The 5.150-5.350 GHz BWA Band

Consistent with WRC 2003 agreements, ICTA opened up this band for mobile licence-exempt use of equipment consist-ent with IEEE 802.11 (Wi-Fi) standards. Given the challenges presented by incumbent radar operations in this band, ICTA determined that Wi-Fi in this band would be limited to indoor use only. Additionally, in order to prevent overcrowding and potential interference to incumbent operations, ICTA required the equipment in this band to use dynamic frequency selection (DFS). This is an automated mechanism that detects the pres-ence of signals from other systems, notably radar systems, and avoids co-channel operation. Equipment must also have trans-mit power control (TPC), a mechanism that regulates a device’s transmit power in response to an input signal or a condition. These capabilities must be certified as part of the equipment registration and approval process that ICTA plans to establish for this band.

Figure 5.6: The 2.5-2.7 GHz BWA Allocation in Mauritius

CHAPTER 5


90

The limitations imposed on the use of this band reflect the tradeoffs ICTA believed were necessary to advance BWA while accommodating the sensitive operations already existing in the band. Both DFC and TPC systems are at early stages in their development and have not yet fully demonstrated their ability to protect incumbent operations. Nonetheless, they are clearly part of the technological advances enabling the rapid deployment of BWA services.

5.7 ConclusionThe ability of broadband wireless access networks to

improve our lives ultimately relies on the amount of spectrum regulators make available for BWA. For the first time in the relatively short history of spectrum management, however, it appears that advances in technology – independent of the actions of regulators – can increase spectrum capabilities and resources. New technologies allow users to do more with the same amount of spectrum and enable new uses for spectrum not previously possible. As these advances become more widely adopted and new spectrum resources become available, regula-tors must consider whether traditional approaches to spectrum management are sufficient to address the resulting challenges and opportunities.

A sensible approach for BWA spectrum licensing calls for granting spectrum licensees unlimited technical flexibility (so long as they avoid harmful interference to adjacent licensees) to create more spectrum capabilities and resources. Licensees

should have enough operational autonomy to enter new lines of business. The pragmatic regulator grants these additional rights to licensees as long as they meet two absolute precondi-tions important to the development of communications mar-kets: to increase competition for broadband communication services and to experience the opportunity cost of using their spectrum assignments.

While contemplating the appropriate regulatory model for the evolving state of spectrum technology, spectrum regula-tors must keep in mind key best practice concepts that have developed around spectrum management over the last three decades. These best practices fostered the widespread adoption and deployment of an earlier generation of wireless services. The same best practices will help deploy BWA networks today, increasing the welfare of consumers in the modern Informa-tion Society.

In reviewing recent comprehensive BWA allocations in Mauritius, one can trace the outlines of the practical tradeoffs that are necessary to achieve the broader policy goals of the regulators. The lesson to be drawn from this type of complex allocation decision is not whether any particular outcome is better than any other. The lesson comes from better under-standing the need to balance the demands of allocating new spectrum rights and resources for new technologies, while simultaneously advancing broader social and policy goals such as benefiting consumers and boosting economic productivity. This balancing act, after all, is the task entrusted to the modern spectrum regulator.

1 Opportunity cost, as defined by WiKipedia, is a term used in economics, to mean the cost of something in terms of an opportunity foregone (and the benefits that could be received from that opportunity), or the most valuable foregone alternative. http://en.wikipedia.org/wiki/Main_Page

2 See http://www.itu.int/ITU-D/treg/Events/Seminars/2003/GSR/WSIS-Statement.html and http://www.itu.int/ITU-D/treg/Events/Seminars/2004/GSR04/consulta-tion.html

91


CHAPTER 6

Voice over Internet Protocol (VoIP) service is often viewed as a “disruptive technology,” meaning that it has the potential to drastically alter the status quo in the global tele com mu ni-cation industry.1 In fact, all of the current market indications show that IP networks and VoIP services will replace tradi-tional PSTN networks and services. ITU expects that by 2008, at least 50 per cent of international minutes will be carried on IP networks, and many carriers will have all-IP networks. Recent trends are certainly headed in this direction. For exam-ple, in the United States, residential VoIP subscriber numbers rose from 150,000 at the end of 2003 to more than 2 million in March 2005. U.S. subscribership is expected to exceed 4.1 mil-lion by the end of 2006, generating over USD 1 billion in gross revenues for the year.2

ITU believes that much of the mobile traffic in the world will become IP-based, as well, and that the introduction of mobile VoIP will influence the shape of the mobile business globally. Wireless-enabled VoIP offers the potential for cheaper voice calling. Users of 3G networks can already use mobile phones to make VoIP calls at cheaper data rates. The growth of mobile VoIP will particularly affect the high-priced interna-tional roaming business.

Today, however, VoIP services have been implemented unevenly around the world. Some countries have legalized and allowed multiple providers, while others have completely blocked the provision of VoIP services. Some countries have only “grey market” VoIP providers. These varying approaches reflect very different perceptions of VoIP in various parts of the world. In some countries, VoIP is seen as an exciting techno-logical development that offers cheaper calling for consumers. But in others, it is seen as an unauthorized threat to the existing order. At the international level, VoIP traffic is often described as “by-pass” or “lost” traffic, but it is driving the development of new service providers in both the developed and developing worlds.

The advent of VoIP has brought new challenges for regu-lators. In developing countries, where the entrenched rights of fixed-line operators are most protected, the main question has been whether to legalize the introduction of VoIP. In more mature and competitive markets, meanwhile, VoIP has raised questions about what aspects of it should be regulated. The countries that have legalized VoIP as part of a broader liber-

alization of their markets have started to accumulate thinking, experience and precedent in this area. But even in these coun-tries, VoIP is a relatively recent development, and there is often little consensus about how to regulate it.

This chapter examines how VoIP services will affect future regulation. Because of the starkly contrasting global perceptions of VoIP, however, it is difficult to present a uni-fied approach to regulating VoIP. Instead, this chapter aims to reflect regulatory experiences from a wide range of countries that are grappling with the transition to VoIP. The sections of this chapter are structured to answer both the broad and spe-cific questions raised by VoIP services, including:• The overall approach to regulating VoIP as a mainstream

service;• How VoIP has changed telephony business models and

the various ways of classifying the services it has created; and

• Other related issues frequently raised in connection with VoIP, such as quality of service; network integrity; com-munication security and lawful interception.

6.1 VoIP: Regulatory Evolution or Revolution?

6.1.1 How VoIP is Changing Voice Business Models

In liberalized tele com mu ni cation environments, it is important to create regulatory frameworks that allow the market to produce sustainable business models. VoIP is a pow-erful service innovation that has the potential to change how existing voice markets operate. Since VoIP service is largely enabled by the existence of IP networks, there is inevitably an overlap throughout this chapter between these two key con-cepts, which remain inextricably linked. It is not yet clear how IP networks will be implemented – or at what speed – but it is important to try and identify key elements of the changing business model in order to understand the policy and regula-tory dilemmas VoIP raises.

Current regulatory practice for telephone service was devised at a time when circuit-switched technology was domi-nant. Historically, regulators treated different types of networks differently. Future regulation should, however, be based on a

Authors: Tracy Cohen, Independent Communications Authority of South Africa (ICASA); Olli Matila,Finnish Communications Regulatory Authority (FICORA) and Russell Southwood, Southwood Consultants

6 VOIP AND REGULATION

CHAPTER 6


92

fundamental recognition of the convergence of tele com mu ni-cations, broadcasting, media and information technology sec-tors. This means that all transmission networks and services increasingly will be addressed by a single regulatory frame-work. For the remainder of this chapter, the term Information and Communication Technologies (ICTs) will be used to reflect this comprehensive, forward-looking approach.

For operators, VoIP represents three broad types of com-mercial opportunities: price arbitrage, savings from new net-work topologies, and new products and services. Each of these is considered in the following subsections.

6.1.1.1 Price Arbitrage

Arbitrage is a term used to describe a situation where one buys something at a cheaper price in one market in order to sell it at a higher price in another. The growth of VoIP for interna-tional calling has been built on the wide gap between retail and wholesale calling prices in many parts of both the developed and developing world. These differences are a function of the

uneven introduction of competition in voice markets around the globe.

In Africa, for example, it may cost a caller (at the retail rate) between USD 0.50 (50 cents) and USD 1 a minute to call Washington, DC, but the international operator may buy the call (at the wholesale rate) for between USD 0.01-0.03 (1-3 cents). Where this circumstance exists, incumbent operators are able to maintain high margins because they either have monopolies or limited competition at the retail level. What is often described as “bypass traffic” or the “grey market” is a proxy for competition, particularly international voice calling. This has already forced incumbent telephone companies to cut their international rates. And because there is less regulatory protection of international voice markets, these rates will con-tinue to decrease.3

Moreover, many incumbent operators are going through the process of “rebalancing” their tariffs in line with the costs of providing services. In the pre-competition days, high inter-national rates were used to cross-subsidize rates on domestic networks. With competition in the largest voice markets driv-

Box 6.1: A VoIP Primer

“Voice over Internet Protocol” (VoIP) is a generic term referring to a technical standard that enables the transmission of voice traffic, in whole or in part, over one or more networks that use Internet Protocol (IP).

Standards or “protocols” for VoIP are still evolving, but two main open protocols and proprietary vendor protocols enable VoIP:

– “H. 323” – The most widely adopted protocol for the transmission of VoIP, this is an ITU legacy standard that builds on earlier protocols for the transmission of voice and video over analogue PSTN, ISDN and ATM networks; and

– “Session Initiation Protocol” (SIP) – An application-layer control protocol, SIP is an end-to-end signalling protocol. SIP facilitates communications between two or more SIP-supported devices, but it is not the only protocol required to make VoIP calls, which take place via additional protocols.1

How VoIP technology works: Voice (or data) is compressed and converted into digital packets that travel over the internet (or a private network utilizing VoIP) and are then converted back at the other end, correcting for echoes from end-to-end delay, for jitter (variability) and for dropped packets. The data packets are non-isochronous and may take many different and independ-ent paths to the intended destination, arriving out of sequence or with different end-to-end delays. VoIP technology makes much more efficient use of bandwidth. Plus, voice is transmitted on IP-based networks at considerably lower cost than calls on circuit-switched networks, which require dedicated connections for the entire duration of the call.

VoIP Applications: The first generation of VoIP services, known as PC-to-PC or Class 3 services, allows individuals only to call other people using the same service (examples include Yahoo! Instant Messenger). Voice signals transmitted are not routed or switched through the PSTN at all. Second-generation VoIP services (PC-to-Phone or Class 2 services) allow calls from PCs to any PSTN telephone number, including local, long distance, mobile, and international numbers (examples include Dialpad, Net2Phone and Skype Out). A third generation of VoIP services (Phone-to-Phone, or Class 1 2 enables use of a traditional tel-ephone to make VoIP calls, using an adaptor at the customer premises. The calls are then routed over an IP network rather than the circuit-switched PSTN.3

1 See Webopedia at http://www.webopedia.com/TERM/I/Internet_telephony.html

2 Class 0 is phone-to-phone over the PSTN. This is David Clarke’s classification system (MIT).

3 A VoIP network requires a terminal or communication end-point, which can be a phone, PC or even a software programme. Terminals are identified by at least one IP address (e.g. [email protected]) and are registered with a server, which stores IP addresses and can map an address to a terminal. The server might also store location, identification and traffic data. Finally, gateways act as bridges between the local PSTN and IP network to allow calls between different networks so that the signaling protocol can be understood between networks and so that IP addresses and regular PSTN numbers are recognizable between networks. Signaling data is exchanged between switched circuit telephone networks and VoIP networks. This information is used to set up, manage and release voice calls, and to support telephony services such as caller ID, toll-free calling, and mobile authentication and roaming services.

CHAPTER 6 93


ing down international rates, cross-subsidization is unlikely to be sustainable. Mandated tariff rebalancing will also end this business model in many countries.4

The VoIP service market has been fostered by the intro-duction of IP networks and the proliferation of Web-based transactions. Grey market operators can simply sign up on a website and gain access to international calling capacity that they can resell to end users. Because these are data services, the calls are not recorded as telephone minutes, and they need not pass through the international gateway of the incumbent operator. While quality-of-service issues do arise, many cost-conscious callers seem willing to make the trade-off between price and quality.

As a commercial opportunity, however, much of this market is entirely price-dependent. As international calling prices plunge, the arbitrage effect may disappear. The legaliza-tion of VoIP services in a wider range of countries is, in effect, the introduction of greater competition that will reduce this price-arbitrage gap. The future of telephony revenue, particu-larly in the international domain, appears to be “low-margin, high volume” rather than “high-margin, low-volume” for this type of calling. Where VoIP has been legalized, VoIP providers appear content to work with lower margins than established operators.

6.1.1.2 Savings from New Network Topologies

Many of the world’s larger operators have been persuaded to consider VoIP because an IP-based network can carry both voice and data in one network rather than two. In this way, operators can invest in a single network that can be used more efficiently for different forms of traffic.

Moreover, IP network deployment costs often come in smaller increments than those required for circuit-switched facilities and dedicated circuits. It is possible to add capacity incrementally, in a manner that will bring a return on invest-ment more quickly than the traditional multi-million dollar tele com mu ni cation infra struc ture investments, which require many years to produce the required return. For example, operators can replace large numbers of traditional switches with fewer “soft switches.” Smaller investments can often be financed from cash flow rather than major external borrow-

ing, and some IP network components can even be purchased in retail electronics outlets rather than as inflated “integrated solutions.”

Not surprisingly there is considerable debate about these cost advantages. Some of the debate centres on reliability and cost of the newer generation of network equipment – including Wi-Fi and Wi-MAX – as part of an IP network roll-out. These new wireless technologies can and are being deployed both to create local loop VoIP access and for backbone links. Again, it is argued that this is being done at prices that are much cheaper than possible with traditional copper or fibre networks. As with arguments about IP networks, the potential cost savings using wireless technologies can be debated. But incumbents world-wide are deploying these technologies, at the same time they are threatened by them.

Similarly, mobile operators that have invested consider-able sums in 3G licences and need to make a return on their investment over 5-10 years are also threatened by the potential of VoIP. They are particularly vulnerable to the effect substitu-tion of mobile VoIP will have on operators’ high-priced inter-national roaming services. This presents a recurring dilemma for regulators. Should they protect the mobile operators’ investments and delay cost-saving innovations for consumers? Or, should they allow wireless-enabled, mobile VoIP to flour-ish, potentially risking thwarting investment in 3G network deployment?

These questions are particularly pertinent for regulators and policy-makers in developing countries. There, the choice is often between defending a government-owned incumbent (for financial and social reasons) and making cheaper commu-nications available to a wider number of people, particularly in rural areas.

6.1.1.3 New Products and Services

The convergence of voice, data and video on IP networks allows users to combine these different forms of traffic and significantly expand the range of product and service offerings. Many operators are now offering the so-called “triple-play” option that combines all three in a single service. Senegal’s Sonatel, for example, has rolled out a “triple-play” service offering voice, -internet access and television programming.

Figure 6.1: Shades of GreyGrey Market Revenues as a Percentage of Overall International Call Revenues in Selected Regions and Countries

Africa 20-30%*Brazil 36%**Colombia 50%***Costa Rica 20%****

* Source: Balancing Act** Source: Abrafi x*** Source: V-P Technology, Orbitel**** Source: Incumbent telco ICE

CHAPTER 6


94

This ability to package services together has implications for competition, as users increasingly seek a single provider and billing option. Triple play offerings have the potential to open up television as a delivery platform for a far wider range of rich, multimedia services. This could overcome, to some extent, the lack of installed, internet-connected computers in developing countries. This may be a solution for the urban poor, but it still will not address lack of internet access in rural areas without electricity or television coverage. Yet, even though the triple play may not be everywhere overnight, it is certainly going to be relevant in the medium-term.

In addition, there are now several VoIP crossover tech-nologies coming into use on mobile phones. For example,

“push-to-talk” services are essentially what most people used to understand as “walkie-talkies.” That is, they enable instan-taneous, direct two-way conversations between two individu-als, using IP software that rides on a mobile phone platform. U.S.-based operator Nextel (based on its origins using a trunked radio network) pioneered the push-to-talk experi-ence, and others started offering the service in 2003. The prod-uct debuted in Europe with Orange’s “Talk Now” offering. Although it requires users to subscribe to a data service from one of the major carriers – and there are some service quality issues – there has been significant take-up. Push-to-talk is suf-ficiently threatening the traditional “walkie-talkie” market that Motorola has produced specially designed, rugged “push-to-talk” phones at the beginning of 2005. Two of India’s mobile providers – Hutchison Essar and Tata Indicom – also launched

“push-to-talk” services in May 2004.

Another area of development will be the current testing of products that integrate cellular and WLAN networks and provide voice from a WLAN device (See Chapter 3). Mobile carriers in the United States and Europe are already rolling out networks that offer 11 mbit/s 802.11 WLAN access. This could eat into their existing data services (and perhaps even undercut the rationale for 3G), but operators know they cannot ignore the technology if competitors deploy it. One European UMTS mobile operator has been sufficiently worried by the impact of VoIP that it has threatened to block Skype calls to its subscrib-ers.5

6.1.2 Changes in the Voice Business Model

These three categories of business categories – arbitrage, new network topologies and new services – have come about from a number of key changes in the underlying business model for voice service.

6.1.2.1 The Impact of IP Network Features

With traditional telephony, “intelligence” in the network is located centrally (in the func tion alities of the switch) and is usually controlled by one operator. Largely “dumb” devices (telephones) are attached to the network, and they have a limited set of functions. The traditional telephone network’s root-and-branch structure means that traffic flows to and from exchanges in ways that reinforce this pattern. For example, traffic for international destinations is commonly routed via a single international gateway. Telecommunication carriers main-

tain bilateral relationships with other carriers and exchange revenue through the international settlements system.

By contrast, IP networks are not controlled by any single entity, other than for the most basic transport to other net-works. The “intelligence” is deliberately designed out of the network architecture. Put simply, the network is “dumb” and the intelligence is at the edge of the network, in the terminal equipment. A computer accessing the network has a far more complex range of service functionality in its application pro-grammes than the network to which it is attached.6

IP traffic is routed via the easiest route, not through cen-tral choke points. International traffic can just as easily flow from an ISP, a cybercafé or a telephone company. Each of these has only to open a network connection and have the required capacity available. In this way, VoIP traffic does not need to go through an international gateway. The network design origi-nated from United States military requirements for redun-dancy and resilience, and that led to a decentralized network without a central focus or control point. And because of its use for research, the internet was designed to be open to users through publicly available standards, making it easy to access.

Due to their open nature, IP networks pose particu-lar security challenges. The PSTN and mobile networks are closed systems with controlled security and privacy. IP net-works, however, have open architectures in which vulnerabili-ties, threats and communication security risks exist in various network elements. Special measures are required for ensuring communications security.

6.1.2.2 The Separation of Retail and Wholesale

Historically, a vertically-integrated organization like the telephone company carried traffic and offered services, usu-ally from a monopoly market position. In a more liberalized market, the same telephone company might be selling inter-national transmission to both external ISP customers and to its own ISP. This can lead to accusations of conflicts of interest. For VoIP service providers, the terms under which they can access broadband networks become a key issue.

With liberalization bringing new market entrants, there has been increasing discussion of separating the “retail” func-tions of the operator from the “wholesale” ones. In other words, there is more talk of service provision and network operation as separate roles. Alternative infra struc ture providers such as util-ity companies have begun to reshape themselves into whole-sale operators by adding bandwidth to their private networks. Meanwhile, ISPs, VoIP service providers and “mobile virtual network operators” or MVNOs (essentially, mobile service resellers) have begun to retail services to end users. The nature of IP networks has encouraged changes in thinking about these two functions.

As a result, many telephone companies have internally separated out their wholesale and retail functions in order to better understand the underlying cost structures of differ-ent parts of their business. In some instances, this has been prompted by regulators seeking to assess the costs and clarify the terms for access to local loops or other network elements.

CHAPTER 6 95


In other cases, the companies themselves have wanted to ana-lyse their costs more closely.

6.1.2.3 Changing Charging Structures

Wholesale traffic rates are largely the same on routes between the more competitive markets in the world. So it is no surprise that several VoIP service operators offer the same or broadly similar retail rates for calls between these countries. For example, Skype sells “Skype Out” minutes to enable sub-scribers to call PSTN phones in other countries. At the time of publication, it cost less than USD 0.02 (2 cents) a minute on Skype to call Australia, Chile, Europe and North America.

Most telephone calls have traditionally been billed accord-ing to where they originated and terminated. Pricing was distance-sensitive, even where the actual cost of terminating the calls was not. This model, however, has been turned on its head by VoIP services. Now, per-minute charges are based on what the market will bear in the terminating country, and those countries with less competition are more expensive to call. But even those countries can be called more cheaply using VoIP than routing through their circuit-switched incumbents’ services.

As discussed in greater detail in Section 6.4.3, number-ing has traditionally reflected geographic location, but with VoIP, this is no longer the case. Many VoIP service providers offer users “virtual” numbers, allowing them to be reached at a

“local” number when they are actually thousands of miles away. For example, a VoIP customer may live in London, where her PSTN dialling code is “207.” Her mother may live in Florida, where she has a “561” PSTN dialling code. The VoIP service provider can give its customer a “561” dialling code that rings on her London line, allowing her mother to call her at the VoIP rate and avoid international long distance charges. Argen-tina’s PVTEL, for example, offers its customers the choice of a Buenos Aires or Miami dialling code.

Because fixed-line and mobile phones (rather than SIP or “soft phones”) are still the dominant form of telephone termi-nal equipment, a hybrid service model is emerging in which operators offer consumers a combination of broadband access, plus an adaptor or VoIP-enabled phone. The package will also include free calling to the company’s other subscribers (as an incentive to recruit new subscribers by “word of mouth” sales) and cheaper domestic and international calls.

6.2 The Pace of VoIP Market Development

All current market indications show that IP networks and services eventually will replace PSTN networks and serv-ices and will alter the mobile business, as well. Moreover, the introduction of IP networks will affect all countries, although the timeline of adoption will vary widely. Some major interna-tional carriers have committed themselves to making the tran-sition to VoIP, including:

• British Telecom (100 per cent conversion planned by 2009);

• MCI (100 per cent of all traffic by the end of 2005);• AT&T (100 per cent by the end 2010); and• Telecom Italia (80 percent of all traffic was VoIP by the

end of 2003).

In Europe, the number of market players offering VoIP increased with astonishing speed during late 2004 and early 2005. According to rough estimates, in March 2005 there were at least 10 VoIP service providers in most Western European countries, and in some the number exceeded 40. This kind of growth appeared to be taking place across the globe. For exam-ple, there were 11 companies in Pakistan offering VoIP by early 2006, and more than 80 VoIP providers were licensed in Malay-sia.7 Since 1 February 2005, all “value-added network” service providers in South Africa were legally allowed to carry VoIP on their networks. While there were no specifically licensed VoIP providers, several companies began offering VoIP solutions there in 2005.

In less liberalized markets, the impact of VoIP wholesale transmission indicated how the market was changing. In 2004, up to a quarter of all operators in Africa were using VoIP to carry at least part of their international traffic (these agree-ments were politically sensitive, so establishing exact numbers was difficult). In late 2005, Telkom Kenya was preparing to offer a VoIP-based international service. Four African carriers

– BTC (Botswana), Mundo Startel (Angola), Telecom Namibia and UTL (Uganda) – announced plans to introduce IP-based networks, and the second national operator in South Africa was expected to, as well.

Mexican incumbent Telmex had already implemented IP for the majority of its core network, and various Mexican car-riers (Alestra, Avantel, Axtel and Protel) have been conducting initial trials pending legislative changes. Marcatel was already offering IP-based long distance services. Oman’s incumbent, Omantel, meanwhile, had committed to creating an end-to-end IP communications services network.

The transition to VoIP is so rapid and far-reaching that it is hard to make definitive statements about its progress. In some countries, legalized VoIP operators are already offering significant cost and service choices for both national and inter-national calls. In others, the process of liberalization has not yet begun, and the only “choice” for consumers is to look to grey market operators. Irrespective of national regulation, there has been rapid growth in VoIP services over the internet. Carriers such as Delta Three, Skype and Vonage have increased their global subscriber bases rapidly over the last three years.

Although there is little market data on the progress of this transition, it is useful to differentiate between the differ-ent types of VoIP service transitions that are occurring. At the wholesale level, there is a well-developed market for the car-riage of international traffic over IP networks. The calls actu-ally originate from a PSTN phone as circuit-switched traffic and are then converted to packet-switched data at gateway switches. They are routed internationally over IP networks to save money and increase efficiency, then converted back to

CHAPTER 6


96

circuit-switched calls at the destination. Domestic calls may be carried over an IP backbone, with the same conversions happening at either end of the call. Such calls may happen in parallel to traditional routing over circuit-switched (PSTN) networks. Packet-switched routing is essentially seamless to the consumer.

At a local level, VoIP is a much more challenging under-taking. Few end users have IP-enabled phones or software yet on their PCs or mobile phones. Also, the im ple men ta tion of enterprise-level, IP-enabled PABXs that can handle both time-division multiplexing (TDM) and IP in corporate markets varies enormously from sector to sector and country to country. Investment in customer premises equipment, both at a con-sumer level (by the individual user) and by companies (at an enterprise level) will take time. An unpredictable mix of cus-tomer choice, service availability and regulatory policy in each country will drive market take-up of IP-based services and equipment. The cost of the equipment needed – particularly at a local level – will only begin to come down once there is a sufficient volume of buyers.

The question posed by these developments is whether the transition to VoIP requires a revolution in regulatory thinking, or whether an incremental approach is wiser. The best way to consider this question is to differentiate between short-term changes, which are largely evolutionary, and long-term changes, which are more far-reaching. The transition from short-term to long-term changes is summarized in Box 6.2. The changes described as “long term” indicate that VoIP represents a major disruptive force for all tele com mu ni cation service providers. In other words, when VoIP becomes the primary transmission mode for voice calls, radical changes will need to be made in many regulatory models.

Clearly, regulators will need new knowledge and expertise. Nearly all regulators and staff are steeped in circuit-switched technology and services. In the future, regulators will need to understand the new, layered IP networks, the service concepts based upon them and the influence these will have on the future shape of the market. And because IP networks and VoIP services change rapidly, regulators are under greater pressure to make swift decisions and decide now on a course of action.

6.3 Grappling with Change: Regulators’ Responses to VoIP

Before looking at the detailed issues that VoIP raises for regulators, it is worth looking at the overall policy and regu-latory responses of different countries so far. Those responses are extremely varied, so this section groups countries under a series of headings corresponding to their approach to VoIP.

Some countries have adopted an incremental, evolu-tionary approach to VoIP regulatory issues, making modest adjustments to their regulatory frameworks. For others, VoIP represents a considerable threat to the established order (par-ticularly the international arbitrage aspect) and remains illegal. Some countries simply have not dealt comprehensively with VoIP yet. In China for example, the basic service operators (China Unicom, China Telecom and China TieTong) have

driven deployment of IP technology. But there is currently no specific VoIP regulation, and VoIP has not been classified as either a value-added network service or as a basic service. Basic tele com mu ni cation licensees are allowed to offer VoIP services and use IP technology in their core networks. China’s govern-ment, however, is considering banning the provision of VoIP services by anyone other than licensed operators. Cybercafés that offer long distance calls are deemed to be operating ille-gally, and basic service licensees are seeking ways to shut them down.8 Currently, ISPs can only offer PC-to-PC VoIP services.

6.3.1 A Liberalized Policy Approach to VoIP

Various countries have legalized VoIP services at differ-ent levels. For example, all forms of VoIP service are legal in Canada, the European Union, India and Korea (Rep.). The fol-lowing paragraphs provide other specific examples:

European Union: The European Regulators Group (rep-resenting regulators from 27 European countries) has agreed to a common statement on the regulatory approach to VoIP. According to the Group, VoIP should be used to enable (for the benefit of consumers) the greatest possible level of innova-tion and competitive entry in the market, while ensuring that consumers are adequately protected. VoIP rights and obliga-tions should be interpreted in accordance with the European regulatory framework, including the policy goals and regula-tory principles existing today. Consumers and service provid-ers should have enough information to make informed choices about services and service providers.

United States: The 1996 Telecommunications Act sepa-rated tele com mu ni cation services and information services and defined them differently. The FCC has formalized a policy of not imposing traditional tele com mu ni cation rules on internet applications, which are considered information services. At the beginning of 2006, the FCC was engaged in proceedings to examine issues raised by IP-enabled services, including VoIP. These proceedings were examining various social issues, such as how to adjust contributions to the Universal Service Fund, and the regulatory classification of services.9

Japan: VoIP is permitted and is subject to minimal regu-lation. The legal framework distinguishes three types of VoIP services based on the quality of the service. Providers that do not need numbers for their operation (that is, providers of PC-to-PC communications) do not have to comply with quality of-service (QoS) requirements. If the provider can ensure a minimum QoS (in terms of end-to-end voice quality and end-to-end voice delay), the provider can receive “050” – prefix numbers. Only if the quality is equal to traditional telephony can the provider use the same numbers as PSTN operators.

Tariffs and access charges for VoIP services are not reg-ulated in Japan. Interconnection is required only if the VoIP provider is a facility-based operator. VoIP providers have to pay access charges to terminate calls on the PSTN. It is also worth noting that Korea (Rep.) adopted a broadly similar approach to VoIP in September 2004.

Canada: Following a public consultation process, the Canadian Radio-television and Telecommunications Commis-

CHAPTER 6 97


sion (CRTC) published a decision in May 2005 that it would only regulate VoIP service when it is provided and used as a local telephone service.10 The CRTC reached its decision based on a principle of “service neutrality,” meaning that subscrib-ers could use VoIP service interchangeably with standard local exchange services.11 VoIP providers were classified in the same category as competitive local carriers.12 The CRTC’s decision called for:

• Registration of VoIP resellers,

• Access to numbers and local number portability,

• Access to directory listings,

• Equal access to interexchange carriers,

• Rules governing “win-back” marketing,

• Comprehensive assessment by VoIP operators of access for the disabled,

• Message relay service,• Privacy safeguards,• Tariff-filing requirements,• Contribution to the national service fund,• Non-dominant carrier regulation,• The development of IP inter con nection interface guide-

lines, and• The regulation of VoIP in areas where local competition

is not permitted (areas served by small ILECs and the Far North).

In line with its approach to retail internet services, the CRTC will not regulate computer-to-computer (peer-to-peer) VoIP services that reside solely on the internet.

Singapore: In June 2005, Singapore introduced a new policy framework for “IP telephony,” addressing the growing trend

Box 6.2: The VoIP Transition

Short to medium-term evolution (evolution from PSTN to IP networks)

Technical developments

– PSTN phone services and VoIP services exist in parallel.

– PSTN – IP network gateways are needed in most cases.

– Standard (E.164) numbers are (mainly) used; ENUM use of E.164 numbers increases.

– Terminals: Adapter + regular phone, IP-phone or a software-enabled phone.

Transition period for the market

– New, inter-modal competition develops, featuring cost-structure advantages, innovative services (in particular, nomadic use of VoIP) and lower charging models.

– Voice traffic is shifting to IP-based traffic and revenues from traditional phone services decline.

Regulatory model

– Changes are required to the current regulatory regimes to account for long-term influences.

– Regulators should balance basic main objectives:

• To enable the development of new services.

• To ensure acceptable social and consumer protection.

Long-term change (towards all-IP networks)

Technical developments

– IP networks and VoIP services are prevalent.

– Subscribers and services are addressed mainly by different types of internet addresses.

– E.164 numbers are likely to prevail at least in the global context.

– New types of terminals (e.g. combined GSM/UMTS/WLAN phones) support VoIP at home and in WLAN coverage areas.

– VoIP is normally one service inside a large service set.

Market structure and competition developments

– Integrated, innovative and personalized services dominate the market.

– Nomadic use is important, increasing the amount of cross-border traffic and services.

– Cost and revenue models of service providers have changed radically.

– There is separation between the transport network and the services delivered on top of that network.

Regulatory models

– New legal frameworks and regulatory models are needed.

CHAPTER 6


98

of using the internet and other IP-based networks to make local and international voice calls. The Info-Communications Development Authority (IDA) planned to issue licences and numbers for IP telephony providers. IDA’s regulatory frame-work called for minimal regulatory obligations. For instance, operators providing PC-to-PC telephony services were not required to provide number portability, emergency service access, directory services, or to conform to QoS levels. But operators must provide clear information to their subscribers regarding whether their offerings allow access to emergency services or meet the minimum QoS levels. Also, facilities-based operators were allowed to provide certain numbers only to users with valid Singapore addresses, safeguarding the integ-rity of Singapore’s national numbering plan.

South Africa: As of 1 February 2005, any holder of a “value-added network service” (VANS) or “enhanced service” licence was allowed to carry voice traffic on its network. Before that date, VANS providers were prohibited from carrying voice traffic. VANS providers can now apply for numbers, spectrum rights and inter con nection with any operator. Various ISPs and VANS operators have begun to offer retail VoIP, with aggressive advertising. There is no rate regulation of VoIP services directly, but the regulator is considering QoS issues and access to emer-gency services. It is worth noting that the planned second net-work operator’s service will be entirely IP-based, because of its deployment of a fibre network.

The Philippines: In August 2005, the National Telecommu-nications Commission (NTC) issued new regulations treating VoIP as a value-added service, for which only registration, not authorization, is required. Commercial VoIP providers with no network of their own are required to enter into inter con-nection agreements with network operators. Although such inter con nection agreements are to be negotiated between the parties, the NTC will intervene where necessary to ensure inter con nection is provided under fair terms. Carriers that have previously received authorizations were not required to register with the NTC when starting VoIP services.13

6.3.2 An Incremental Approach to VoIPMeanwhile, other countries have taken a more cautious

approach, making incremental changes rather than introducing comprehensive new frameworks.

India: VoIP has been legal since April 2002, under the designation of “Internet telephony,” which covers (1) PC-to-PC voice (both within the country as well as abroad), (2) PC-to-phone, and (3) “IP based H.323/SIP terminals.” Internet telephony through PCs or IP-based terminals is available also through India’s public “tele-info” centres and internet kiosks. Facility-based operators can provide internet telephony and use VoIP technology to manage their networks. Furthermore, the Telecom Regulatory Authority of India (TRAI) has issued regulations on quality for VoIP international long distance calls, differentiating between two quality levels: toll quality and below-toll quality. TRAI does not regulate tariffs of VoIP serv-ices offered by ISPs.

Bolivia: VoIP is considered telephony. In January 2005, a Bolivian ISP, Unete, announced an investment of USD 5 mil-

lion to launch a national and international long distance voice service.

Ecuador: VoIP providers are required either to have a licence for local or long distance public telephony or to estab-lish resale agreements with licensed operators. In February 2005, the regulator CONATEL published regulations covering cybercafes and tele-centres. The regulations limit the number of PCs that can be used for VoIP services to no more than 25 per cent of the total – or one PC if the cybercafé has only two or three computers.

Honduras: The regulator has allowed VoIP services, pro-vided that operators contract with the monopoly incumbent, Hondutel. The organizations that do this are described as “sub-operators” and they can use their own networks to sell other licensed services.

6.3.3 VoIP ConsultationsMany countries have not resolved how they will regulate

VoIP and are engaged in public consultation proceedings.

Chile: In July 2004 the Chilean regulator, SUBTEL, launched a public consultation, indicating that all services offered through some part of the PSTN network should be under PSTN regulations. But services provided entirely over the internet would not be subject to the same conditions. The regulator suggested creating a broadband voice licence that would cover VoIP. Some operators responded that the proposed classification was too rigid and potentially problematic in an increasingly converged environment. The Chilean incumbent argued that the introduction of VoIP would rob it of income while benefiting only a small group of the population.

Colombia: In June 2004, the Ministry of Communications issued a consultation document on VoIP services. The consul-tation was completed, but no action had been taken by August 2005. Prior to any results of the consultation, rules required operators to obtain a basic PSTN-service licence, but the use of a PC to make calls over the internet was not restricted. The Ministry was seeking to include VoIP in the existing public telephony category, but several parties in the consultation sug-gested that VoIP merited a new service category.

Jordan: In May 2005, the Telecommunications Regulatory Authority (TRA) of Jordan issued a consultation document on the delivery of IP voice services. This raised several issues for comment, including (1) whether or how to distinguish among different types of voice services that are identical to the con-sumer; (2) the provision of information to equipment purchas-ers and potential users; (3) the roles of network operators and service providers in maintaining network integrity; (4) geo-graphic and non-geographic numbering; (5) emergency service requirements; (6) inter con nection; (7) using class licensing for VoIP service providers; and (8) quality of service issues.

Hong Kong, China: In June 2005, Hong Kong’s Office of the Telecommunications Authority (OFTA) published a state-ment on the “Regulation of Internet Protocol (IP) Telephony.”14 This statement outlined OFTA’s position following a consulta-tion process begun in October 2004. OFTA said service-based providers should be allowed to compete with facility-based

CHAPTER 6 99


operators. Moreover, it upheld the principle of technological neutrality. Therefore, OFTA decided to introduce two differ-ent licenses for VoIP providers:

• Class 1 services – IP telephony offered with service attributes similar to those of conventional telephone services; and

• Class 2 services – Those that do not have the same attributes as conventional telephony.

Class 2 services were subjected to minimal regulation, although service providers were required to inform customers about the limitations of their services. Class 1 service provid-ers would have to fulfil basic telephone licensing conditions. Meanwhile, Algeria, Israel, Taiwan, China and Trinidad and Tobago have also begun consultation proceedings on VoIP. Kenya has issued guidelines legalizing various categories of VoIP, following public consultation.

6.3.4 Where VoIP is Illegal

VoIP remains illegal in quite a few countries.15 Where it is, governments adopt a variety of strategies to eliminate grey market operators. Some regulators seek to ban websites that allow users to make international calls. Others periodically confiscate the equipment of grey market operators. Some juris-dictions back up these sanctions with jail sentences.

Before the end of Panama’s telephone service monopoly in 2003, the Public Services Regulator required all ISPs to block IP ports identified with VoIP services. In addition, some-times telephone companies filter (block) VoIP services on their own. For example, one ISP in Mexico filters out VoIP service providers including Skype. Operators in Kenya have also fil-tered VoIP traffic. In almost all of the countries where incum-bents provide estimates of the grey market, the amount of “lost” traffic indicates that few of these blocking strategies work com-pletely. Usually, enterprising consumers and companies are able to access some form of VoIP service.

6.3.5 Classifying VoIP Services for Regulation

As can be seen from examining the current state of VoIP regulation, a key question for regulators is how to classify vari-ous forms of VoIP, IP telephony or “Internet telephony.” Any rigid classification set at this juncture is unlikely to be stable, given the pace of technological and market-driven change. Any classification used also depends on national policies and legisla-tion. Box 6.3, however, offers three general categories.

Below are various examples of how different countries have treated classification of VoIP services:

The European Union: The regulatory framework addresses the question of how communication services should be clas-sified in two ways. The Universal Service Directive classifies services for the purpose of consumer and social protection. For market and competition management, the need for regulation is assessed by analysing whether VoIP services have “significant market power” (known in some other countries as “domi-nance”) in one or more “relevant markets.” The services cov-ered by the Universal Service Directive are divided into two categories:

• Electronic Communication Service (ECS) – includes services provided for remuneration and consisting wholly or mainly in the conveyance of signals on Electronic Communications Networks. ECS is treated with lighter regulation.

• Publicly Available Telephone Service (PATS) – con-sists of a service that includes all the following functions: available to the public; used for originating and receiving national and international calls; and provides access to emergency services through a number or numbers in a national or international telephone numbering plan. PATS attracts more regulation and obligations. The main obliga-tion compared with ECS is the provision of emergency calls.

The EU regulatory framework seeks to be technology-neutral, although some of its rules are based on traditional telephone technologies. Debate on the classification of VoIP services is ongoing. Meanwhile, different European govern-ments seem to be interpreting the regulatory framework for VoIP services in divergent ways. These vary from a flexible reading of the rules to a strict interpretation of the framework’s wording. The following approaches are evident:

• One approach, apparent in the United Kingdom, is a flex-ible classification regime, in which service providers, not regulators, decide in which regulatory category they want to be classified;

• Another approach distinguishes different types of services based on a strict interpretation of the PATS definition. That is, PATS obligations are imposed on a voice teleph-ony service only if all four parts of the PATS definition are fulfilled;

• A compromise approach is to classify a service as PATS if it is available to the public for originating and/or receiv-ing national or international calls through a standard telephone number. Access to emergency services is not regarded as being an integral part of the PATS definition, and whether to mandate such access can be decided sepa-rately.

For competition policy purposes, each EU national regu-lator assesses how different VoIP services fit into the EU’s

“relevant markets” categories. At present, there is very limited practical experience and precedent for VoIP, and the EU is cur-rently discussing market analyses that include VoIP services.

Canada: The classification of VoIP services stems from discussion of emergency call services. From this perspective, there are currently three different types of VoIP service offered to customers: fixed, nomadic and foreign exchange. Users of fixed VoIP service can only place a telephone call from the location where the service is provided. Users of nomadic VoIP service can make calls from any location where users can access their service via the internet. Foreign exchange VoIP service allows users in one exchange to receive telephone calls dialled as local calls in another exchange that they have selected (for example, a customer can be reached in Ottawa by calling a Halifax local telephone number). These different types of VoIP services have different obligations regarding emergency calls.

CHAPTER 6


100

United States: There has been substantial debate about how to classify (and therefore regulate) VoIP services. The country’s Communications Act (as amended by the 1996 Telecommu-nications Act) applies more stringent regulation on “tele com-mu ni cations services” services than on “information services,” and there traditionally has been reticence to overly regulate internet-related services. From the FCC’s perspective, services that are only provided over the internet (like MSN Messen-ger and Skype) are classified as information services, and even where VoIP services have a gateway to exchange traffic with the PSTN, they should only be regulated in relation to emergency calls and lawful interception.

6.4 Crafting New Regulatory Approaches to VoIP

6.4.1 Balancing Different Policy Needs

The transition to IP networks and VoIP services tends to produce conflicting policy approaches in different countries. This section defines some of these conflicts, which tend to reflect opposing policy goals and sector objectives.

The main challenge is to balance short- and long-term policy and regulatory approaches. In some countries, VoIP is seen as a major threat to established operators because it under-cuts their domestic and international long distance rates and radically reduces their revenues. But strict regulation of VoIP in the short-term to protect the incumbent’s revenues might harm the long-term development of the sector. Often, this drive to protect the incumbent stems from a particular social policy, such as extending universal access. But regulators need

to consider not just the established operator’s point of view, but also the welfare of end users and potential new market entrants. The impact of lower prices from competition brought about by VoIP will directly benefit consumers. Competition will also help to increase the number of overall subscribers and boost the volume of usage on all networks. In these circumstances, it is important that policy decisions be based on trend data and that the regulator analyse different market scenarios.

Regulation should promote competition, but it also needs to protect consumers and take social concerns into account, including universal access objectives. Regulators may find themselves torn between the need to react quickly to new con-cerns and stepping back to see the shape and dynamics of the emerging market. Many regulators are guided by a principle of limiting regulatory obligations so as not to discourage market entry or the introduction of innovative new services. This leads to efforts to implement “light-touch” regulation in order to encourage market and technological experimentation. For example, some developing-country regulators have encouraged the use of free or low-cost spectrum, or reduced licence fees, in areas where there are little or no voice or data services.

A vital task for regulators and policy-makers is to manage the transition to the new world of IP networks. The most basic questions to address include:

• How long should a PSTN network be maintained and supported?

• How much time is needed to make changes to existing legislation and regulations, and how can legal stability be maintained in a time of technological flux?

Box 6.3: Classifi cation of VoIP Services

Regulators around the world face the challenge of trying to differentiate between types of VoIP services and capabilities. One approach to classification of VoIP follows:

Category I

VoIP offerings that do not require regulation because there is no clear service provider. This would cover PC-based VoIP com-munications installed by end users via software downloads (Examples include GIZMO, Yahoo Instant Messenger and Skype).

Category II

VoIP offerings that are outside the scope of regulation because they are not directly offered to end users or do not impact the PSTN, including:

– Corporate private networks, where VoIP is used to provide internal communications,

– IP technologies used within a public operator’s core network, but which do not affect the retail services offered to end users.

Category III

This category would cover publicly available services provided to end users. There are many different kinds of publicly avail-able VoIP service offerings, and the regulatory treatment depends on the nature of the service being offered and relevant national legislation.

A large number of national regulatory agencies have carried out consultations on VoIP issues but have not reached any final classification decisions. Many are trying to adapt service classifications provided under existing tele com mu ni cations legislation for VoIP services. There appears to be a consensus, however, that VoIP services residing solely on the internet (PC-to-PC calling) should not be regulated. Global discussion of these issues focuses on approaches to VoIP services that are similar to those described in category III above. The basic regulatory question that hovers over the whole discussion is whether or not VoIP can be regarded as a substitute for PSTN telephony.

CHAPTER 6 101


• How quickly should competition policy change to reflect convergence on IP-based networks?

Governments do not have the luxury of setting the pace at which they will resolve these questions. Market changes are already under way, and they are likely to proceed whether they are legally sanctioned or manifested in grey market activities. Unfortunately, there is no consensus on how to address regu-latory change prompted by the emergence of VoIP, either at a global or a national level. The following subsections examine a number of potential responses in key issue areas.

6.4.2 Market Entry

In countries where tele com mu ni cation competition is already legal, regulators will have to manage the range of issues associated with the transition of services to IP networks. In those countries where the PSTN remains an exclusive pre-serve of the incumbent, regulators face a more complex task. These regulators have to manage two simultaneous, parallel processes: (1) the transition from a monopoly to a liberalized market, and (2) the transition from a circuit-switched PSTN to an IP-based environment.

In the past, regulators largely dictated market changes, and did so at a measured pace. In the future, however, regulators will face a rapidly changing market based on a new type of infra struc ture. They will need to deal with a large number of new services entering the market that may never have existed before. The policy and regulatory framework may be influ-enced or changed by international market developments that regulators cannot control or even influence. So it is becom-ing increasingly difficult to plan or orchestrate market devel-opments in advance. There are widely differing strategies for addressing market entry.

The European Union: Here, the approach is to facilitate easy market entry by requiring individual licences only for scarce resources like spectrum. For other network or service devel-opments, only a notification or registration is required. The EU also adopts a technology-neutral approach, leaving market players to decide themselves what technology to deploy.

South Africa: The government has created conditions for easier market entry in Value-Added Network Service (VANS) or enhanced services, but it retains control over the market structure for fixed, mobile and satellite services. Under the current Telecommunications Act, only VANS and “Private Telecommunications Network Service” (PTNS) licences may be issued on a non-exclusive basis. Fixed-line service has been dominated by the incumbent, Telkom, with a second entrant set to commence operations in 2006. Mobile voice and data services are supplied by three operators.

Until 1 February 2005, Telkom had a full monopoly on facilities provisioning, and the resale of spare capacity for VANS and PTNS was prohibited. Since then, VANS licensees may obtain alternative facilities from other operators. New regulations for VANS gave providers the right to access spec-trum, apply for numbering allocations and inter con nect with other operators. Numerous VoIP providers are emerging in the South African market as a result, although the regulator has

not yet finalized a policy on numbering and spectrum access for VANS.

Other countries restrict market entry of new VoIP opera-tors in various ways. In many countries, ease of market entry depends on whether VoIP is defined as a voice or informa-tion/data service. Where VoIP has not yet been categorized, the issue is often under extended debate.

Brazil: The regulator, ANATEL, has not defined VoIP as a tele com mu ni cation service, a value-added-service or a technol-ogy. If VoIP is considered a data service, operators will need a licence for multimedia communication services. And in order to initiate and terminate calls outside of a private network, operators need a licence for public switched fixed telephony, which entails certain requirements for coverage and QoS. Moreover, the licence application process is more complex.

Some countries, such as Colombia, Egypt and Nigeria, are pragmatic about certain types of VoIP such as like PC-to-PC telephony, which is regarded as a use of personal computing equipment that would be almost impossible to control. Other countries, such as Guinea, have legalized the use of VoIP over virtual private networks (VPNs), which, again, is hard to detect and control. In Nigeria, the regulator has stated that VoIP is legal, provided operators obtain the appropriate licences. As a result, one satellite operator with an international licence is offering a VoIP service to its customers. Yet, other countries completely restrict any form of VoIP market entry and seek to control grey market operators through a variety of strategies.

6.4.3 Numbering Resources

Telephone numbers can be used for several different pur-poses. They can be used to differentiate between services and inform users of tariff categories like premium call services. Numbers can also be used as a tool to control markets by set-ting restrictions on the use of certain numbers. So access to numbers – or more correctly, withholding of access to them

– can become a barrier to market entry.

VoIP calls can be routed to end users in several different ways: IP addresses, SIP addresses, H.323 addresses or E.164 numbers. Traditionally, E.164 numbers have been needed to originate and receive voice calls, but they may lose their dominant position in the future and become just one of many options. E.164 numbering ranges are usually divided into sev-eral generic types, indicating the services that may be offered using them. Geographic numbers or special number series are regarded as most relevant for VoIP services. Also, mobile, per-sonal and corporate numbers can be used to address VoIP sub-scribers. These are, however, seen as less attractive in many countries because users will associate them with high retail calling prices.

The current position on the availability of geographic numbers for VoIP services varies among countries. The main argument in favour of allocating geographic numbers to VoIP services is that they offer the best support for competition, especially when combined with number portability. The main arguments against this approach have been the nomadic nature of VoIP and potential for exhausting geographic numbering

CHAPTER 6


102

resources. There are three ways to allocate geographic num-bers in order to support VoIP services: • Allowing nomadic access in a limited area; • Allowing nomadic access countrywide but requiring some

relationship with the geographic area of the number; or • Removing any requirement for a relationship to a geo-

graphic location.

Regulators may also open new number ranges for nomadic VoIP services. Broadly speaking, there are three types of possi-ble new number ranges: (1) a general-purpose number range; (2) a number range for nomadic services, and (3) a number range for Electronic Number Mapping (ENUM)-based or similar, software-based services.

ENUM is a protocol developed by the Internet Engi-neering Taskforce (IETF) that defines a domain name system (DNS)- based architecture and protocol aimed at using a tel-ephone number to look up a list of IP service addresses (e-mail, IP phone addresses, URL, SMS, etc). The idea of ENUM is to use an E.164 number as the key to identify the available com-munication services to contact a person. For VoIP, it is used to route a VoIP call to an IP network based on the receiver’s E.164 number.

The idea of creating new number ranges to accommodate VoIP seems to have mainly been motivated by a number of factors, including avoiding the impression that these are high-tariff numbers like those for mobile phones. Also, regulators may want to keep existing number ranges intact and avoid having them depleted.

In Europe for example, the number ranges open for VoIP use vary due to different numbering policies and the regula-tions related to them. The geographic number ranges are open for VoIP services in most countries. But some countries list a number of requirements that have to be fulfilled in order for VoIP providers to obtain numbers.

Meanwhile, the cost of numbers (national numbering fees) can be a significant barrier to market entry in some coun-tries. Geographical numbers are typically allocated in blocks (normally blocks of 1000 or 10,000). Where geographic num-bers are sold with limited nomadic capability, VoIP service pro-viders may need to get number blocks that cover the whole country. The costs of doing so can be high enough to consti-tute an entry barrier for small providers. In Europe, numbering costs vary greatly.

Number Portability is a key enabler of competition because it allows users to retain their telephone numbers when they change service providers. Regulators must decide whether VoIP providers are included in number portability systems, making them co-equal with all other competitors.

6.4.4 Emergency CallingThe nomadic capability of VoIP services poses a problem

for the provision of emergency calls because it breaks the link between the calling party and the location. In the PSTN envi-ronment, calls routed to the emergency response centre carry information on the location of the caller. With VoIP calls, how-ever, there is – at least currently – no automatic way to convey

location information about the user calling an emergency serv-ice. Technical solutions, however, are expected to emerge from the current stand ardization work that is being undertaken.16

The problems of handling emergency calls from VoIP users can be divided into two categories: (a) emergency calls made within a country and (b) cross-border emergency calls. The first category is likely to be less problematic, due to the likelihood of increased cooperation between service providers covered by national regulation in a single country.

Cross border VoIP emergency calls raise more difficult issues, because an emergency call needs to be identified as such. Globally, there are more than 60 national emergency call numbers (for example, “911” in the United States and “112” in the EU). For an emergency call to be routed to the right coun-try and the correct emergency centre, intensive international agreement, cooperation and arrangements will be required. Cross-border emergency calling is mostly an issue, of course, in places such as Europe, which are interlaced with multi-ple national borders. In North America, by contrast, the vast majority of emergency calling is entirely domestic, and cross-border arrangements are less of an issue.

From a consumer point of view, the best possible situa-tion is to be able to reach emergency services from any hand-set, including a VoIP-enabled phone. Technical solutions will be required to overcome any limitations in routing of emer-gency calls in a VoIP environment. A key question is what will happen with VoIP connections during electricity outages, espe-cially where VoIP is the only option for making a call in an emergency situation. Various principles may be considered to address this issue:• If VoIP handsets or terminals are likely to be used for

making emergency calls, they should be equipped to do so;

• An emergency call from a VoIP handset or terminal should reach an emergency centre in the country in which the call originates;

• Where possible, an emergency call from a VoIP handset or terminal should reach the specific emergency call centre that is responsible for receiving emergency calls for the area in which the caller is located;

• The VoIP call made to an emergency centre should carry Calling Line Identification (CLI), which can be used to call back the person reporting the emergency if the person is disconnected before full information has been pro-vided;

• Where possible, the number provided by CLI for an emer-gency call from a VoIP handset or terminal should not be linked to location information that is incorrect or mislead-ing.

The following examples of VoIP emergency calling sug-gest various approaches:

European Union: The Universal Service Directive allows for technical feasibility when imposing obligations relating to the provision of location information. It states that location information must be “handled in a manner best suited to the national organization of emergency systems and within the

CHAPTER 6 103


technological possibilities of the networks.” Authorities should “make caller location information available to authorities han-dling emergencies, to the extent technically feasible, for all calls to the single European emergency call number 112.”

How the Directive is actually applied varies, depending on national emergency systems and the capabilities of the net-works involved. Similarly, related legal requirements – such as providing caller location information, routing calls to an appropriate emergency centre and providing CLI – vary greatly between countries. Where VoIP services are regulated, some countries have set the same legal requirements for both VoIP and PSTN calls, although they provide a temporary techni-cal reprieve from these requirements for VoIP calls. Currently, nomadic VoIP service providers can only meet national legal requirements for emergency services in a few countries.

Canada: CRTC decided in April 2005 to require provid-ers of fixed (non-nomadic) VoIP service to provide the same level of emergency service that incumbent telephone compa-nies provide to their existing customers (either “Enhanced 911” or “Basic 911” service). Implementation was made mandatory within 90 days.17 CRTC required providers of nomadic and

foreign exchange VoIP services to provide an interim solution equivalent to basic 911 service. In addition, the Commission required all VoIP providers to inform their customers about any limitations of their offerings in providing access to 911 service. The VoIP service providers must secure the customer’s express ac knowl edgement that they are aware of these limita-tions before starting service.

United States: The FCC issued an order in May 2005 requiring inter con nected VoIP providers to provide Enhanced 911 Service (See Boxes 6.4 and 6.5). The order affects providers of services that are functionally equivalent to circuit-switched telephony. For example, the FCC included service provider Vonage in this category, because its service enables customers to send and receive calls from the PSTN. The order did not impose E911 obligations on providers of other IP-based serv-ices, such as instant messaging or internet gaming. Although these services may contain a voice component, the FCC said, customers of these services cannot use them to exchange calls with the PSTN. The FCC has also stated its intention to adopt, in a future order, an advanced E911 solution that automatically determines a customer’s location.

Box 6.4: Emergency Services under the North American Numbering Plan

Both Canada and the United States are among those countries that follow a regional numbering scheme, known as the North American Numbering Plan (NANP). Consequently, both countries have a standard emergency calling number: “911.” The exist-ing local telephone networks currently provide two types of 911 services: Enhanced 911 and Basic 911. Enhanced 911 (sometimes known as “E911”) service automatically sends customer location information to an emergency centre, where an operator dispatches a response service. E911 allows dispatchers to send responders to the correct location even if the caller is incapacitated, agitated or otherwise unable to provide an accurate description of the location. The older Basic 911 service, meanwhile, connects the caller to a central call centre, which then connects the call to the correct emergency response centre. At that point, the caller must be able to identify or accurately describe the location in order for responders to be dispatched to the right place.

Box 6.5: The FCC’s Enhanced 911 Service Order1

Provisions of the U.S. Federal Communications Commission’s May 2005 order on VoIP access to emergency services in-clude:

• Interconnected VoIP providers must deliver all 911 calls to the customer’s local emergency operator. This must be a standard, rather than an optional feature of the service.

• Interconnected VoIP providers must provide emergency operators with Calling Line Identification and location information for their customers (an Enhanced 911 service) wherever the emergency operator is capable of receiving it. Although the customer must provide the location information, the VoIP provider must provide the customer a means of updating this information, whether he or she is at or away from home.

• Interconnected VoIP providers must inform their customers, both new and existing, of the Enhanced 911 capabilities and limitations of their service.

• The incumbent local exchange carriers are required to provide access to their Enhanced 911 networks to any requesting tele-com mu ni cation carrier. They must continue to provide access to trunks, selective routers, and E911 databases to competing carriers. Interconnected VoIP providers must submit to the FCC a letter detailing their compliance no later than 120 days after the effective date of+ the order.

1 Federal Communications Commission, 911 Services, at http://www.fcc.gov/911/enhanced/

CHAPTER 6


104

6.4.5 Universal Access and Universal Service

As customers switch to VoIP and disconnect from the PSTN, regulators in many countries fear that a decline in revenues from traditional carriers will undercut funding for universal service funds, which often rely on mandated carrier contributions. Of course, this begs the question of whether VoIP providers should be included in universal service funding arrangements. There are two issues at stake here: (1) potential contributions to universal service funds by VoIP providers, and (2) whether VoIP can be an effective tool to deliver cheaper calling to a wider number of consumers.

United States: The FCC is confronting a multifaceted chal-lenge to the financial health of the Universal Service Fund (USF), stemming from lower wire-line telephony revenues. The Commission is considering overhauling the contribu-tion formula, which has been revenue-based. One suggestion is to require contributions for any service that requires use of a telephone number – which could take in some VoIP serv-ices. While the service providers would make the contributions, they could pass the costs along to customers (as carriers now commonly do) in retail fees or charges.18

Various other countries that allow VoIP, including the Czech Republic, Mauritius, the Slovak Republic and Venezuela, levy universal service/access contributions on operators. In Canada, the CRTC has ruled that if the VoIP service provided allows for access to and/or from the PSTN, the service is considered eligible for contribution requirements, even if the customer uses the service to make peer-to-peer (PC-to-PC) calls.

Regulators have to consider the potential consequences of their decisions on funding of universal service programs. If they continue to levy contributions on a narrow base of incum-bent or traditional telephony revenues, they risk overtaxing a declining base as subscribers flee to VoIP. But if they broaden the base to include VoIP providers, they may raise the bar for market entry by potential voice service competitors.

Indeed, IP-based transmission is arguably less costly and more efficient than circuit-switched transmission. Using the same bandwidth, a VoIP network can carry many times the number of voice calls as a circuit-switched network, making the transport cost, per bit of information, lower on packet-switched networks. For that reason, VoIP can be embraced as a new tool in achieving universal service and universal access, at least from an affordability standpoint. The Association of Infocentres of El Salvador (Infotel), for example, is launching a VoIP service for international calls using pre-paid cards. The service will be available in 41 of the Association’s centres. The initiative is supported by the Salvadoran regulator as a means of reducing international calling costs. Similarly, the state-owned Telecommunications Office (Telof) in the Philippines plans to launch VoIP services in un-served rural areas.19

6.4.6 Competition

Consideration of how to regulate the VoIP market raises the question of what kind of competition issues might arise specifically in relation to VoIP. There are several. For example, the development of VoIP service depends greatly on the avail-

ability of broadband access. Regulators have to ensure open, non-discriminatory and fair-priced access for ISPs wanting to resell broadband access.

Another challenge is to prevent incumbent operators from blocking access to VoIP services by closing ports used by VoIP services or refusing to lease facilities. There is no consensus at this juncture about which regulatory tools should be used to ensure fair competition. But it is important for regulators to prevent anti-competitive behaviour against VoIP providers as they apply downward pressure on tariffs.

European Union: Market regulation is based on the cir-cumstance in each market. The tele com mu ni cation sector is divided into 18 “relevant markets,”20 which can be analysed by the national regulator in each member country. If robust com-petition is present, no regulation is applied, but if the regula-tor finds an operator with “significant market power,” it will implement regulatory remedies to prevent that operator from abusing its market power.

Several European countries are in process of considering whether VoIP should be included in the relevant market cov-ering fixed telephony. Again, results vary considerably, but in some countries, VoIP services with a gateway to the PSTN are considered part of the fixed telephony market for the purposes of competition analysis. The whole question of where VoIP services will be placed in terms of “relevant markets” will soon be considered by the European Commission.

Canada: CRTC also has decided that VoIP should be seen as part of fixed telephony market. But it will only regulate VoIP services when they are provided and used as a local telephone service. This decision was aimed at building sustainable com-petition in the local telephone market. Incumbent local loop carriers (which are held to possess market power) are prohib-ited from pricing their local services below cost as a way to stifle competition.

6.4.7 Interconnection

6.4.7.1 VoIP’s Challenge to Current Models

VoIP services challenge current inter con nection models in several respects, both from an economic and structural per-spective. Again, it is helpful to separate what may happen in the short-term, during the transition period to IP networks, and in the longer-term, when most networks have become IP-based. In the short-term, inter con nection will be mainly between IP networks and the PSTN. But over the long term, inter con-nection will increasingly occur directly between IP networks.

Short-term inter con nection issues: Current inter con nection models are increasingly based on cost-oriented charges. VoIP may disrupt this because of the difference in investment scale between IP-based networks and PSTN networks. Although views differ as to the exact scale of difference, the cost of IP networks is significantly lower. So there is a possibility of vari-ations in call-termination rates based on cost differences. There are not likely to be any issues when the call comes from an IP-based network to the PSTN, because the receiving PSTN operators normally charge the same termination fee regard-less of which network the call comes from. But when the call

CHAPTER 6 105


originates on the PSTN and terminates on an IP network, the termination cost is difficult to determine. The relevant ele-ments to assess the actual costs are unclear. Thus PSTN-to-IP inter con nection is likely to generate a difficult debate between the different players, requiring regulatory oversight and inter-vention.

Long-term inter con nection issues: In the long term, when IP-to-IP inter con nection is predominant, the application of cur-rent telephony inter con nection models will create a number of problem areas. These are examined below, and it should be noted that these issues apply to all IP services and not just to voice calls.• Support of new IP-based services: IP-based networks

are expected to support new services, including third-party services. The existing, usage-based format for inter-con nection would call for inter con nection agreements and charging arrangements with each service provider, at each inter con nection point. But there is a basic technical problem: there is no way (at least in the near future) to transmit the charging information between IP networks. These kinds of practical issues could constitute a barrier to the roll-out of new services if changes are not made in the inter con nection model.

• Changes to cost structures: Developments in tech-nology and huge economies of scale have resulted in the substantial decline of core or backbone network costs. The existing regulatory and commercial models assume an expensive backbone network, which is why there is so much emphasis on competition and pre-selection for long distance and international calls. With less-expensive IP backbones, there is a need to consider adopting capac-ity-based charging instead of usage-based prices for basic connectivity. A simple approach would be to make one of two changes: either separate charges for services and con-nectivity, or adopt “bill and keep” inter con nection.

• Changes to the retail market: The existing retail market is changing. Call prices are dropping, and some opera-tors are starting to offer flat-rated tariffs with unlimited call volumes for a fixed subscription. This sets up the risk of arbitrage. Operators would benefit from having inter-con nection arrangements that better match the structure of the retail charges. This change is causing many com-mentators to say that the days of time-based call charges are rapidly disappearing. But there are still a number of high-priced, time-based charging structures (for mobile calls and premium-rate services), and providers of these services will likely seek to continue these revenue streams or find substitutes, such as metered charges for video or music downloads.

6.4.7.2 New Approaches to Interconnection

These developments suggest the need to look for new approaches to inter con nection. A variety of ideas are under discussion, and there is a lively debate between the “telecom world” and the “internet world” on two basic approaches:a) An open, internet-type approach, in which the separation

of service provision and connectivity occurs as it does on

the internet. This would require separate consideration of:• Interconnection (inter op erability) at the service level,

where services are charged on the “bill and keep” prin-ciple (peer to peer), and

• Interconnection at the connectivity level, where charg-ing between networks is based on capacity charging or another similar method;

b) A next-generation network (NGN) architecture approach, in which network operators have more control over, for example, service quality (such as providing different cat-egories of guaranteed bandwidth), security (such as cus-tomer ID, authentication and security tunnelling) and charging for services by third parties. NGN architecture includes additional software, which is not present on the basic internet network, such as IP Multimedia Subsystem (IPS), which controls the inter con nection of services to networks. This means that in NGN architecture – since network and application services are separate – network operators can get a share of revenues from application services.

In the future, there will likely be more than one inter con-nection model. Market players may be able to choose the one that best fits their needs and situations. During the transition period to IP networks, new inter con nection models will per-haps exist in parallel with older arrangements, just as the inter-net’s current charging model works alongside the telephony charging structure.

6.5 End User and Consumer Issues Because VoIP services ride on an entirely new network

and transport model, there are inevitably issues with regard to duplicating or recreating some of the salient features of the old environment. These issues relate to quality of service (QoS), network security, privacy and access for legitimate law enforce-ment purposes. This section examines how regulators and other government authorities are addressing these issues.

6.5.1 Quality of ServiceOne of the requirements for the deployment of VoIP net-

works is the ability to offer toll-quality service equivalent to the existing PSTN. The QoS standard for VoIP can be defined in several ways, depending on whether it is considered from an end user or a technical perspective. The end users’ perceptions of service quality can be measured through subjective quality assessment. The most common consumer-based method to quantify QoS is the “Mean Opinion Score” (MOS) developed by ITU-T.

6.5.1.1 End-to-End Quality

Discussions of VoIP QoS typically highlight the issue of increased end-to-end delay and discuss the effects of this delay in interfering with the normal cadence of voice conversations. People expect their conversations to be in real time. If delays occur, people begin to talk over each other and conversations begin to break down.

CHAPTER 6


106

This problem is actually a familiar one, however, because of satellite latency on international calls and strained reception on mobile phones. The delays and dropped calls in the mobile environment are, in fact, more marked than those stemming from VoIP use, and they are tolerated by users as the price for mobility and convenience. Moreover, mobile phones are the default level of QoS expectation in most developing countries, where mobile telephony is the dominant voice medium. This may be particularly true in Africa, for example. And although VoIP call quality is still inferior to analogue or circuit-switched systems, many PSTNs in developing countries offer call qual-ity well below what is experienced in developed countries, anyway. Seen in this light, QoS is a relative concept.

Delay is still a major issue for digital voice transmission, but other parameters need to be included in QoS for voice transmission evaluation. The combination of these parameters will therefore define the end-to-end quality:

– Jitter, which is the variation in the time between packets arriving, caused by network congestion, timing drift, or route changes. A jitter buffer can be used to address this;

– Packet loss, which introduces audio distortions;

– Speech coding and decoding, which generate an approxi-mation of the original signal.

6.5.1.2 QoS in Practice

From the consumer’s perspective, QoS is really about a series of trade-offs involving lower costs and other advantages. Extending the comparison with mobile phones, consumers appear willing to pay more for the added value of mobility than for fixed-line service. Similarly, a consumer using VoIP is often willing to trade call quality for lower-cost (or free) calls. In developing countries, grey market users are making exactly the same trade-off.

From this perspective, QoS for VoIP seems more like a function of market forces, and in fact, many operators make great efforts to maintain the QoS at the highest possible level, so as to attract subscribers. If consumers are unhappy with the quality of any particular service, they will cease to use it. Many VoIP applications may not normally meet the mission-critical voice-quality standards required by corporate clients – unless, of course, service-level agreements are in place to provide (for a price) higher-level quality performance and network integ-rity.

If VoIP QoS appears to be a matter of choice, trade-offs and market demand, it begs the question of whether QoS is a major issue. Still, regulators are examining it, particularly in developing countries, where VoIP is emerging as a low-cost communication solution.

6.5.2 Regulating Network Integrity and Security

6.5.2.1 Network Integrity

The term network integrity refers to the inherent reliability of a network and its resilience to external threats, such as natu-ral disasters or malicious acts. The main regulatory issue arises from the fact that VoIP can be provided independently of the

underlying network access. This network/service independ-ence has several implications:• VoIP can be provided over an access network without the

operator of that network being aware, or having any con-trol over, the voice service provided;

• VoIP can be provided over an IP network using any access technology;

• VoIP can be provided over an IP network at any location.

This fundamental independence from network operations has the potential to raise complications for regulators, because existing network integrity requirements usually were devel-oped when the network and the service were not independent. VoIP services have introduced the possibility that calls can be provided independently of the access network provider.

The European Union’s network integrity requirements are set out in the Universal Service Directive, which calls upon member countries to take “all necessary steps” to ensure the integrity of the public telephone network at fixed locations. The member countries are supposed to ensure that providers of publicly available fixed telephony “take steps to ensure unin-terrupted access to emergency services.” This raises questions about how VoIP services fit into these requirements:• Are “nomadic” VoIP services (ones that can be used over

any IP access network) provided at a “fixed” location?• What are the implications of network/service independ-

ence?• What are “all necessary steps” to ensure the integrity of

the public telephone network at fixed locations in a VoIP context?

These issues are under discussion in European countries. Thus far, there is no common European regulatory approach covering VoIP and network integrity.

6.5.2.2 Electronic Communication Security

Electronic communication security is a broad subject that can be used to address a large variety of issues. There are eight dimensions of network security:• Access control,• User authentication,• Non-repudiation,• Confidentiality,• Communication security,• Data integrity,• Availability, and• Privacy.

On application layers of an IP network, the operation of each application (web browsing, e-mail, domain names) brings with it its own security questions. In each case, specific actions are taken to minimize risks, such as filtering software for e-mail.

The PSTN is, by nature, a closed network with con-trolled security and privacy. An IP network is based on open network architecture, with resulting communication security vulnerabilities that can affect VoIP as much as any other service.

CHAPTER 6 107


Box 6.8 discusses some mechanisms that can be used to guard against attacks.

6.5.3 Lawful Interception

There are times when every nation’s law enforcement authorities may have justifiable and legal cause to intercept

tele com mu ni cation traffic. The regulatory aspects of lawful interception of VoIP services are complex. In the future it will be vital for law enforcement agencies to be able to monitor and intercept internet-based voice traffic, but the VoIP serv-ices sector is much more fragmented than the large telephony operators and involves multiple types of network platforms.

Box 6.6: SPIT: A Looming Issue

An issue that has recently begun to emerge and may require further consideration is “Spam over Internet Telephony,” or “SPIT.” Related to the larger problems of Spam and “SPIM” (Spam over Instant Messenger), SPIT is essentially junk mail in voice form – unsolicited voice messages and unwanted advertising or marketing. This problem has a unique element, in that VoIP has a broadcast capability. Views differ on the future scope of the problem. Some network security software products are being developed to incorporate SPIT-blocking technology in future releases. But some operators and analysts are less concerned about SPIT, because messages have to be streamed to a network, as opposed to simply being mass emailed.

At the same time, others have noted that standard content filters used for spam would be very difficult to apply to voice. The variability of phrases and pronunciation make algorithms difficult to write, and the technology lends itself to extremely cost-effec-tive solutions for telemarketers. While SPIT is just barely emerging, it is not a new problem. It can be addressed as part of a general approach to voice security in the IP space that addresses multiple problems and vulnerabilities – including viruses and “denial of service” (DoS) attacks – that have been identified with Session Internet Protocol (SIP).

Box 6.7: Threats to VoIP Networks and Publicly Available Services

• Distributed Denial of Service (DDoS) attacks threaten the availability of the VoIP network by flooding it with unnecessary data or attacking the key network elements. DDoS attacks are typically launched from a large number of compromised client machines and are difficult to defend against in the light of VoIP QoS requirements;

• Thefts of call information can occur by breaching vulnerable VoIP signalling servers. The call information can be as valuable as the content, so it is likely to be a target for attackers;

• Conversation eavesdropping or recording can occur by breaching VoIP network gateways or other key network points. Software plug-ins required to sniff out VoIP traffic are available on many open-source websites for free;

• “Call hijacking” or “man-in-the-middle” attacks can occur. These scenarios involve rerouting the connection or modifying call parameters;

• Identity “spoofing” can occur through manipulating Caller ID;

• Attacks can be made against the terminal equipment software, devices or network servers themselves. The software on these devices can be vulnerable to the same types of vulnerabilities that affect all operating systems software.

Box 6.8: Defense Mechanisms against Security Attacks

Communication security depends on both the actions of the end user and the security practices of the VoIP service provider. Security is always a compromise between the utility and cost of the service and the protection mechanisms that are available. In order to mitigate security risks, the VoIP service provider should consider at least the following measures:

• VoIP networks should be logically separated from other IP services and applications;

• VoIP servers should be hardened and treated using the same security precautions as any other servers that contain confiden-tial information and offer network services;

• VoIP networks should be redundant, in order to ensure the availability of the service. The VoIP network has to be resistant to denial-of-service attacks. This is especially essential for emergency services;

• Encryption of VoIP traffic can be used whenever reasonable. Encryption can be implemented on the application, transport or network levels;

• Network devices should be configured properly to restrict unnecessary traffic toward VoIP systems and to ensure the opera-tion of VoIP services.

CHAPTER 6


108

The issue of how to manage lawful interception is likely to become more urgent, as operators move to “next-generation network” architectures.

Several countries are considering measures to give their security services powers to intercept e-mails and monitor internet traffic. Increasing pressure is being placed on national regulatory bodies to make sure that operators enable surveil-lance and retain call and traffic information. As more voice traffic moves to IP-based networks, the same pressures will be brought to bear on VoIP providers.

For example, in August 2005, the United States FCC ruled that providers of certain broadband and inter con nected VoIP services must be prepared within 18 months to accom-modate law enforcement wiretaps. Under the Communica-tions Assistance for Law Enforcement Act, or CALEA, the FCC has determined that VoIP services can essentially replace the tele com mu ni cation services currently subject to wiretap rules, including circuit-switched voice service and dial-up internet access. As a result, any VoIP provider inter con nected with the PSTN must be wiretap-ready by early 2007.21 CALEA requires the Commission to preserve the ability of law enforce-ment agencies to conduct court-ordered wiretaps in the face of technological change.

Wiretapping, however, raises serious concerns regard-ing individual privacy rights. In terms of lawful interception, regulators might play a useful role in helping to determine the balance between the rights of the individual citizen and the requirements of government to monitor traffic. At a practical level, regulators can also help find a balance between the obli-gations and requirements of law enforcement agencies and the needs of service providers. Below are some of the issues that are raised by lawful interception obligations:• Costs: The cost of complying with wiretap obligations can

be significant. In some countries, the government shares the costs of lawful interception with smaller operators or service providers, but where these arrangements are absent, regulators need to be sensitive to the fact that for smaller ISPs or VoIP service providers, the cost of purchasing the

necessary equipment to be able to provide access to law enforcement agencies can be prohibitive.

• Area of Responsibility: Another potential problem area is delineating re spon sibilities for im ple men ta tion and com-pliance between the regulators and law enforcement agencies. This can lead to difficulties in establishing tech-nical specifications, determining service provider re spon-sibilities and applying remedies for non-compliance.

• Standards: Lawful interception, especially of cross-border services, is highly dependent on stand ardization bodies such as the European Telecommunications Standards Institute. Unfortunately, although standards for lawful interception in traditional circuit-switched networks are well defined, there is still a long way to go before inter-ception standards for VoIP are stand ardized.

6.6 Conclusion

The im ple men ta tion of regulatory approaches to VoIP will remain highly uneven for many years to come. In fact, they may never be uniform. But there is at least a consensus that regulators should address IP-based networks and services. The countries that have begun to do so are generating precedents and providing guidance for others to follow.

It appears that at least minimal additional regulation of VoIP may be required to ensure quality, security, network integrity, inter con nection, access to emergency services and further competition in global telecoms markets. VoIP services offer a truly exciting technological development that may yet unlock affordable communication solutions for much of the developing world. Regulators can act to ensure that they assist in this common goal.

VoIP is a particularly important opportunity for develop-ing countries to provide voice and other services more cheaply than with traditional PSTN networks. The increased availabil-ity of cheaper services will broaden access to a larger number of citizens, providing another avenue for closing the Digital Divide.

CHAPTER 6 109


1 Clayton Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (New York: Harper Business, 1997). Christensen coined the term “disruptive technologies,” which are said to satisfy existing customer needs at a drastically lower cost and are simpler and easier to use than previous ones. Disruptive technologies cannot at the point they are introduced into the market, compete against the traditional products such that they can acquire a large market share.

2 See http://www.telegeography.com/press/releases/2005-05-31.php3 For example, from May 2005, Senegal’s incumbent operator, Sonatel, cut its international call rates to USD 0.19 a minute at weekends and USD 0.31 cents during

office hours. Sonatel was one of early adopters of VoIP for international calling.4 In South Africa, as part of its tariff rebalancing, the monopoly operator Telkom reduced international calls overall by approximately 36 per cent over 2004-2005.

Telkom’s licence required it to ensure tariff rebalancing was completed by the end of its five-year monopoly. 5 See http://www.20min.ch/tools/suchen/story/27383201.6 It is also noteworthy that the establishment of the World Trade Organization (WTO) and the GATS Framework Agreement has facilitated a much wider and more

complex exchange process, and facilitated a regime change from one typified by an exchange of traffic, to one characterized by (multilateral) trade in services, and a consequent shift away from individual, bilateral country negotiations for market access and network inter con nection.

7 The Pakistan carriers use VoIP as licensed long-distance or local loop carriers.8 Chinatechnews, “MII: No Plans Soon To Lift Ban Over VoIP”, July 22, 2005, online at http://www.chinatechnews.com/index.php?action=show&type=news&id=2813;

Chinatechnews, “China Netcom Prepares To Crackdown On Illegal Phone Kiosks”, June 20, 2005, online at http://www.chinatechnews.com/index.php?action=show&type=news&id=2716

9 IP-Enabled Services, WC Docket No. 04-36, Notice of Proposed Rulemaking, 19 FCC Rcd 4863 (2004)10 “Regulatory framework for voice communication services using Internet Protocol”, Telecom Decision CRTC 2005-28, 12 May 2005.11 Local VoIP services are defined as those which use telephone numbers that conform to the North American Numbering Plan and provide universal access to and/or

from the Public Switched Telephone Network (PSTN), along with the ability to make or receive calls that originate and terminate within an exchange or local calling area.

12 “Local Competition”, Telecom Decision CRTC 97-8, 1 May 1997.13 Memorandum Circular, No. 05-08-2005, Subject: VOICE OVER INTERNET PROTOCOL (VOIP), National Telecommunications Commission of the Philip-

pines (last visited at http://www.ntc.gov.ph/whatsnew-frame.html)14 REGULATION OF INTERNET PROTOCOL (IP) TELEPHONY, Statement of the Telecommunications Authority, 20 June 2005, OFTA http://www.ofta.gov.

hk/en/tas/ftn/tas20050620.pdf15 China’s government is considering banning the use of unregulated VoIP services. Report in Fierce VoIP, 13 September 2005.16 Intrado has introduced a new service that will enable VoIP providers to offer 911 services in compliance with the recent FCC mandate. The company claims that its

V911 Mobility Service will let wireless VoIP providers nationwide position their services as primary line replacements. The solution is designed to accommodate each jurisdiction’s 911 regulations while supporting static, mobile and out-of-area phone numbers. Intrado supplies 911 integration by transmitting routing instructions for the local 911 service through the wireless VoIP provider’s modem. The service redirects VoIP 911 calls over the existing wireline service and offers live call-center support 24 hours a day. See http://lists.fiercemarkets.com/c.html?rtr=on&s=69l,f8zf,lcs,bhks,bxd0,jti9,1419

17 “Decision on 9-1-1 Emergency Services for VoIP Service Providers”, Telecom Decision CRTC 2005-21.18 Federal Communications Commission, Federal-State Joint Board on Universal Service Seeks Comment on Proposals to modify the Commission’s Rules relating

to High Cost Universal Service Support”, FCC 05J-1, CC Docket No. 96-45, 17 August 2005.19 Telegeography, “Telof to launch VoIP in unserved areas”, May 16, 2005. Available at http://www.telegeography.com/cu/article.php?article_id=720320 The list is provided in the European Commission’s recommendations. 21 Federal Communications Commission, First Report and Order and Further Notice of Proposed Rulemaking, FCC 05-153, ET Docket No. 04-295 RM-10865, 23

September 2005. Available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/FCC-05-153A1.pdf

111


CHAPTER 7

Author: John G. Palfrey, Jr., Executive Director, Berkman Center for Internet & Societyand Clinical Professor of Law, Harvard Law School

7 STEMMING THE INTERNATIONAL TIDE OF SPAM

The anti-spam laws enacted around the world so far have been largely unsuccessful in stopping spam.1 In almost every instance, anti-spam statutes have been directed at sanctioning spammers for their bad acts. An increasing number of coun-tries and other jurisdictions have created such laws or applied to spam their existing, generally applicable laws concerning data protection, consumer protection, and protection against fraud. Yet, in many cases, these laws have missed their target entirely, with no perceptible impact on actual spammers. Even worse, the laws have often had negative side effects, in the form of transaction costs, ad min is trative costs, and a chilling effect on legitimate senders of e-mail.

No matter what kind of law is enacted or applied, anti-spam measures require well-conceived, targeted, and coor-dinated enforcement mechanisms in order to be effective. Without a doubt, anti-spam investigations are invariably com-plicated and expensive, presenting challenges for any coun-try seeking to enforce anti-spam laws. Even the U.S. Federal Trade Commission, with its substantial resources, has brought only approximately 70 cases against spammers. For developing countries that have limited human and financial resources for such work, anti-spam laws can be rendered nearly meaningless because of the enforcement challenge.

Cross-border cooperation and enforcement is not only desirable, but also essential to spam fighting. But the variety of anti-spam laws and underlying legal systems on the books of various countries makes collaboration extremely difficult. The challenge of fighting spam through law – to be sure, only one of the potential modes of regulation – calls for new thinking and increased emphasis on international harmonization and collaboration. The only effective means of combating spam is likely to be a combination of approaches. As noted in the Chairman’s report of the ITU 2004 Global Symposium for Regulators (GSR),2 a multi-pronged approach to dealing with spam is an appropriate measure.

This chapter primarily takes up the question of what – beyond coordinating with technologists and other countries’ enforcement teams and educating consumers – legislators and regulators might consider by way of legal mechanisms. First, the chapter takes up the elements that might be included in an anti-spam law. Second, it explores one alternative legal mechanism which might be built into an anti-spam strategy,

the establishment of enforceable codes of conduct for Internet Service Providers (ISPs). Third, the chapter also examines a variant of the legal approach where ISPs are formally encour-aged by regulators to develop their own code of conduct. ISPs should be encouraged to establish and enforce narrowly-drawn codes of conduct that prohibit their users from using that ISP as a source for spamming and related bad acts, such as spoofing and phishing, and not to enter into peering arrangements with ISPs that do not uphold similar codes of conduct. Rather than continue to rely upon chasing individual spammers, regulators in the most resource-constrained countries in particular would be more likely to succeed by working with and through the ISPs that are closer to the source of the problem, to their cus-tomers, and to the technology in question. The regulator’s job would be to ensure that ISPs within their jurisdiction adopt adequate codes of conduct as a condition of their operating li-cense and then to enforce adherence to those codes of conduct. The regulator can also play a role in sharing best practices among ISPs and making consumers aware of the good works of the best ISPs. While effectively just shifting the burden of some of the anti-spam enforcement to ISPs is not without clear drawbacks, and cannot alone succeed in stemming the tide of spam, such a policy has a far higher likelihood of success in the developing countries context than the anti-spam enforcement tactics employed to date.

7.1 The Spam Problem

The problem of spam is well established. The extent of the problem is plain to anyone who relies upon electronic mail (email) for communications. Email and related forms of mes-saging such as “blogs” (short for “Web logs”) and short mes-saging service (SMS), have become an important and popular means of communication in cultures around the world. These services are cheap, they have global reach, and they are playing a key role in the development of e-commerce. The proof of their value is found in their extraordinary global adoption rate, whether in the form of an e-mail client (such as Microsoft’s Outlook, Eudora, Thunderbird, or others) or hosted services (such as Microsoft’s Hotmail, Outblaze, Yahoo! Mail, Google’s Gmail, Wanadoo or Noos in France, among others).

CHAPTER 7


112

But the openness that has made e-mail and its close cous-ins such tremendously easy ways to connect is also emerging as their greatest vulnerability. A combination of economics, tech-nologies, and online behaviour norms has made the incremen-tal cost of sending a spam message nearly zero, while promising senders a profitable potential return.

At first glance, the economics seem baffling. How can it possibly be worthwhile to send out grammatically challenged messages about low-cost pharmaceuticals or pirated software

– offers that the vast majority of recipients ignore and quickly transfer to their “junk mail” folders? Part of the answer is the tiny marginal cost of sending spam messages. Because they cost nearly nothing to send, the response rate does not need to be very high. And it turns out that enough people do respond to make the endeavour worthwhile to the spammer. Astonish-ingly, the Business Software Alliance (BSA) has found that 22 per cent of British consumers they surveyed purchased soft-ware through spam.3 Rates for the other five countries BSA surveyed were similarly high. The bottom line is that spam per-sists because it is profitable. Unless enough consumers become educated to avoid or reject spam, the best way to reduce spam may be to raise the risks and costs to the spammer.

Right now, the costs seem to be landing on consumers. Every major, credible report on this topic suggests that more than half of the e-mails sent today are spam, and some suggest that spam comprises between 70 and 90 per cent of all e-mails sent.4 The costs of this scourge are borne not by the spammers, but by those who run networks, employers and the individu-als who receive the messages. Spammers – and those who use spam to perpetrate related frauds – take advantage of the open design of IP networks to render e-mail costly and nearly unus-able for some businesses and consumers.

7.1.1 Legislative ResponsesThe “extremely rapid growth” of spam5 has led to the

enactment of more than 75 specific laws,6 such as the well-regarded Australian law, the United States’ CAN-SPAM Act of 2003 and comparable legislation in several dozen countries around the world.7 These laws have, to date, been unable to stop spam. Accounts vary somewhat in terms of rates of growth, but there is no persuasive evidence that the growth of spam has abated in the wake of anti-spam legislation.8 In fact, most indicators point in the other direction.9

Spam is best viewed not as a nuisance, but in the context of cybersecurity. Spam is bad enough as a drain on productivity and a daily annoyance. But few people consider that spam is enormously costly to ISPs and others who maintain the net-work at various levels. Meanwhile, its negative impact is grow-ing by virtue of the bad things it brings with it. Spam is the preferred delivery mechanism for a range of Internet security threats: viruses, “phishing” and “pharming,”10 scams with end-less permutations, and advance fee frauds, to name a few.11 Spam is also undercutting the efforts of developing countries to persuade new users to rely on digital communications.

Bill Gates, who is arguably the world’s most powerful technologist, promised to lead the charge against spam and to end it within two years of the January 2004 World Economic

Forum meeting in Davos, Switzerland.12 He is not alone in having fallen short in this goal. In fact, most major, well-inten-tioned ISPs and e-mail service providers, along with many technology start-ups, have devoted many millions of dollars to spam-fighting measures. Standards bodies have sought to improve protocols to snag more spam. User education cam-paigns have been launched. And governments around the world have come together to enforce their spam laws and to cooperate more effectively with one another. The problem continues despite these many efforts, suggesting that new solu-tions must emerge and that existing efforts must be better pur-sued and coordinated.

Some of the most effective recent efforts have been those lawsuits undertaken by ISPs under a private right of action in spam legislation. In the United States, the CAN-SPAM Act of 2003 enables ISPs to sue spammers directly. AOL, Microsoft, and Earthlink – very large-scale providers of electronic mes-saging services – have each brought actions under this statute, as well as under state-level computer crime and common law statutes. This has resulted in multi-million-dollar judgments and settlements against “spam king-pins” who abuse their net-works.13 Microsoft won a USD 7 million judgment that may well have put an end to one spamming operation that alleg-edly distributed more than 38 billion unsolicited messages per year.14

These lawsuits – although few and far between, and limited to certain jurisdictions – represent a ray of hope that enforcement by ISPs, with help from customers, might get the job done against spam. Indeed, the success of these efforts suggests that ISPs could become the most valuable players in the effort to end spam. The challenge for lawmakers is how to create a fair, effective regulatory regime that takes advantage of ISPs’ ability to help end spam without placing an undue burden on law-abiding companies.15

7.1.2 A Model Law: One of Several Ways To End Spam

7.1.2.1 A Combination of Approaches Is Needed

The persistence of spam problem has led policy-makers, technologists, academics, and many others to come up with a wide range of possible strategies to end it. The least intru-sive approach, most consistent with the end-to-end principle of network design, is to leave the job to end users, through simple technologies such as spam filters on e-mail clients. The improvement of authentication, accreditation, and identity management technologies ought to help make user-level con-trols more effective over time.16 At Davos in 2004, Mr. Gates described Microsoft’s pursuit of solutions to complement these user controls.17 One approach calls for a combination of law, code, markets, and norms.18

Meanwhile, the chairman’s report of the ITU Thematic Workshop on Countering Spam in 2004 contains a range of proposals, suggesting an intersection of many methods of spam-fighting.19 This comprehensive, five-part approach calls for a combination of:

CHAPTER 7 113


• Strong, enforceable legislation;• The continued development of technical measures;• The establishment of meaningful industry partnerships,

especially among ISPs, mobile carriers and direct market-ing associations;

• The education of consumers and industry players about anti-spam measures and Internet security practices; and,

• International cooperation among government, industry, consumer, business and anti-spam groups, for a global and coordinated approach to the problem.

In fact, virtually every major report on spam calls for a combination of approaches to combat the problem, rather than a single, “silver-bullet” solution. This chapter does not take up in detail each of these anti-spam tools, but rather focuses on legal strategies, emphasizing those that are relevant to develop-ing countries.

Anti-spam laws are perceived today to be a necessary tool for all countries, if for no other reason than that they help facilitate international cooperation in combating spam. Even the most ardent supporters of user controls and market solu-tions agree that governments have a role to play in tracking down and punishing the worst offenders, such as those who use spam to commit fraud. The existence of inter op erable anti-spam laws creates a common baseline for international enforcement. A developing country may not be able, by itself, to enforce its anti-spam law, but that law can provide the basis for regional and multinational enforcement actions.

A country with experience enforcing anti-spam legislation may wish to provide human resources to conduct an anti-spam investigation and enforcement action that leads to another country. In the absence of anti-spam legislation, however, such international cooperation is not possible on a systemic basis. Anti-spam laws are increasingly viewed as one of several necessary tools for most countries.

7.1.2.2 The Effect on Developing Countries

Spam is arguably a bigger problem in developing countries than in wealthier countries, where anti-spam mechanisms are more robust. Many developing countries do not yet have anti-spam laws,20 and those that do often do not have resources to enforce them.21 Meanwhile, the effects of spam are often rela-tively more costly in developing countries. ISPs are frequently deluged by spikes in spam, which lead to network slowdowns and breakdowns.22

Moreover, many people in developing countries send emails from shared Internet connections and equipment, such as at cybercafés or other public access centres. These services ordinarily rely on hosted email services with limits on inbox sizes. Accessing email becomes too expensive if per-minute charges paid to cybercafé owners are consumed by cleaning spam from their inboxes. Even worse, legitimate emails are bounced because the limited space of their inboxes is con-sumed by spam.

Officials from developing countries often point to the fact that most spam still comes from the United States and other wealthy countries, which have done little to help developing

countries cope with the problem. In addition, they note that the resources of regional bodies such as the OECD are not consistently available to developing countries. This leaves them at a comparative disadvantage in fighting spam.

The answer for developing countries is not simply to copy anti-spam laws enacted in developed countries. That approach is unlikely to be effective. Anti-spam laws aimed at sanction-ing spammers may be of little use in developing countries if the spammers are outside their jurisdiction. The challenge is to tailor legislation to patterns of usage in developing countries and to consider all avenues to combat spam, such as imple-menting enforceable codes of conduct for ISPs.

7.1.3 An Alternative Mechanism: Enforceable Codes of Conduct

In addition to enacting anti-spam legislation, developing countries could require ISPs to establish an industry code of conduct on spam. The enabling legislation for such a code could stipulate that the nation’s regulatory agency would enforce the code against any ISP that ma teri ally violated it.23 Such a proposal cuts jarringly across the grain of most inter-net regulation to date. As essential players in developing ICT-powered economies, ISPs have generally been left alone by legislatures, ad min is trative agencies, and judges. They may be licensed and overseen by regulators in some contexts, but ISPs have largely been immune from prosecution for bad acts com-mitted by people through their services.

7.1.3.1 Elevating the Role of the ISP

Ideally, it is not an ISP’s job to be a gatekeeper. The ISP should pass all packets from sender to receiver, with end users deciding what to send and what to receive. Any departure from this model should be undertaken only when serious circum-stances warrant it. In addition, regulation should be handled with a light touch, and any burdens placed on ISPs should not be starting points for more intrusive regulation.

It is essential to ac knowl edge how the internet has changed since its inception. We use the network far differently than any of its early architects could possibly have imagined. The

“community” of users is now more far-flung than it ever was, and they no longer expect to know one another, as the earliest academics and military users did. The internet’s architecture is a victim of its own success. The conventional wisdom that no intelligence should be built into the heart of the network

– the so-called end-to-end principle – is still held dear by many technologists, but it is no longer fully reflected in reality. A large number of control points have been built into the net-work – often to deal with massive problems like spam.24

ISPs still enjoy broad immunities in many jurisdictions from claims based on what others do on their networks. For example, they rarely face copyright violation or defamation claims. But they are increasingly called upon to play a role in protecting and policing the internet. There are substantial risks associated with placing such jobs in the hands of ISPs – par-ticularly to civil liberties – so any legislation that mandates a greater supervisory role must be carefully drafted so as to miti-gate these risks.

CHAPTER 7


114

7.1.3.2 Establishing an Industry-Led Approach

Countries should work to establish an industry-led regu-latory approach that provides a mechanism for regulators to step in against the worst spam abusers. This proposal is not meant to presage a wholesale shift in the role of ISPs. Nor is it meant to indicate a rejection of the end-to-end principle as a preferred design matter. ISPs already bear the brunt of the costs of spam. The role of the law and the regulator should not be to burden ISPs further, especially given the constraints they already face.25

Rather, the goal is to reduce spam in a way that protects responsible ISPs. As the internet has developed into a complex network of networks, ISPs are positioned, for good or ill, as key gatekeepers. ISPs that implement responsible, effective anti-spam measures, while preserving the civil liberties of their users in a manner that is consistent with local law, should be rewarded for their good behaviour. One means of rewarding those responsible ISPs is for regulators to hold their irresponsi-ble competitors accountable. This would create a level playing field for responsible ISPs.

ISPs are no strangers to fighting spam. ISPs around the world have taken an active role in attacking spam at the source, before it clogs their customers’ inboxes. Anti-spam measures implemented by ISPs cover a wide range. Many ISPs partici-pate in industry-wide working groups, such as the Messaging Anti-Abuse Working Group.26 Many also work with standard-setting organizations developing technical solutions.27

ISPs’ initiatives are often geared toward improving security and decreasing the vulnerability of users and of their networks. When they succeed, it can often be a strong selling point for them. For example, Google’s Gmail, a free Web-based e-mail service, removes hyperlinks from messages that the service believes to be phishing attempts.28 The large U.S.-based ISP Earthlink requires all e-mail messages to be routed through its mail servers, in order to reduce the impact of “zombie” net-works. Earthlink also mandates that users’ e-mail programs submit passwords to transmit messages.29

While these methods can reduce the burden of spam, their effect is minimal if consumers do not also take steps at the “client” level of the network. End users may not update their own virus software automatically or regularly. Or, they may download programs that contain “malware” and “spyware” that compromise their computers, posing a risk not only to themselves but to other users worldwide, since their PCs may be hijacked to relay spam to other unsuspecting consumers.

Governments and ISPs both have incentives to end spam.30 ISPs bear a large amount of the cost of spam and get nothing in return – unless they are charging a premium to spammers in exchange for sending spam out on their behalf. ISPs also are relatively close to the problem. After all, spammers need ISPs to get access to the internet to dump their messages. While spammers are increasingly sophisticated in evading tracking, a concerted effort among cooperating ISPs (and possibly law enforcement officials and end users) can find the worst offend-ers. The routing of spam can be traced and mapped at a net-work level.31 While ISPs are often short on cash flow, many do

have the financial and human resources to play a key role in the anti-spam fight.

National laws can mandate the development of codes of conduct by and for ISPs. Adherence to the code could be a licence condition, or it could be implemented through a rule-making proceeding, via a common set of regulations that applies to ISPs whether licensed or authorized, much as operators are required to provide interconnection, the rules for which are spelled out in interconnection regulations with industry partici-pation. The law would give ISPs the first opportunity to craft the code, outlining acceptable behaviour for ISPs and their customers. Preferably, the code would prohibit spam, phishing, spoofing on the ISPs network, and similar practices. It could also suggest or endorse the best use of spam filters and other technological tools for customers and ISPs to fight spam. The regulatory agency would approve and, in many cases, enforce the code.

Under such codes, ISPs would commit themselves to denying service of any kind to spammers, phishers, spoofers and other bad actors who violate these policies. Such codes of conduct would be led by industry and made functionally con-sistent among all players across the industry, but as part of a process that is grounded in law and provides a role for regula-tors. The regulator would be empowered to approve the code and to enforce the code if the ISP deviates from its terms in material fashion.

Regulators are better able to do their job under this sce-nario, as compared to the straight enforcement role against spammers, since the regulators would primarily interact with ISPs. The ISPs are largely running legitimate businesses, are incentivized to help solve the problem (so long as they are not cheating), and are easy to find relative to the spammers, who are often not in the same country and are constantly hiding behind technological smoke and mirrors. The ISPs, in turn, would be responsible to keep tabs on those customers who are engaged in illegal activity and to spurn offers for premium pay-ments to provide spammers with an onramp to the internet.

This mechanism would empower the regulator to apply a default code of conduct where ISPs fail to develop one or until an acceptable policy is set forth by the ISP. Such a mechanism would also include the regulator’s certification of the code which ISPs could use in their advertisements, to ensure cus-tomers that the ISP is taking all available steps to protect its customers, and the network at large, from spam. The system would also involve a reporting mechanism so that victims of spam, phishing, spoofing and the like can report such activity either to the ISP or the regulator for follow-up investigation and action.

An enforceable code of conduct is not without drawbacks. The code must be narrowly tailored to curb spam and related bad acts. It should not be used as a back-door measure to over-burden ISPs, such as by:• Imposing anti-spam obligations where no technical solu-

tion yet exists (as with many anti-spoofing requirements);• By using anti-spam measures as a means to limit legiti-

mate political discourse or other protected speech; or• By infringing on the privacy interests of citizens.

CHAPTER 7 115


It is essential that the industry develops and approves the code of conduct – or, at a minimum, collaborates with regula-tors in this task. Industry “buy in” is important, because the code will require frequent updating to reflect new develop-ments in spamming practices and anti-spam technologies.

7.1.3.3 Voluntary Codes of Conduct

As an alternative to a mandated code, enforced by regula-tors, governments might encourage ISPs to develop their own, industry-enforced codes of conduct. In fact, many ISPs are taking this step without any encouragement. Terms and norms are often built into “acceptable use” policies for customers and peering arrangements.32 Under this voluntary model, regula-tors could advise the industry in developing the codes. It could then help consumers find the ISPs that have developed or signed on to those codes. If a vibrant ISP market emerges, con-sumers could then choose ISPs that have proactively tried to fight and reduce spam.

Finally, regardless of whether ISPs are compelled to estab-lish codes of conduct or do so voluntarily, regulators have an important role to play in educating and raising awareness. Consumers, businesses, ISPs and cybercafé operators need information on technical solutions such as spam filters, as well as warnings about viruses and fraudulent activities that have been detected. There is much to be gained from government-industry collaboration in protecting consumers from spam.

7.2 An Outline of a Model Law

7.2.1 The Context for a Model Anti-Spam Law

Representatives of many countries, particularly in devel-oping regions of the world, have sought a model law for combating spam. The topic was discussed intensively at two international gatherings hosted by ITU. The first, held in the summer of 2004, was devoted to the issue of spam, while the other, a year later, focused on cybersecurity. This chapter draws upon the many resources developed to date, in an attempt to create a model anti-spam law. There are multiple potential benefits of such a document:• Clear guidelines – Email senders that want to comply

with legal requirements could more easily learn what rules apply to them and could then follow them more consist-ently.

• Jurisdictional Consistency – Enacting a similar, model law in many jurisdictions would free ISPs and email send-ers from having to attempt the near-impossible task of tai-loring messages for recipients in different jurisdictions.

• Easy adoption – Legal systems that do not yet have laws governing spam would have a ready-made model to implement, reducing the burdens of drafting, im ple men ta-tion, and coordination.

• Enhanced enforcement – Regulators could enforce laws more effectively and easily since their systems would share harmonized definitions of offences, burdens of proof, and

exceptions. Greater harmonization would make broad-based cooperative arrangements more likely to arise.

• Stronger norms – Broad international consensus on the meaning of spam, and what constitutes unlawful abuse of electronic communication, would strengthen norms that deplore such conduct.

• Fewer havens for spammers – As more governments adopted the model law, spammers would have fewer friendly locations to establish operations. This would increase their costs and reduce the financial incentives to engage in massive spamming.

• Increased sharing of best practices – Since legal sys-tems would share harmonized provisions, regulators and enforcers could more easily collaborate to develop and share best practices for implementing spam laws.33

Even well-crafted anti-spam laws, implemented in every jurisdiction, will never get the job done alone. But anti-spam legislation can be a useful element of a coordinated anti-spam strategy. A good anti-spam law should distinguish between good actors and bad actors and mete out punishment accord-ingly. Moreover, if spammers were liable for each spam mes-sage they send, the level of fines would increase exponentially, according to the scale of the spam operation.34 Enforcement is the key – and the most difficult element – particularly in devel-oping countries.35

The development of a model anti-spam law should be collaborative and inclusive. As with any model law (or any offi-cial document with the force of law) an anti-spam law must be flexible enough to dovetail with existing laws, including anti-fraud, consumer-protection, tele com mu ni cation and inter-net-specific laws and regulations. One relevant example is the process that the United Nations Commission on International Trade Law (UNCITRAL) undertook in establishing its Model Law on Electronic Commerce (1996).36 UNCITRAL’s e-com-merce model law does not specifically address spam, which did not exist as in 1996 as the huge issue that it is today. Anyone designing an anti-spam model law should also consider the broad range of laws on the books today in many countries, containing variations that are worth considering but that are too numerous to be included in this chapter.37

Most of the existing anti-spam laws are directed at con-trolling spammers’ behaviour. This seems appropriate, since spammers directly cause the problem. But the current slate of laws has failed even to curb the growth of spam, much less to reduce the problem.38 Why have they failed? Some observers argue that the countries generating the largest proportion of the world’s spam have done too little at home to stop the prob-lem.39 Those making this argument especially criticize reliance upon “opt-out” rules that allow spam unless consumers spe-cifically ask not to receive it. Even then, opt-out rules are not enforced aggressively enough.

It is not enough to blame the greatest spam-producing nations, though. No country in the world – including those lauded as the most effective in combating spam – has made significant inroads using classic enforcement mechanisms. Of course, it would help if governments updated their laws in

CHAPTER 7


116

light of their apparent inadequacy, but that takes time. Other observers suggest that anti-spam laws should be focused not on the spammers themselves, but rather on the (often dodgy) companies for whom the spam is sent. 40

The primary issue is that little emphasis is placed on inves-tigation, enforcement powers, or resources. It is not that hard to build and win a case. Most spammers and their clients even-tually can be found, with enough hard work and cooperation. The problem is that each investigation is so time-intensive and costly that police and prosecutors often decide that the costs outweigh the benefits. One of the core tenets of the model law described below is that it emphasizes creating a framework for national enforcement, international coordination, and distrib-uted monitoring through the ISP code of conduct.41

7.2.2 Elements of a Model Spam Law

The draft model law presented in this section as an anno-tated outline roughly follows the structure of the Australian anti-spam law, which is widely regarded as one of the most well-conceived statutes of its kind in the world.42 This section describes the key elements of a model law, offering suggestions for options at each stage of the drafting process.

One threshold issue is whether the law will be an “opt-in” or an “opt-out” statute. An opt-in statute makes it illegal to send spam unless a recipient has affirmatively agreed to receive it. Often, only tacit acceptance is required, such as the existence of an ongoing business relationship of some kind. An opt-out statute, on the other hand, permits spam unless the recipient has specifically informed the spammer that he or she does not want to receive it.

The decision to choose an opt-in or opt-out approach will reverberate throughout the law from that point onward. For instance, in an opt-out system, the provision to establish an

“unsubscribe” function will be more essential and take on a dif-ferent character than in an opt-in law, which presumes that the receiver already gave a green light before receiving any spam messages.

One deficiency of many spam laws is a lack of clear defini-tions. The draft model law, below, seeks to head off variations among definitions adopted in different jurisdictions, because these variations could undermine international cooperation on enforcement.

Draft Model Law

Section 1: Introduction and Definitions

The law should clarify that it establishes a scheme for regulating commercial e-mail and other types of commercial electronic messages.

Annotation: The introduction section of the law ought to set forth the definitions, which take on special significance in the anti-spam context. On the one hand, the terms must be broad enough to encompass emerging types of spam as they develop. On the other hand, the provisions must be precise enough to be clearly understood.

In addition, since anti-spam statutes can affect civil liberties such as free speech and personal privacy, definitions may play a pivotal role in determining whether the statute is permissible under a country’s constitutional framework or sufficiently protective of citizens’ rights.

The following are some of the key terms to be included in the definitions section of the model law, (although this is not a complete list):

• Address-harvesting software. The law should define what types of computer applications used to harvest e-mail addresses are banned under the statute.

Annotation: An important question for any anti-spam law is whether or not to include a prohibition on the use of, or trafficking in, technologies that support spamming, such as address-harvesting software. If such as ban is included in the law, the term must be carefully defined so as to avoid banning useful technologies of gen-eral applicability that may be used for address-harvesting. Another approach is not to ban any technology, but rather to bar its use for gathering e-mail addresses for spamming.

• Authority, or Regulator. The law should specify the entity or individual that has jurisdiction over the anti-spam law. Countries vary as to the precise placement of this authority, which might be vested in the tele com mu-ni cation regulator, the consumer protection authority, the trade regulator, or another authority.

Annotation: If multiple regulators are tasked with enforc-ing anti-spam rules, a precise division of re spon sibilities should be established, either in the definitions section or, more likely, in the enforcement-related provisions.

• Authorization. The law should clarify what it means for an individual to authorize sending a message that could be defined as spam.

Annotation: This definition may take on greater or lesser sig-nificance depending on whether the law is designed as opt-in rather than opt-out. Depending upon the nature of the law adopted and the use and definition of the term “consent,” this definition might not be necessary.

• Commercial. The law must specify with precision what constitutes a message sent for commercial purposes. Commercial messages sent to recipients with whom they do not have a previous commercial relationship are likely to serve as the core, prohibited type of content..

Annotation: One key issue facing development of a useful model law is variation in the treatment of speech rights in dif-ferent countries. In Australia and the United States, for instance, legislators and regulators have stayed clear of regulating unsolicited political messages in light of constitutional protections for political speech. Most anti-spam laws focus not on the content of the mes-sage, but rather on the intent of the sender. Spam legislation varies as to whether or not it applies only to commercial messages, but it is important to define what constitutes “commercial” in any event.

• Consent (or, Affirmative Consent). The law should clearly state what the recipient must do to signal willing-ness to receive e-mail from a particular sender. The law could use the term affirmative consent, which means that (A)

CHAPTER 7 117


the recipient expressly consented to receive the message, either in response to a clear and conspicuous request for such consent or at the recipient’s own initiative; and (B) if the message is from a party other than the party to which the recipient communicated such consent, the recipient was given clear and conspicuous notice at the time the consent was communicated that the recipient’s electronic mail address could be transferred to another party for the purpose of initiating commercial electronic mail mes-sages.

Annotation: This definition should be coordinated with the definition of the term “authorization,” as needed.

• Electronic message. The law should specify what con-stitutes an electronic message. In the Australian statute, an electronic message is a message sent using (a) an Inter-net carriage service or (b) any other listed carriage serv-ice. Also, an email message is sent to an electronic address in connection with (1) an e-mail account; (2) an instant messaging account; (3) a telephone account; or (4) a simi-lar account.

Annotation: An important area to consider is what applica-tions the anti-spam statute covers. The best anti-spam laws will be general enough to cover ICT-based unsolicited messaging in formats that have yet to be devised, as well as those that exist today. Short Messaging Service (SMS) text messages on cellular phones, spam over the instant messaging protocol (“spim”), web blogs (especially in the comments fields), spam over Internet telephony (SPIT), voice messaging over Internet telephony and Really Simple Syn-dication (RSS) are important current variants of traditional e-mail spam that drafters may wish to keep in mind.

• Evidential (or evidentiary) burden (or, burden of proof). The law should define carefully which party bears the burden of producing evidence.

Annotation: One of the key problems that enforcement authorities face is a high burden of proof placed upon the prosecu-tion in instances where they must show conclusively that a user did not opt-in to receiving spam. Virtually no individual can prove the negative – that they never entered into a commercial relationship, or never once hit “OK” in a click-through contract. To place the burden on the regulator to prove this negative is to hamstring her or him in the enforcement process.

• Internet service provider (or Internet carriage service; Internet content provider; E-mail service provider; Telecommunications service; or the like depending upon jurisdiction). The law should define what type of service the statute covers. The essential part of the definition is that the covered party provides a con-nection between an end-user and the internet, for a fee.

Annotation: In many jurisdictions, a wide range of definitions for ISPs are established by various internet-related laws, so special care should be taken to harmonize definitions across statutes, for clarity’s sake. U.S. law, for instance, has more than 40 potential definitions for terms that resemble “Internet service provider.” 43 The elimination of ambiguity is particularly important for this model law, which contemplates setting an affirmative requirement for ISPs to develop an enforceable code of conduct.

• Send. The law should clarify that the definition of “send” includes attempts to send.

Section 2: It is unlawful to send unsolicited commer-cial electronic messages

Annotation: The scope of what type of message is unlawful to send, combined with the definition of the terms of what is banned, is a crucial element of any spam law. Countries vary widely in terms of whether messages beyond “unsolicited commercial e-mail” are included under the law. For instance, non-commercial bulk e-mail is included in the definition of “spam” in some anti-spam legislation and not in others. This is also the juncture at which each country must decide whether to join the opt-in or opt-out camp. Virtually all anti-spam laws focus upon the act of sending (or attempting to send) as the core, operative offence. An additional prohibition for this section might be to hone in on the act of paying someone to send unsolicited commercial electronic messages on one’s behalf. Some states also bar the sending of unsolicited charitable and issue-oriented (political) messages, but that step is dangerous and not advocated here, given the importance of political speech to well-functioning government systems.

Section 3: Commercial electronic messages must include accurate sender information

Commercial electronic messages must include infor-mation about the individual (or organization) who (or that) authorized sending the message.

Annotation: The law might also require that commercial email be identified as an advertisement, by requiring that “ADV” or the like be included in the header. The law could also require commercial email to include the sender’s valid postal address. Some activists have also called for the requirement that senders label sexu-ally explicit messages in the subject line. The labeling requirement is hotly contested by e-mail marketers, who fear that ISPs or indi-viduals will filter out all such messages, even if they are legitimate commercial offers.

Section 4: It is unlawful to include false information in any commercial electronic messages

Commercial electronic messages must not include false information. That includes an email’s “from,” “to,” and rout-ing information, which should include the originating domain name and email address. The subject line cannot mislead the recipient about the contents or subject matter of the message.

Annotation: Most experts contend that an anti-spam law ought to contain such a ban on inclusion of false information as a supplement to other provisions, such as the outright bar against sending an unsolicited message. Without the general ban on unsolic-ited emails, this accuracy requirement can be criticized as effectively permitting spam that is unwanted but accurate. Much of the criti-cism leveled against the U.S. CAN-SPAM Act of 2003 has fol-lowed this argument.

CHAPTER 7


118

Section 5: It is unlawful to send a commercial elec-tronic message without a simple means for recipients to indicate that the recipients do not wish to receive any further commercial electronic messages from the sender

Commercial electronic messages must contain a func-tional “unsubscribe” or opt-out facility. If a recipient exercises the right to request no further emails, the sender must be bound to honour that request. In an opt-in regime, an unsub-scribe provision would basically ensure that any recipient who had previously opted in could reverse that decision and opt out at any time.

Annotation: In the United States, a sender must provide a return email address or another internet-based response mechanism that allows a recipient to ask the sender not to send future email messages to that email address. The sender must honour that request. Any opt-out mechanism a sender includes must be able to process opt-out requests for at least 30 days after commercial email is sent. When a sender receives an opt-out request, the law allows 10 busi-ness days to stop sending email to the requestor’s email address. A sender may not help another entity send email to that address, or have another entity send email on its behalf.

Also, it is illegal for a sender to sell or transfer the email addresses of people who choose not to receive that sender’s email, even in the form of a mailing list, unless a sender transfers the addresses so another entity can comply with the law. These provi-sions, while sensible, are believed to have a very low rate of compli-ance. Most critics also believe that unsubscribe responses by recipients are frequently used to bolster spamming lists, since the spammers then know that the email has reached a real recipient.

Section 6: The use of, and trafficking in, address-har-vesting software and the resulting lists of electronic mail addresses are prohibited.

Address-harvesting software must not be supplied, acquired, trafficked in, or used. An electronic address list pro-duced using address-harvesting software must not be supplied, acquired, trafficked in, or used.

Annotation: There is a wise presumption generally against banning general-purpose technologies. Any provision of this sort ought to exempt the makers of general-purpose technologies (for instance, a spreadsheet or software enabling a user to write a simple program that could scrape information from the Web) that might be used by spammers to harvest e-mail addresses. The law might also include a prohibition against hacking into databases of e-mail addresses, although in many jurisdictions such acts would be cov-ered under statutes related to computer crimes, larceny, trespassing or other offences.

Section 7: Remedies include civil penalties, injunctions, and criminal penalties

The main remedies for violation of the law would be civil penalties and injunctions. Criminal penalties, including impris-onment, are also sometimes sought when false representation,

use of another’s computer to perpetrate a fraud, or similar acts are involved.44

Annotation: The law might also include a provision making it a criminal offence for an ISP knowingly to accept premium pay-ments from spammers who use the ISP’s network to send their spam. Similarly, the law might include a provision that makes the knowing hiring of a spammer to send out unsolicited commercial e-mail a criminal offence.

Section 8: Causes of Action

This section would establish a cause of action for regula-tors against anyone hiring a spammer to distribute bulk email for them (i.e., the owner of a website to whom a spammer is paid to direct traffic, or the party seeking to drive up the value of a certain equity offering, etc.)45. The law might also include additional causes of action, enabling ISPs, enforcement officers in lower jurisdictions, and harmed individuals to initiate cases.

Section 9: International Cooperation

The law should create a mechanism for international infor-mation sharing and, possibly, formal cross-border enforcement support. These rules would simplify the process for exchang-ing information and encourage exploration of memoranda of understanding (MOUs) and similar means of cross-border cooperation.

Annotation: Much of the emphasis of far-sighted regulators in recent years has been on improving cross-border enforcement efforts. The U.S. Federal Trade Commission has been encouraging the U.S. Congress to pass legislation to make such cooperation more likely to succeed. Consider also the work of the International Con-sumer Protection and Enforcement Network, which involves dozens of countries in “sweep days” to rid the internet of scams.46

Section 10: Jurisdiction

An effective anti-spam law might include provisions designed to assist enforcers by resolving jurisdictional ambi-guities.

Annotation: Such a provision could simply clarify what it means for a message to originate or be received within that country and how the regulator will treat such situations. On a more elaborate level, in the United States, the state of Washington’s anti-spam law established a database that includes many of the e-mail addresses in that jurisdiction. The purpose is to protect the state’s residents.47 A list of that nature, held in one place, however, could be an attractive target for hackers. This concern is mitigated by the fact that spam-mers apparently do not have much of a problem coming across large swaths of e-mail addresses through other means.

Section 11: Enforceable Codes of Conduct by ISPs.

An effective anti-spam law might include sections related to the development and enforcement by regulatory authorities

CHAPTER 7 119


of industry-derived and implemented Codes of Conduct for ISPs.48 Such provisions might include:a) An introduction, explaining the intention to establish such

codes of conduct.b) A provision granting regulators authority to require all

ISPs to develop a code of conduct for that jurisdiction.c) A description of the multi-stakeholder process involved in

developing codes of conduct, including what groups will represent the interests of consumers and industry.

d) A provision establishing a registration process for codes of conduct.

e) A provision enabling consumers to access registered codes of conduct.

f) A provision enabling the regulator to draft a code of con-duct in the event that industry cannot agree or otherwise fails to develop one.

g) A provision enabling the regulator to reject a proposed code of conduct in the event that it lacks appropriate com-munity safeguards.

h) A description of the process for the regulator to issue a warning to an ISP for apparent breach of the code prior to taking an enforcement action.

i) A provision granting power to the regulator to enforce the code in the event of breach by the ISP.

Annotation: A similar structure is set forth in Part 6 of Aus-tralia’s Telecommunications Act of 1997 covering industry codes of conduct (see Box 7.1). There are several issues to be considered, many of which are set forth in the section that follows. The law would need to establish a deadline for compliance and provide for periodic updating of the code. One option would be to task an industry association (if one exists in that jurisdiction) to develop the code. The next decision would be whether all ISPs have to comply with a code developed by the association. The enabling provisions for the code might allow ISPs to opt out of a code developed by the association and register a separate code with the regulator, provided the ISP’s self-developed code sufficiently protects the public interest.

7.3 Codes of Conduct The primary goal of a code of conduct is to ensure that

ISPs that provide a route to the internet – the source ISPs – are taking adequate steps to keep spammers off the network. The effect of the code should be to level the playing field for ISPs that are actively seeking to rid the network of spam instead of profiting from sourcing it. While there are many risks in regu-lating ISPs more extensively than they have been in the past, a carefully balanced set of provisions will benefit not just cus-tomers, but all well-intentioned ISPs, too.49

In virtually all instances, industry knows better than most regulators what technical solutions to spam exist and can be implemented.50 Regulators have a role to play in ensuring that industry does all that it can to put technical and policy solu-tions in place and to share best practices.

The use of industry codes of conduct is a promising mechanism that has been under-utilized in the anti-spam fight.

A similar strategy has been used for a variety of other issues, such as inter con nection, number portability, and other techni-cal coordination issues. If combating spam is not in the remit of the tele com mu ni cation regulator, a similar mechanism could be established for consumer protection authorities, data protection authorities or other similar bodies. For the purposes of this chapter, the code of conduct has been included in a model anti-spam law, but such a set of provisions could easily fit within other sections of a country’s legal codes, such as the tele com mu ni cation laws and regulations. The code of conduct does rely, however, upon core elements of an anti-spam stat-ute.

7.3.1 Procedural Steps Toward an Enforceable Code of Conduct

Industry codes of conduct should be developed in a spirit of minimal regulation of the internet and as a measure of private and public sector cooperation to address the growing problem of spam. The process of drafting a code likely would include several key steps:

• The relevant industry member or members are granted the first chance to develop their own code of conduct, based upon the stated goals of the enabling law or regula-tions. The process by which a code is drafted should be set forth in the law or regulations so as to ensure broad and open participation by key stakeholders.

• Where appropriate, the regulator can help by sharing best practices. This can be done, for example, through the use of ITU’s Global Regulators Exchange (G-REX)51 or face to face meetings such as ITU’s annual Global Symposium for Regulators (GSR). Regulators may also be able to tap into international resources such as the OECD’s Spam Toolkit, which is under development. A draft is accessible at http://www.oecd-antispam.org

• The relevant industry members present the draft code to the regulator for its approval.

• A new body, or an existing regulator with relevant exper-tise, takes re spon sibility for the ad min is tration and regis-tration of the code.

• If the industry fails to develop a code, or if the code is not deemed acceptable, the regulator has the power to step in to draft or revise it, ensuring that sufficient anti-spam measures are being taken by ISPs, network operators and other potential spam carriers.

• The industry members are expected to enforce the code against their customers and those with whom they peer. The enforcement is meant to prohibit the worst acts of spamming, not to encourage an ISP to monitor messages any more than they already do. The expectation is that ISPs would only need to take reasonable measures, such as investigating when they receive an unusually large

CHAPTER 7


120

numbers of complaints against a single customer or when the regulator passes along such complaints.

• The regulator or ad min is trator provides a mechanism for handling end users’ complaints against ISPs for failure to live up the code.

• If industry members fail to enforce the code, the regulator is empowered to take action against non-compliant ISPs. Possible sanctions include fines, harsher licensing require-ments, or lawsuits.

Annotation: One issue to consider is which parties would have a right of action to sue a non-responsive ISP. For instance, consumers who have experienced damage by spam or phishing could be given the right to go to court to sue ISPs directly for violating the code of conduct. Also, regulators could require ISPs to include in their customer contracts binding agreements to honour the code. This would allow consumers and companies to sue not only under an anti-spam law, but also pursuant to laws governing breach of contract.

• The code could also create a “certification” or “accredi-tation” system, allowing ISPs to publicly advertise their compliance with the code. The accredited ISPs would be able to display a “trust mark” signifying their status, help-ing consumers to make reasonable decisions about which service to choose.52

• The code should also include a mandatory review or “sunset” provision to ensure that the rules remain effective and appropriate in a fast-changing technological and legal environment.

7.3.2 Elements of a Model Industry Code of Conduct

Like a model law, an industry code of conduct should be developed in an inclusive, collaborative atmosphere, designed to elicit the best thinking from a range of experts and concerned stakeholders.53 The code should set forth the re spon sibilities of ISPs and other actors with sensitivity to local concerns. But it should also take into account the cross-border nature of the problem. Key elements of a model industry code of conduct might include:• A series of common definitions that correspond to the

definitions in the enabling law.• Procedures ISPs should follow in dealing with obvious

spam that comes into the ISP’s sub-network (including procedures relating to the provision or use of filtering software).

• A commitment not to serve individuals or companies that send unsolicited commercial email in bulk, and to termi-nate those clients when complaints and subsequent inves-tigations reveal that they have been spamming through the ISP’s network. This should also include a commitment to refuse payment, or any enticement of a premium payment, offered by a known spammer for any service.

• A commitment to give ISP subscribers information about the availability and use of software for filtering spam at the client level. ISPs should also commit to helping sub-scribers prevent their computers from being infected by

worms, “Trojans” and other malware that turns comput-ers into spam “zombies.”

• A commitment to assist in developing and evaluating fil-tering software that gives end users a maximum level of control over what to accept and to reject.

• Suggested best practices that ISPs can implement, as appropriate, in order to minimize or prevent spam. At present, such suggested best practices might include some of those set forth in the London Action Plan.54

The London Action Plan stemmed from a July 2004 meeting of “government and public agencies from 27 countries responsible for enforcing laws concerning spam.” They gener-ated several recommendations affecting:• The optimal configuration of servers and other network

devices to minimize or prevent spam;• A commitment to taking meaningful zombie-prevention

measures;55 and, • A statement of principles for entering into peering

arrangements only with ISPs that adhere to the full code of conduct.

The provisions of codes will no doubt change rapidly as the nature of the problem changes. Today, up to half of all spam is sent through “zombie” computers, suggesting that it is vital to help end users prevent the hijacking of their computers. Once this loophole is closed, spammers are sure to look for other mechanisms, and codes will have to be updated accord-ingly. The enabling law should be flexible enough to accom-modate changes in the technological landscape.

7.3.3 Hazards of Enforceable Codes of ConductAdopting a regime of enforceable codes of conduct for

ISPs is not without hazards. A well-designed policy, however, should be able to mitigate these risks, which are worth explor-ing here.

The purpose of industry codes of conduct should be to give ISPs incentives to exclude spammers from their networks, not to over-regulate ISPs. Nor should regulators use codes to deputize ISPs to overzealously block email or monitor conver-sations. Codes should be strictly limited to requiring ISPs to shut down spammers. They should not be employed for other objectives, such as shutting down email with what the govern-ment considers unpalatable political messages or for surveil-lance of a country’s citizens. The risk is that empowering ISPs as gatekeepers will lead them to avidly look into the nature of messages sent across their networks.

This potential pitfall points back to the importance of defining spam in the anti-spam law. A properly crafted law should rule out abuses of authority in the name of preventing spam. Regulators should clearly focus on the goal of weeding out the worst, most obvious cases of spamming, rather than on pressuring ISPs to shut down legitimate e-mailers.

Another risk in establishing an enforceable code of con-duct stems from political realities. In many countries, ISPs have enjoyed broad immunity from regulation and may oppose any spam-related re spon sibilities. More often, the ISP may be a monopoly, state-owned provider that generates important rev-

CHAPTER 7 121


Box 7.1: Australia Telecommunications Act 1997 – SECT 117

Registration of industry codes

1) This section applies if:

a) the ACMA is satisfied that a body or association represents a particular section of the telecommunications industry or the e-marketing industry; and

b) that body or association develops an industry code that applies to participants in that section of the industry and deals with one or more matters relating to the tele com mu ni cations activities or e-marketing activities, as the case may be, of those participants; and

c) the body or association gives a copy of the code to the ACMA; and

d) the ACMA is satisfied that:

i) in a case where the code deals with matters of substantial relevance to the community-the code provides appropriate community safeguards for the matters covered by the code; or

ii) in a case where the code does not deal with matters of substantial relevance to the community-the code deals with the matters covered by the code in an appropriate manner; and

e) the ACMA is satisfied that, before giving the copy of the code to the ACMA:i) the body or association published a draft of the code and invited participants in that section of the industry to make

submissions to the body or association about the draft within a specified period; and

ii) the body or association gave consideration to any submissions that were received from participants in that section of the industry within that period; and

f) the ACMA is satisfied that, before giving the copy of the code to the ACMA:i) the body or association published a draft of the code and invited members of the public to make submissions to the

body or association about the draft within a specified period; and

ii) the body or association gave consideration to any submissions that were received from members of the public within that period; and

g) the ACMA is satisfied that the ACCC has been consulted about the development of the code; and

h) the ACMA is satisfied that the Telecommunications Industry Ombudsman has been consulted about the develop-ment of the code; and

i) the ACMA is satisfied that at least one body or association that represents the interests of consumers has been consulted about the development of the code; and

j) in a case where the code deals with a matter set out in paragraph 113(3)(f)-the ACMA is satisfied that the Privacy Commissioner has been consulted by the body or association about the development of the code before the body or association gave the copy of the code to the ACMA; and

k) the ACMA has consulted the Privacy Commissioner about the code and consequently believes that he or she is satisfied with the code, if the code deals directly or indirectly with a matter dealt with by:

i) the National Privacy Principles (as defined in the Privacy Act 1988 ); or

ii) other provisions of that Act that relate to those Principles; or

iii) an approved privacy code (as defined in that Act) that binds a participant in that section of the tele com mu ni cations industry or the e-marketing industry; or

iv) provisions of that Act that relate to the approved privacy code.

2) The ACMA must register the code by including it in the Register of industry codes kept under section 136.

3) A period specified under subparagraph 1) e) i) or 1) f) i) must run for at least 30 days.

4) If:

a) an industry code (the new code ) is registered under this Part; and

b) the new code is expressed to replace another industry code;

the other code ceases to be registered under this Part when the new code is registered.

Note: An industry code also ceases to be registered when it is removed from the Register of industry codes under section 122A. http://www.austlii.edu.au/au/legis/cth/consol_act/ta1997214/s117.html

CHAPTER 7


122

enues for the government, giving it substantial clout in policy councils. Again, it may resist any attempts to further regulate it.

Meanwhile, there are costs associated with any new ad min-is trative mechanism, even one as simple as the development, registration, and updating of a code of conduct. Potential costs should be factored into the cost-benefit analysis when consid-ering adopting such a regime.

Additionally, adding intelligence to the middle of the net-work, and encouraging gatekeepers to use this intelligence, is sub-optimal from a network design perspective. Like regulators in developing countries, ISPs may themselves face resource constraints to enforce their code. ISPs may or may not see suf-ficient incentives to do so. ISPs often have to balance multiple interest and desires regarding spam, including:• A desire to attract and retain bad-acting but paying cus-

tomers,• A desire to avoid the cost of transmitting spam through

their networks, and• A desire to avoid the regulatory risks and costs of trans-

mitting spam.

ISPs may over-enforce the provisions of their own code, resulting in messages not getting delivered to recipients. This would be a far worse outcome, many argue, than dealing with the current deluge of spam. ISPs may also not be as sensitive to the rights of free expression, and most speech protections do not extend to non-state actions, often allowing private actors to block otherwise protected speech.

Meanwhile, ISPs would likely pass anti-spam costs along to end users, perpetuating the already-vicious cycle of spam-mers making the rest of the internet’s users pay for their bad acts. In a developing country context, high internet access costs are already a major barrier to widespread ICT adoption. Cost concerns, however, should be seen in the context of the spam problem itself, which is adding to the cost of internet access and helping criminals to perpetrate fraud and disseminate destructive viruses. These network ills are bad not only for consumers, but for ISPs themselves.

Any legal and regulatory approach should seek to mitigate these drawbacks. On balance, however, many jurisdictions will likely find enforceable codes of conduct to be a sound policy choice, because they distribute part of the enforcement burden to stakeholders closest to the source of the spam problem – the ISPs and the end users.

7.4 Education and Awareness

The ideal solution to spam would involve no new law whatsoever. If consumers and businesses could take spam fighting into their own hands, the problem would be solved at the lowest cost and at the quickest rate. The brunt of anti-spam enforcement would be borne at the furthest edges of the net-work and in the most distributed manner possible. Those who pay the true costs of spamming – the end users – would ideally take the lead in combating spam, while regulators focused their enforcement resources on the largest, most complex cases.

Regulators would still have an important role to play, how-ever, in educating consumers, businesses and ISPs about the dangers of spam and the steps they can take to protect them-selves against it.56 The London Action Plan includes some sug-gestions:

• Regulators should develop a plan for consumer and ISP education, posting information on their websites and developing print ma teri als for distribution to cybercafé owners, consumers, businesses and ISPs.

• Regulators should provide a simple method for consum-ers to make complaints about spam.

• Regulators should create a special “combating spam” page on their websites, providing information about anti-spam practices and products. The web page should host prac-tical advice on spam filters, warnings about phishing attempts, viruses and scams carried out using e-mail and other important tips for consumers. Examples of websites in use today include:

– Industry Canada’s page on “Recommended Best Practices for Internet Service Providers and Other Network Operators”: http://e-com.ic.gc.ca/epic/internet/inecic-ceac.nsf/en/gv00329e.html

– Recommendations of the Commission Nationale de l’Informatique et des Libertés in France (CNIL République Française): http://www.cnil.fr/index.php?id=1539

– Guidance provided by the Korea Spam Response Centre of the Korea Information Security Agency, an affiliated agency of the Ministry of Information and Communication: http://www.spamcop.or.kr/eng/m_3_2.html

– The United States’ Federal Trade Commission’s spam education pages: http://www.ftc.gov/bcp/conl ine /pubs /buspubs / secureyourserver.h tmhttp: / /www.ftc .gov/bcp/conline/edcams/spam/secureyourserver/index.htm

• Regulators should also consider their ability to play a central role in coordinating the sharing of best practices among ISPs, especially in contexts where political will or resources do not exist for the regulator to take an enforce-ment role. The regulator can also help educate ISPs about some relatively simple technical measures. Specific meas-ures include the latest information related to the blocking of open relays,57 focus on “botnets,”58 and slowdowns of traffic on port 25 that might make an enormous differ-ence, particularly in developing countries.

Consumer and ISP education is a necessary component of spam-fighting strategies, but efforts in this field have had little effectiveness to date. This is not due to any fault in the outreach techniques themselves, but rather due to the limited vigour with which they have been pursued. It is challenging to communicate technical information to a lay audience. Moreo-ver, education efforts cannot succeed in isolation, without other effective technological and regulatory measures. Substantially greater efforts in this area are warranted and would pay large dividends.

CHAPTER 7 123


1 Despite passage of many dozens of anti-spam statutes in jurisdictions across the globe, the problem has continued to worsen. See, e.g., David E. Sorkin, “Spam Legisla-tion in the United States,” The John Marshall Journal of Computer and Information Law, Volume XXII, Number 1, at 4 (2003) (“…it is generally agreed that legislation has failed to solve the spam problem.”) See also, Matthew Prince, “How to Craft an Effective Anti-Spam Law,” WSIS Thematic Meeting on Countering Spam, July 2004, ITU Discussion Paper, at 10, at http://www.itu.int/osg/spu/spam/contributions/Background%20Paper_How%20to%20craft%20and%20effective%20anti-spam%20law.pdf (“Few people would dispute that around the world the first generation of anti-spam laws has been an unqualified failure.”).

2 http://www.itu.int/ITU-D/treg/Events/Seminars/2004/GSR04/index.html3 Business Software Alliance, 1 in 5 British Consumers Buy Software from Spam, Dec. 9, 2004, at http://www.bsa.org/uk/press/newsreleases/online-shopping-tips.cfm.4 For instance, e-mail security provider IronPort Systems asserts that 72 per cent of e-mail sent is spam. See http://www.ironport.com/company/pp_sci-tech_today_08-

10-2005.html. 5 http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=108_cong_public_laws&docid=f:publ187.108.pdf 6 See Matthew Prince, “How to Craft an Effective Anti-Spam Law,” supra note 1, at 3. 7 For the most comprehensive resource on the world s anti-spam laws, see Christina Bueti, “ITU s Survey on Anti-Spam Legislation Worldwide,” July, 2005, at http://

www.itu.int/osg/spu/cybersecurity/docs/Background_Paper_ITU_Bueti_Survey.pdf. 8 AOL claims that spam is down 85 per cent from two years ago, based upon consumer complaint information. However, such a claim does not account for the ef-

fectiveness that their filters may have achieved on behalf of customers, nor the changing perceptions of consumer about how much spam is acceptable. The same article that reported AOL s claim of less spam concludes, “But statistics show that the amount of spam is still huge – even worse than it was when the federal act [the CAN-SPAM Act of 2003] was introduced two years ago.” See http://www.crmbuyer.com. See also http://www.washingtonpost.com/ (27 December 2004). There is a dearth of reliable industry-wide data, which is not surprising in light of the distributed nature of the problem and the competition between ISPs to provide the best anti-spam services to consumers.

9 For a review of some of the many recent spam statistics, see Bueti, “ITU s Survey on Anti-Spam Legislation Worldwide,” supra note 5; see Michael Geist, “Untouch-able: A Canadian Perspective on the Anti-Spam Battle,” June, 2004, at 2, at http://www.michaelgeist.ca/geistspam.pdf; see also, Derek Bambauer, John Palfrey, and David Abrams, “A Comparative Analysis of Spam Laws: the Quest for Model Law,” June 2005, at 7 – 8, at http://www.itu.int/osg/spu/cybersecurity/docs/Back-ground_Paper_Comparative_Analysis_of_Spam_Laws.pdf.

10 “Phishing” refers to a scam in which perpetrators send an email purporting to be from a legitimate business (such as a bank) and ask recipients to provide personal (often financial) information. Victims believe they are complying with a bona fide request, when they are being tricked into providing information to thieves. “Pharm-ing” refers to a scheme in which victims clicking on a website are unknowingly diverted to a duplicate or fake website, where they can be fleeced.

11 See Chairman s Report, ITU WSIS Thematic Meeting on Cybersecurity, June – July, 2005, p. 2, point 12, at http://www.itu.int/osg/spu/cybersecurity/chairmansreport.pdf (citing a speech by Spamhaus CEO Steve Linford).

12 http://news.bbc.co.uk/2/hi/business/3426367.stm.

7.5 ConclusionDespite the challenges that are bound to lie ahead, regula-

tors should encourage the adoption of an anti-spam law that is harmonized, as much as possible, with those of other countries. Such an anti-spam law might involve creating an enforceable code of conduct for ISPs, placing the re spon sibility for miti-gating spam closer to where the technical expertise lies. The problem with anti-spam laws enacted to date is that they have failed to create an enforceable regime or to bridge the divide between governments and the technologists who have the real expertise to solve the problem. While it is an imperfect remedy,

an enforceable code of conduct could help to erase the short-comings of earlier anti-spam laws.

The effort to fight spam is not going to succeed through pursuit of any one, single strategy. Success will be based on international cooperation and a range of shared strategies, including legal and regulatory mechanisms, technical improve-ments, market forces, and consumer-oriented solutions. The development of ISP codes of conduct, and their enforcement by regulators, can help stem the tide of spam and ma teri ally reduce spam’s costs to ISPs and consumers.

CHAPTER 7


124

13 The AOL legal department posts decisions and litigation to their website at http://legal.web.aol.com/decisions/dljunk/. See also http://www.theregister.co.uk/2005/08/10/aol_spam_sweepstake/ (regarding the AOL gold bars raffle, in which they planned to give away the assets seized from a major spammer).

14 See http://abcnews.go.com/Technology/PCWorld/story?id=1029922&ad=true. 15 This discussion paper uses the term “regulators” in the broad sense to include any governmental entity that has been given the mandate to combat spam. Thus, the

term “regulators” for this chapter may mean national tele com mu ni cations or ICT regulatory authorities, consumer protection authorities or data protection ad min-is trations.

16 See David R. Johnson, Susan P. Crawford, and John G. Palfrey, Jr., The Accountable Net: Peer Production of Internet Governance, 9 VA. J. L. & TECH. 9 (2004).17 BBC, supra note 10.18 The four modes of Internet regulation were popularized in Lawrence Lessig s ground-breaking book, Code and Other Laws of Cyberspace, in 1999 (New York: Basic

Books). 19 See http://www.itu.int/osg/spu/spam/background.html and, in particular, the Chairman s Report, at http://www.itu.int/osg/spu/spam/chairman-report.pdf. 20 Ibid., point 24 at 4.21 It should be noted that even the United States Federal Trade Commission, which is a relatively well-funded regulatory body, had only brought “over 70 cases” as of

July, 2005. In light of the billions of spam messages per day, the notion that such an enforcement effort is unlikely to have much effect undoubtedly is apparent to many governments choosing whether or not to devote resources to fighting spam locally. Ibid., point 19, at 3.

22 See Suresh Ramasubramanian, “OECD Task Force on Spam Report: Spam Issues in Developing Countries,” May, 2005, at http://www.oecd.org/dataoecd/5/47/34935342.pdf.

23 See http://www.itu.int/ITU-D/treg/related-links/links-docs/Spam.html for a list of voluntary and enforceable ISP codes of conduct.24 See generally Jonathan Zittrain, “Internet Points of Control,” 43 Boston College Law Review 653 (2003). See also, J.H. Saltzer, D.P. Reed, and D.D. Clark, “The

End-to-End Argument in Systems Design,” at http://www.reed.com/Papers/EndtoEnd.html and “The End of the End-to-End Argument” at http://www.reed.com/dprframeweb/dprframe.asp?section=paper&fn=endofendtoend.html (“But in many areas of the Internet, new chokepoints are being deployed so that anything new not explicitly permitted in advance is systematically blocked.”)

25 See John Spence, “Pennsylvania and Pornography: CDT v. Pappert Offers a New Approach to Criminal Liability Online,” 23 J. Marshall J. Computer & Info. L. 411 (Winter, 2005) (a good general discussion of the role of ISPs in the network and the difficulties they face).

26 http://www.maawg.org/about/roster/27 Many technical working groups have focused on anti-spam-related standards, technologies, and best practices. The IETF, ISOC, and other groups have supported

efforts that have involved rep re senta tives of ISPs, including the now-scuttled MARID Project (see http://www.internetnews.com/bus-news/article.php/3407431), which was preceded by the Anti-Spam Research Group (at http://asrg.sp.am/).

28 See Renai LeMay, Gmail Tries Out Antiphishing Tools, CNET NEWS.COM, Apr. 4, 2005, at http://news.com.com/Gmail+tries+out+antiphishing+tools/2100-1029_3-5653794.html.

29 See Anick Jesdanun, Battle Against Spam Shifts to Containment, ASSOCIATED PRESS, Apr. 15, 2005, at http://finance.lycos.com/qc/news/story.aspx?story=48398343.30 Consider the remarks of Randall Boe, executive vice president of AOL, when he said that “Spam has become the single largest customer problem on the Internet.”

(Quoted in Thomas Claburn, “Four Big ISPs Sue Hundreds of Spammers,” 10 March 2004, Information Week, at http://www.informationweek.com/).31 As one illustration of the fact that spam can be traced, see http://www.channelregister.co.uk/2005/09/20/spam_map/.32 Consider, for instance, that MAAWG is already promoting industry-wide codes of conduct. See http://www.maawg.org/about/. 33 Bambauer, Palfrey, and Abrams, “A Comparative Analysis of Spam Laws: the Quest for Model Law,” supra note 9, at 11.34 Prince, “How to Craft an Effective Anti-Spam Law,” supra note 1, at 4.35 Ibid., at 6. Mr. Prince argues: “The most effective anti-spam laws are action laws that focus on the problems prosecutors face and work to resolve them. If we want

anti-spam laws to be effective, our job must be to identify the costs faced by prosecutors and craft laws to reduce those costs.”36 Accessible online at http://www.uncitral.org/pdf/english/texts/electcom/05-89450_Ebook.pdf.37 See http://www.itu.int/osg/spu/spam/ for a catalogue of existing anti-spam laws on the books in jurisdictions around the world.38 Many analysts predicted the failure of these laws at the time they were passed. For one example of a United States-based consultancy, consider Gartner s report,

Maurene Caplan Grey, Lydia Leong, Arabella Hallawell, Ant Allan, and Adam Sarner, “Spam Will Likely Worsen Despite US Law,” 3 December 2003, at http://www.gartner.com/resources/118700/118762/118762.pdf.

39 See BBC News, “US Still Leads Global Spam List,” 7 April 2005, at http://news.bbc.co.uk/1/hi/technology/4420161.stm (citing a study by security firm Sophos that the US is responsible for sourcing 35 per cent of the world s spam).

40 See the FAQ page for the Coalition Against Unsolicited Commercial Email, at http://www.cauce.org/about/faq.shtml#offshore.41 One interesting, as-yet-theoretical variant to the state-focused enforcement mechanism is the “bounty hunter” system proposed by Prof. Lawrence Lessig of Stanford

Law School. Prof. Lessig has “bet [his] job” on the notion that such a distributed system, established by law but pushing out enforcement authority to netizens, would work if enacted. See http://www.lessig.org/blog/archives/000787.shtml.

42 The Australian law, which took effect in 2003, can be found online (in an unofficial version) at http://scaleplus.law.gov.au/html/pasteact/3/3628/0/PA000260.htm43 For example, the text of the Communications Decency Act Section 230 in the United States provides immunity to the providers of “interactive computer services”

for the content published on their network. These providers are defined as follows: “The term ̀ interactive computer service means any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server, including specifically a service or system that provides access to the Internet and such systems operated or services offered by libraries or educational institutions.” http://www.fcc.gov/Reports/tcom1996.txt. By contrast, the term “Internet access service” in the CAN-SPAM Act of 2003, as stated in the Telecommunications Act of 1934, as amended, reads: “The term Internet access service means a service that enables users to access content, information, electronic mail, or other services offered over the Internet, and may also include access to proprietary content, information, and other services as part of a package of services offered to consumers. Such term does not include tele com mu ni cations services.” http://www4.law.cornell.edu/uscode/html/uscode47/usc_sec_47_00000231----000-.html.

44 Geist, “Untouchable,” supra note 8, at 17 (for a discussion of civil and criminal sanctions common in anti-spam legislation).45 For discussion of the effectiveness of such a measure, see Prince, “How to Craft an Effective Anti-Spam Law,” supra note 1, at 9.46 http://www.icpen.org/.47 For discussion of the effectiveness of the state of Washington s use of such a measure in the United States, see Prince, “How to Craft an Effective Anti-Spam Law,”

supra note 1, at 6 and 10.

CHAPTER 7 125


48 For the full text of the Australian Telecommunications Act of 1997 that contains such provisions, see http://www.austlii.edu.au/au/legis/cth/consol_act/ta1997214/s117.html et seq.

49 The Australian Direct Marketing Association (ADMA) has also established a Code of Conduct. Where such an organization exists, such a code is another logical, parallel step. Many countries will not have such an entity in place, in which event a legal provision mandating a parallel process of this sort would not make sense.

50 Consider the findings of the New Zealand regulators with respect to the most effective mode of enforcement: “A civil penalty regime where the emphasis is on ISPs/carriers taking action in response to customer complaints is considered to be the best approach. This is because most spam in New Zealand originates from overseas and the ISP/carrier will often best be placed to put in place the appropriate technical measures to deal with it. In addition, if spam is originating from an address/number hosted by another ISP/carrier in New Zealand, then the user s ISP/carrier can approach the sender s ISP/carrier and seek action by that ISP/carrier against the sender. If complaints cannot be satisfactorily resolved in this way then the user s ISP/carrier can forward the matter on to the enforcement agency to consider whether an investigation or further action is appropriate.” Ministry of Economic Development (NZ), “Legislating against Unsolicited Electronic Messages Sent for Marketing or Promotional Purposes (Spam) – Enforcement Issues – Cabinet Paper,” at http://www.med.govt.nz/pbt/infotech/spam/cabinet/paper-two/paper-two-03.html#P31_3192.

51 G-REX is an online discussion platform reserved for policy-makers and regulators> For more information, see: http://www.itu.int/ITU-D/grex/index.html.52 See http://www.truste.org/. 53 The process under way at the Messaging Anti-Abuse Working Group may well provide extremely useful guidance on this front, both as a matter of process and of

substance. See http://www.maawg.org/news/maawg050711. 54 See http://www.ftc.gov/os/2004/10/041012londonactionplan.pdf. See also, for particular suggestions, http://www.ftc.gov/bcp/conline/edcams/spam/zombie/index.htm.

For a letter sent to 3,000 ISPs, as part of this initiative, see http://www.ftc.gov/bcp/conline/edcams/spam/zombie/letter_english.htm. 55 The specific suggestions for such zombie-prevention measures will vary over time. Some initial recommendations, derived as part of the London Action Plan meeting

and related efforts, include: 1) blocking port 25 except for the outbound SMTP requirements of users authenticated by the ISP to run mail servers designed for client traffic and other carefully accredited purposes; 2) exploring im ple men ta tion of Authenticated SMTP on port 587 for clients who must operate outgoing mail servers; 3) applying rate-limiting controls for email relays; 4) identifying computers that are sending atypical amounts of email, and take steps to determine if the computer is acting as a spam zombie. When necessary, quarantining the affected computer until the source of the problem is removed; 5) providing, or pointing customers to, easy-to-use tools to remove zombie code if their computers have been infected, and provide the appropriate assistance; and, 6) the shutdown of open relay servers after appropriate notice and inquiry. Regarding the first of these suggestions, related to port 25, Industry Canada (in a separate context), recommends, “ISPs and other network operators should limit, by default, the use of port 25 by end-users. If necessary, the ability to send or receive mail over port 25 should be restricted to hosts on the provider s network. Use of port 25 by end-users should be permitted on an as-needed basis, or as set out in the provider s end-user agreement / terms of service.” http://e-com.ic.gc.ca/epic/internet/inecic-ceac.nsf/en/gv00329e.html.

56 The New Zealand regulators note: “The enforcement agency would be seen as also having a role in educating users/consumers on how to deal with spam in conjunc-tion with the industry as well as a role in educating business and other organisations on how to comply with the legislation along with the Ministry of Economic Development, which will be responsible for administering the legislation, and organisations such as the Direct Marketing Association.” Ministry of Economic Development (NZ), “Legislating against Unsolicited Electronic Messages Sent for Marketing or Promotional Purposes (Spam) – Enforcement Issues – Cabinet Paper,” supra note 47, at http://www.med.govt.nz/pbt/infotech/spam/cabinet/paper-two/paper-two-03.html#P31_3192.

57 For a description of open mail relays and their importance to the spam issue, see http://en.wikipedia.org/wiki/Open_mail_relay. 57 For a definition of botnet, see http://en.wiktionary.org/wiki/botnet.

127


CHAPTER 8

8 MAKING BROADBAND WORK FOR ALL

Regulators and policy-makers view the advent of broad-band networks and services as both a challenge and an oppor-tunity. What, really, is “broadband”? To the layman, it often translates as “fast internet access.” But as the edition of Trends has illustrated, the broadband revolution is a multi-layered phenomenon, with technological, economic and social aspects that will become more apparent as this decade wears on. One thing is clear: broadband is not a passing phase or a high-end niche market. It is literally the future of tele com mu ni cations. Policy-makers and regulators around the world will have to come to grips with it and learn to exploit its opportunities – in other words, learn how to make broadband work for all.

8.1 What Is Broadband?

In the most basic definition, broadband refers to an array of digital, packet-switched network technologies that allow the transport of digital bits at high speeds. These technologies are both wireless and wire-line, and they include both upgrades to existing networks (for example, xDSL or 2.5G networks) and entirely new infra struc ture (such as all-fibre networks, WLANs and 3G systems). Generally, networks with bandwidth capaci-ties of 256 kbit/s or more can be termed “broadband,” although that threshold may well shift higher as new technologies push the envelope on throughput.

What may be more important than network capacity is what one can do with broadband networks. Instead of the old, single-purpose networks, broadband networks can carry any combination of voice, data and multimedia (graphics, video and audio), in any format. Indeed, broadband networks are already generating new permutations on old media: audio “podcasts” downloaded to portable players from websites, chat functions incorporated into online video games. The list goes on. The evolution of new applications is suddenly without boundaries, and human ingenuity is now free to pursue services and appli-cations that will improve lives and bolster economies.

The obvious conclusion is that the term “broadband” does not just mean an interesting set of network technologies. It is an entirely new paradigm, potentially as different from stand-ard voice telephony as telephony is from the telegraph services of 150 years ago. Never before has there been such power to combine images and information in ways that can actually aug-

ment the user’s experience into something more enriched than actually being there.

8.2 Why Should I Care about Broadband?

Broadband networks amplify the internet, certainly, but they go much further than that. A doctor 10,000 kilometres away can check the progress of a patient’s care using a remote live-camera feed, with an in-screen box showing the patient’s records to facilitate diagnosis and treatment. A master of tradi-tional dance in one country can teach students the intricacies of an ancient dance step to émigrés in a hundred other coun-tries at the same time, ensuring its survival for generations to come. Such moments are as ephemeral as the flap of a butter-fly’s wing – and just as priceless. Now imagine such moments taking place literally millions of times each day across disci-plines ranging from dance to dentistry to demographics.

The single most important thing to absorb about broad-band technologies is that they drive intelligence and ingenuity to the edge of networks. More than ever before in the history of tele com mu ni cations, it will be not so much the network but rather the people connected to it that count. ICT technol-ogy may never catch up to human creativity and diversity, but broadband networks will allow it to remain closely linked. The power of computing to generate and organize knowledge – or to germinate and nurture art – will suffocate without the media to convey it from one person to another. Broadband networks empower individuals and groups to create and collate, innovate and inspire, without restrictions of time and distance.

As they empower individuals, broadband capabilities will increase the potential for generating content that will be rel-evant, meaningful and understandable to communities. The key to sustainable network services is demand. And the key to demand is providing useful, culturally sustaining content, in local languages, about local circumstances as well as global realities. Although it certainly will not happen overnight, there is no reason why individuals in the remotest areas cannot even-tually become broadcasters in their own communities, educa-tors in their own homes, and performing artists for worldwide audiences.

CHAPTER 8


128

8.3 How Can I Get Broadband?

For increasing numbers of consumers, the answer is that they already have it – and are likely to get more of it. In many countries, broadband is now available in several different user niches: at home on a desktop PC or with a Wi-Fi-equipped laptop in the airport. For these lucky users, the future will be about convergence onto multiple platforms – interactive digital televisions, broadband mobile phones and streaming video on computers, just to name a few – and inter-modal competition.

For developing countries, the key to the broadband future is the flexibility of the technologies coupled with the declining costs of the network topologies. Broadband technologies can increasingly be either fixed or mobile, and they can convey any mélange of voice, data and multimedia content – all at a lower marginal cost than earlier-generation, circuit-switched tele com-mu ni cations networks. Moreover, advances in infra struc ture

– particularly with wireless networking standards in the Wi-Fi and WiMAX families – will allow more and more broadband capabilities at lower cost.

In the context of developing markets, broadband will clearly have to dovetail with the prevalent mobile flavour of the tele com mu ni cation sectors there. For one thing, the age of the ubiquitous wire-line network may never arrive in many countries. And certainly, the broadband revolution will not be obliged to wait around for it. Access networks are likely to con-tinue to be mobile and wireless – including broadband wire-less access (BWA) technologies. Mobile and fixed wireless will likely converge, bringing a broadband capability to markets that are essentially mobile, and will continue to be. For transport and backhaul, a combination of network types will have to be employed. This will undoubtedly include terrestrial wireless links, some satellite hops and lit fibre.

Non-traditional suppliers will have to be part of the mix in developing countries. For backhaul, this may mean infra-struc ture sharing between the fibre networks of universities and the transport and energy sectors as well as leased fibre from the private sector. In the access network sector, the decentral-ized nature of broadband networks will enlist smaller-scale, local and regional operators – community groups, universi-ties, municipalities and entrepreneurs – to set up “hotspots” or wide-area networks that can be linked back to larger opera-tors’ networks. The network can be built from the periphery inwards, by local operations providing local content and serv-ices, generating demand from the ground up.

8.4 What Can Regulators Do?

Regulators and policy-makers around the world are reviewing their laws and regulations to judge whether they provide a proper environment to help speed the opportunities and benefits of broadband networks and the new services and applications that ride on them. At this juncture, responses have been varied – everything from initiating an overhaul of licens-ing and market-entry policies to doggedly restricting access to VoIP services. As with other aspects of sector liberalization, many governments feel the instinct to protect incumbents and ensure that they take the lead in broadband investment.

It may do little good, however, to liberalize traditional telephony markets, or even mobile service markets, while pro-tecting incumbents from inter-modal or broadband competi-tion. Growing numbers of countries are adopting regulations that allow for open market entry at all levels and layers, includ-ing applications and services such as VoIP and internet access. Of course, in markets undergoing a transition to competition, it is important for regulators to work towards a level playing field for providers of IP-based services. This will often mean taking steps to ensure that network operators inter con nect and provide open access to support infra struc ture and some net-work capabilities.

In many areas of regulation, it may be possible to “regu-late down” rather than “regulate up” to achieve competition. That is, as competition becomes viable, and market forces begin to discipline operators’ behaviour, it may be wiser to reduce the regulatory burden on all operators – incumbents and new entrants alike – rather than imposing the same regula-tory structure on competitors that has always been applied to the monopoly incumbent. Regulators must strike a balance between giving incumbents too much latitude to obstruct competition, on the one hand, and so many restrictions that it stifles broadband investment, on the other.

One key area of regulatory practice, of course, is spectrum management, and the role of spectrum management takes on greater importance with the rise of broadband wireless access (BWA) technologies. At the same time that spectrum-hungry BWA equipment is being developed and deployed, other advances, such as in the computing power of processors in radio equipment, are causing more ferment in the spectrum community than perhaps at any time in its history. The tradi-tional paradigm of spectrum planning, allocation, assignment and monitoring reflect largely the technological limits of an earlier era. New radio technologies may free regulators from some legacy spectrum management practices that may become obsolete, or worse, too rigid for a wireless broadband world. Technological developments such as spectrum-hopping, adap-tive and directional antennae use and other techniques promise to shift interference management from governments to opera-tors and even to end-user terminals to allow greater sharing and reuse of existing spectrum.

Spectrum regulators have begun to respond by granting licensees with more flexibile use of technologies so that opera-tors can deliver the services the market demands. This tech-nology and service “neutrality” reinforces the overall trends of decentralizing control over networks. This will allow the application of BWA technologies, for example, to be more agile in responding to local market realities. Indeed, in developing countries generally, and in remote areas within developing countries, in particular, where spectrum scarcity is far less an issue than in developed countries, there is no reason why spec-trum management policies could not be tailored to local reali-ties, which may include a less dense environment of spectrum use, allowing greater power and range for wireless systems that do operate there.

In the end, perhaps the best analogy of the future broad-band wireless environment is the highway analogy. Anyone

CHAPTER 8 129


can get on the highway, as long as they obey the rules of the road. That is, as long as operators obey rules designed to limit harmful interference, entry barriers can be lowered to allow as many operators as demand warrants. In congested areas, of course, those rules may need to be more stringent – as speed limits are on urban highways. In the end, however, it makes little sense to restrict market entry when technical demands do not require it. That would be akin to establishing a nationwide limit of cars allowed on any highway anywhere in the country, based solely on the potential for congestion at the single busi-est spot, at rush hour, in the country’s capital city.

In the final analysis, policy-makers and regulators need to address the broadband revolution in three ways. First, they need to accept the reality of its coming and embrace its poten-tial. Second, they need to radically revise the way they under-stand tele com mu ni cations, to understand all the options they might pursue in accommodating the new networks, services and applications. Finally, and most importantly, they need to plan, with the advice and input of their industries, academic institutions and civil societies, how to make broadband work for all.

REGULATORY TABLES 131


REGULATORY TABLES*

Table 1 Countries with a separate Regulatory Authority ................................................................ 133

Table 2 Status of the main fixed-line operators ............................................................................ 169

Table 3 Level of competition ....................................................................................................... 203

* Regulatory profiles as well as contact details for policy-makers and regulators are available on the TREG website at: http://www.itu.int/ITU-D/treg.This data is extracted from the ITU World Telecommunication Regulatory Database.



WORLD

Name of AuthorityYear

createdFinanced byLegal document that created

the authority

Country

1. Countries with a separate Regulatory Authority

Reports to Is it a collegial body?

AfghanistanAfghanistan Telecom Regulatory Authority (ATRA) Website: http://www.trb.gov.af

2003 Presidental Decree Sector Ministry (annual report)

Yes: 5 MembersRegulatory fees: 100%

AlbaniaTelecommunication Regulatory Entity Website: http://www.ert.gov.al

1998 Law 8288,18.2.1998 on TelecommunicationRegulatory Entity

LegislatureOther: Council of Ministers

Yes: 5 Members

Financial income: 44.4%Other: 0.1%

Licence fees: 17.1%Spectrum fees: 38.3%Numbering fees: 0.1%

AlgeriaAutorité de régulation de la poste et des télécommunications (ARPT) Website: http://www.arpt.dz

2000 Loi n° 2000-03 Legislature Yes: 7 Members

Other: 20.6% Various products, research fees, ICT training and standardization

Award/auction of other licence: 17.7%Spectrum fees: 60.1%Numbering fees: 1.5%

Angola1

Institut Angolais des Communications (INACOM) Website: http://www.inacom.og.ao1999 Décret nº 12/99 du 25 juin Sector Ministry

(annual report)Yes: 5 Members

Other: 3%

Award/auction of mobile licence: 30%Award/auction of other licence: 2.0%Spectrum fees: 65%

Source: ITU World Telecommunication Regulatory Database.2004 data pre-2004 data1 2

REGULATORY TABLES


134

WORLD



the authority

Country



Argentina2

Comisión Nacional de Comunicaciones (CNC) Website: http://www.cnc.gov.ar 1990 Decreto 1185 de fecha 22 de

junio de 1990Yes: 8 Members

Contributions from operator turnover: 42%Financial income: 0.5%Other: 2.5%

Spectrum fees: 55%

AustraliaAustralian Communications and Media Authority (ACMA) Website: http://www.acma.gov.au

1997 Austel - Telecommunications Act 1989. ACCC - Trade Practices Act 1974. ACA - ACA Act 1997. Australian Communications and Media Authority Act 2004

LegislatureSector Ministry (annual report)

Yes: 7 Members

Contributions from operator turnoverOther: ACMA and ACCC telco regulatory costs are recovered from carriers via carrier licence fees, based on each carrier's % of total carrier eligible revenue. Costs of administering numbering arrangments are recovered through numbering charges. 'Once off' fees (eg. spectrum auction) are paid directly to general goverment consolidated revenue. At the time of auction, a recurrent annual tax is set for each band in the spectrum.Bidders who use the particular band are taxed pro-rata for the megahertz population in the band that they use, i.e. their share of the total band tax.

Licence feesGovernment appropriationSpectrum feesNumbering fees

Austria1

Telecom-Control-Commission (Telekom-Control-Kommission, TKK) Website: http://www.rtr.at1997 Telecommunications Act 1997 Sector Ministry

(annual report)Yes: 3 MembersContributions from operator

turnover: 70,5 %Other: 29,5% Broadcasting activities are financed by broadcasting companies (Data for 2003)




WORLD



the authority

Country



BahamasThe Public Utilities Commission Website: http://www.pucbahamas.gov.bs

1999 Public Utilities Commission Act, 1993

Other: The PUC is required 3 months after the end of the financial year to transmit an Annual Report to the Governor General and Prime Minister

Yes: 3 Members

Contributions from operator turnover: 45.9%

Licence feesGovernment appropriation: 19.9%Spectrum fees: 34.2%

BahrainTelecommunications Regulatory Authority Website: http://www.tra.org.bh

2002 The Telecommunications Law No reporting requirements

Yes: 5 MembersLicence fees: 100%

BangladeshBangladesh Telecommunication Regulatory Commission (BTRC) Website: www.btrc.org.bd

2002 BangladeshTelecommunication Act, 2001


Yes: 5 MembersOther: 100% Govt. of Bangladesh

BarbadosFair Trading Commission Website: http://www.ftc.gov.bb

2001 Fair Trading Commission Act 2000-31

Sector Ministry (annual report)Other Ministry

Yes: 11 Members

Contributions from operator turnover

Government appropriation


REGULATORY TABLES


136

WORLD



the authority

Country



Belgium2

Belgian Institute for Postal Services and Telecommunications Website: http://www.bipt.be1993 Act of 21 March 1991 on the

reform of some economic public companies. Act of 17 January 2003 concerning the statute of the Belgium Post and Telecommunication regulator

LegislatureSector Ministry (annual report)Other Ministry

Yes: 4 Members

Other: 4% Office of the ombudsperson

Award/auction of mobile licence: 2%Licence fees: 7%Spectrum fees: 81%Numbering fees: 6%

Belize2

Public Utilities Commission (PUC)1988

Benin2

Direction de la Politique des Postes et Télécommunications (interim)2002 Ordonnance n° 2002- 003 du

31 janvier 2002 Sector Ministry (annual report)

Yes: 5 MembersOther: The Authority is not yet operational.

BhutanBhutan Communications Authority Website: http://www.bca.gov.bt

2000 Bhutan Telecommuncations Act 1999

Sector Ministry (annual report)

No: DirectorGovernment appropriation: 100%

Bolivia2

Superintendencia de Telecomunicaciones Website: http://www.sittel.gov.bo1995 Ley 1600 (Ley SIRESE) No:

TelecommunicactionSuperintendent

Spectrum fees: 100%




WORLD



the authority

Country



Bosnia and Herzegovina2

Communications Regulatory Agency (CRA) Website: http://www.cra.ba1999 Decision combining the

competencies of the Independent Media Commission and the TelecommunicationsRegulatory Agency issued by Office of the High Representative (OHR) in 2001

Other: The Council of Ministers

No: Chief Executive Officer (CEO)

Licence fees: 70%Spectrum fees: 7%Numbering fees: 23%

BotswanaBotswana Telecommunications Authority (BTA) Website: http://www.bta.org.bw

1996 Telecommunication Act 1996 (No 15 of 1996)


Yes: 5 MembersLicence fees: 91%Spectrum fees: 9%

Brazil2

Agência Nacional de Telecomunicações do Brasil (Anatel) Website: http://www.anatel.gov.br 1997 Law Nº 9.9.472 of July 16,

1997No reporting requirements

Yes: 5 Members

Financial income: 17.1%Other: 0.1% Management services, homologation

Award/auction of mobile licence: 48%Award/auction of other licence: 8.6%Spectrum fees: 0.9%Regulatory fees: 25.2%Fines/penalties: 0.62%

Brunei Darussalam1

Authority for Info-communications Technology Industry of Brunei Darussalam (AiTi)Website: http://www.aiti.gov.bn

2003 Authority for info-communications technology industry of Brunei Darussalam Order, 2001

Other: Report to Minister of Communications

Yes: 5 MembersConfidential


REGULATORY TABLES


138

WORLD



the authority

Country



BulgariaCommunications Regulation Commission Website: http://www.crc.bg

2002 Telecommunications Law Other: To the National Assembly, the President of the Republic, the Council of Ministers, the Council for Electronic Media.

Yes: 5 Members

Contributions from operator turnover: 15,3%

Award/auction of mobile licence: 57,6%Licence fees: 1,6%Spectrum fees: 22,7%Numbering fees: 2,6%Fines/penalties: 0,2%

Burkina Faso1

Autorité Nationale de Régulation des Télécommunications Website: http://www.artel.bf1998 Loi n° 051/98/AN du 04

décembre 1998Sector Ministry (annual report)

No: Director General

Contributions from operator turnover: 1%Other: 4% Approval and authorization fees

Award/auction of other licenceLicence feesSpectrum fees: 45%Numbering fees: 15%Regulatory fees: 35%Fines/penalties: 1%

Burundi2

Agence de Régulation et de Contrôle des Télécommunications 1997 Décret Présidentiel n° 100/182 Sector Ministry

(annual report)No:Administrator/Director-General




WORLD



the authority

Country



Cameroon1

Agence de Régulation des Télécommunications1998 Loi n° 98/014 du 14 juillet 1998 Other: Submits an

annual public report on activities


Contributions from operator turnover

Licence feesSpectrum feesNumbering feesRegulatory feesFines/penalties

CanadaCanadian Radio-television and Telecommunications Commission (CRTC) Website: http://www.crtc.gc.ca

1976 Telecommunications Act and CRTC Act


Yes: 7 Members

Other: Telecom activities are funded by fees. Cable activities are funded by entities regulated under the Broadcasting act.

Regulatory fees: 100%

Cape VerdeInstituto das Comunicações e das Técnologias de Informação (ICTI) Website: http//www.icti.cv

2004 Résolution nº 1/2004 du 19 janvier 2004


Yes: 3 Members

Central African Rep.1

Agence chargée de la Régulation des Télécommunications (ART)

Chad1

Office Tchadien de Régulation des Télécoms (OTRT) Website: http://www.otrt.td1998 Loi 009/PR/98 du 17 août

1998Sector Ministry (annual report)Other: Chairman of the Board of Directors

Contributions from operator turnover: 13,7%Other: 0,5% Transfers, fixed assets

Spectrum fees: 51,6%Regulatory fees: 34,5%


REGULATORY TABLES


140

WORLD



the authority

Country



ChileSubsecretaría de Telecomunicaciones Website: http://www.subtel.cl

1977 Decreto Ley Nº 1.762 LegislatureSector Ministry (annual report)Other Ministry

No: Subsecretariat of telecommunications

Other: 54% Public Treasury

Award/auction of mobile licence: 0.5%Spectrum fees: 45%Fines/penalties: 0.5%

ColombiaComisión de Regulación de Telecomunicaciones (CRT) Website: http://www.crt.gov.co

1994 Ley 142 de 1994, Decreto 1130 de 1999

LegislatureSector Ministry (annual report)Other Ministry

Yes: 3 MembersContributions from operator turnover: 100%Financial income

Congo (Dem. Rep.)1

Autorité de Régulation de la Poste et des Télécommunications du Congo ( A.R.P.T.C.)Website: http://www.arptc.cd/

2002 Loi n° 014/2002 du 16 octobre 2002

Other: President of the Republic

Yes: 7 Members

Other: Regulatory fee

Numbering fees

Costa RicaAutoridad Reguladora de los Servicios Públicos (ARESEP) Website: http://www.aresep.go.cr

1963 Ley 258 de 1941; Ley 3226 de 1963: facultades regulatorias; Ley 7593 del año 1996 se transformó en la ARESEP

LegislatureOther: Annual activities report

Yes: 5 MembersOther: 100% Regulatory fees covered directly by the operator




WORLD



the authority

Country



Côte d'Ivoire1

Agence des télécommunications de Côte d’Ivoire (ATCI) Website: http://www.atci.ci1995 Loi n° 95-526 du 07 juillet

1995 portant Code des Télécommunications

Sector Ministry (annual report)Other: Ministry of the Economy and Finance

Yes: 10 Members

Contributions from operator turnover: 11.96%Other: 4.61% Various

Award/auction of mobile licence: 8.74%Spectrum fees: 69.7%Numbering fees: 4.85%Fines/penalties: 0.08%

Croatia1

Croatian Institute of Telecommunications Website: http://www.telekom.hr2000 Law on Telecommunications Other: Annual

report to the Government

Yes: 7 Members

Contributions from operator turnover: 25%Other: 1% technical examinations and additional incomes

Spectrum fees: 50%Numbering fees: 24%

CyprusOffice of the Commissioner of Elecronic Communications & Postal RegulationWebsite: http://www.ocecpr.org.cy

2002 The Telecommunications and Postal Services Regulation Law of 2002

Other: Annual Report to the Head of State

No: Commissioner of ElectronicCommunications and Postal Regulation

Financial income: 3.4 %Other: 0.02% Postal licence fees

Licence fees: 75.2%Government appropriation: 8.3%Numbering fees: 8.5%Fines/penalties: 0.5 %

Czech RepublicCzech Telecommunication Office Website: http://www.ctu.cz

2000 Act on Electronic Communications and on Amendments to Other Acts No. 127/2005 Coll.,

Other: Annual Report is presented to the Governmentand the Parliament

Yes: 5 MembersGovernment appropriation: 100%


REGULATORY TABLES


142

WORLD



the authority

Country



DenmarkNational IT and Telecom Agency (NITA) Website: http://www.itst.dk/

1991 Finance Act 2002 Sector Ministry (annual report)Other: Annual Status Report. Annual activity account on financialcircumstances and performance.

No: Director-General

Other: 35% Various services

Government appropriation: 65%

Dominican Rep.1

INDOTEL Website: http://www.indotel.org.do1998 Ley General de

TelecomunicationcesNo. 153-98

Other: Annual report to the Executive for presentation to the National Congress of the Dominican Republic

Yes: 5 MembersContributions from operator turnoverOther: In addition, revenues are received corresponding to the use of the radio spectrum public domain; the fees established, as appropriate, under the procedures for the issue of concessions and licences, in accordance with the regulations; from the returns generated by its own capital resources; from budgetary appropriations assigned to it, as appropriate, by the central Government; and from any other possible sources.




WORLD



the authority

Country



EcuadorConsejo Nacional de Telecomunicaciones (CONATEL) Website: http://www.conatel.gov.ec

1995 Ley Reformatoria a la Ley Especial de Telecomunicaciones. R.O. 770 del 30 de agosto de 1995

No reporting requirements

Yes: 7 MembersAward/auction of mobile licence: 10%Award/auction of other licence: 10%Licence fees: 10%Spectrum fees: 70%

EgyptNational Telecommunication Regulatory Authority (NTRA) Website: http://www.tra.gov.eg/

1998 Presidential Decree Sector Ministry (annual report)Other: NTRA Board of Directors

No: Executive President

Financial income

Licence feesSpectrum feesRegulatory feesFines/penalties

El SalvadorSIGET Website: http://www.siget.gob.sv

1996 Ley de creación de la SIGET Legislature Yes: 3 Members

Financial income: 3.6%Other: 22.75% Accumulated surpluses, electricity charges and miscellaneous charges

Award/auction of other licence: 1.54%Licence fees: 5.54%Spectrum fees: 66.03%Numbering fees: 0.54%

EritreaCommunications Department Website: http://www.cd.gov.er

1998 The Communications Proclamation No. 102/1998

No: Director GeneralGovernment appropriation: 100%


REGULATORY TABLES


144

WORLD



the authority

Country



EstoniaEstonian National Communications Board Website: http://www.sa.ee

1998 Government of the Republic Act


No: Director GeneralOther: 100% State Budget

Ethiopia1

Ethiopian Telecommunications Agency Website: http://www.telecom.net.et/~eta/ 1996 Telecommunication

proclamation No. 49/1996LegislatureSector Ministry (annual report)

No: General ManagerGovernment appropriation: 100%

Finland1

Finnish Communications Regulatory Authority Website: http://www.ficora.fi 1988 Act on Telecommunications

AdministrationOther: Half-yearly financial report to the sector Ministry with evaluation of achievement of targets


Other: 47% refund for collection of television fees 34%, domain name fees 9%, other fees 4 %


FranceAutorité de Régulation des Communications Electroniques et des Postes (ARCEP)Website: http://www.art-telecom.fr/

1997 Loi de réglementation des télécommunications n° 96-659 du 26 juillet 1996

LegislatureSector Ministry (annual report)Other: Annual report to president of France





WORLD



the authority

Country



GabonAgence de Régulation des Télécommunications (ARTEL) Website: http://www.artel.ga

2001 Loi N°005/2001 Sector Ministry (annual report)Other MinistryOther: ARTEL also reports to finance ministry. It may also report to audit office for financial matters

Yes: 6 Members

Contributions from operator turnover: 4%The Administrative and Financial Directorate of ARTEL.

Licence fees: 70%Government appropriation: 17%Spectrum fees: 12%

GambiaPublic Utility Regulatory Authority (PURA) Website: http://www.pura.gm/

2004 PURA Act 2001 LegislatureOther: Secretary of State for Finance and Economic Affairs, who is responsible for the Administration of the Act. He in turn will submit it to the National Assembly


GeorgiaGeorgian National Communications Commission Website: http://www.gncc.ge

2000 Georgian Law on Telecommunications and Post

Other: Annual Report to the President of Georgia

Yes: 3 Members


REGULATORY TABLES


146

WORLD



the authority

Country



GermanyFederal Network Agency for Electricity, Gas, Telecommunication, Post and Railway (Section 115)Website: http://www.bundesnetzagentur.de

1998 Telecommunications Act of 25 July 1996

LegislatureOther: Annual Report [to the interested public; the reporting to Legislature is made by the bi-annual Activity Report (cf. Sections 121 and 122 of the Telecommunications Act)].

No: President

Other: 9.3% amateur radio, digital signature, radio equipment and telecommunications terminal equipment, operator's certificate in the aeronautical mobile service, measurements for third parties (i.e. environmental impact measurements, Leeheim Satellite Monitoring station, etc.) and post (postal licenses)

Government appropriation: 47%Spectrum fees: 32%Numbering fees: 13%Fines/penalties: 0.7%

Ghana1

National Communications Authority Website: www.nca.org.gh (this is under construction)1997 National Communications

Authority Act 524 of 1996Sector Ministry (annual report)

Yes: 7 Members

Financial income: 2%

Licence fees: 16%Spectrum fees: 46%Regulatory fees: 31%Fines/penalties: 5%

GreeceNational Telecommunications and Post Commission, Greece (EETT) Website: http://www.eett.gr

1992 Law 2075/1992 Sector Ministry (annual report)

Yes: 9 Members

Contributions from operator turnover: 12.34%Financial income: 0.64%Other: 72.87% EETT's reserves from previous years

Licence fees: 0.48%Spectrum fees: 8.55%Numbering fees: 4.71%Fines/penalties: 0.41%




WORLD



the authority

Country



Grenada2

National Telecommunications Regulatory Commission Website: http://www.ectel.int/grd/2001 Yes: 5 Members

Guatemala2

Superintendencia de Telecomunicaciones Website: http://www.sit.gob.gt/1996 Ley General de

Telecomunicaciones (Decreto 94-96 del Congreso de la República)

No: Superintendent

Other: 90% Fines, 3%; bank interest, 81%; administrative charges, 6%

Award/auction of mobile licence: 10%

Guinea2

Direction nationale des postes et télécommunications1992 Sector Ministry

(annual report)No: National Director of Posts and Telecommunications

Other: The Regulatory Authority (DNPT) is financed out of the Public Treasury.

Guinea-Bissau1

Institut des Communications de la Guinée-Bissau (ICGB) Website: http://www.icgb.org1999 Décret-loi nº 3/99, 20 août Sector Ministry

(annual report)Other Ministry

Yes: 3 MembersLicence fees: 3%Government appropriation: 14%Spectrum fees: 62%Regulatory fees: 21%

Guyana2

Public Utilities Commission


REGULATORY TABLES


148

WORLD



the authority

Country



HaitiConseil National des Telecommunications (CONATEL) Website: N/A

1969 Décret-loi de création du CONATEL 27 Septembre 1969



Contributions from operator turnoverOther: Type approval

Licence feesSpectrum fees

Honduras2

Comisión Nacional de Telecomunicaciones Website: http://www.conatel.hn1996 Ley Marco del Sector de

TelecomunicacionesYes: 5 Members3 principals, 2 substitutes


HungaryNational Communications Authority Website: http://www.nhh.hu

1999 The present structure is created by Act C of 2003 on Electronic Communications

LegislatureOther: Annual Report to the Government

Yes: 7 Members

Contributions from operator turnover: 13%Other: 18% V.A.T. reimbursed


Iceland2

Post and Telecom Administration Website: http://www.pta.is1997 Law on the Post and Telecom

Administration, no. 147/1996No: Managing Director

Other: 35% Operating fees

Award/auction of mobile licence: 2%Licence fees: 19%Spectrum fees: 44%

IndiaTelecom Regulatory Authority of India Website: http://www.trai.gov.in

1997 Telecom Regulatory Authority of India Act-1997

LegislatureOther: Annual Report to the Parliament

No: ChairpersonGovernment appropriation: 100%




WORLD



the authority

Country



Indonesia1

Directorate General of Posts and Telecommunications Website: http://www.postel.go.id2003 Minister of Communications'

Decree No.31 regarding Establishment Indonesian TelecommunicationRegulatory Body



Iran (I.R.)Communications Regulatory Authorithy Website: http://www.cra.ir

2003 The law of duties and powers of the Ministry of ICT, 2003


Yes: 7 MembersOther: 100% At present from the Government Fund. The following possible revenues are transferred to the Government account: License fees, Spectrum fees, Numbering fees,

IrelandCommission for Communications Regulation (ComReg) Website: http://www.comreg.ie

2002 Communications (Regulation) Act 2002


Yes: 3 Members

Contributions from operator turnover: 30%Financial income: 3%Other: Cable and MMDS, Wireless Telegraphy Fees etc 2%, Postal Levy 4%

Licence fees: 61%

Italy2

Autorità per le Garanzie nelle Comunicazioni (AGCOM) Website: http://www.agcom.it 1998 L. 31 luglio 1997, n. 249 Legislature Yes: 9 Members

Contributions from operator turnover: 39%Financial income: 1%



REGULATORY TABLES


150

WORLD



the authority

Country



JamaicaOffice of Utilities Regulation Website: http://www.our.org.jm

1995 Office of Utilities Regulation Act

Other: Annual report is submitted to Legislative Branch through the Minister of Development

Yes: 3 Members

Financial income: 2.36%Other: 6.63% License application processing fees (6.43%) and grants (0.2%)

Regulatory fees: 91.01%

JordanTelecommunications Regulatory Commission (TRC) Website: http://www.trc.gov.jo

1995 Telecommunication Law No. 13 of 1995 and its amendments of 2002

Other: Report to the prime minister

Yes: 5 MembersLicence fees: 82.377%Spectrum fees: 16.234%Fines/penalties: 1.389%

Kenya1

Communications Commission of Kenya Website: http://www.cck.go.ke1999 Kenya Communications Act,

1998Sector Ministry (annual report)

Yes: 11 Members


Spectrum fees: 80.7%Regulatory fees: 0.6%

Korea (Rep.)Korea Communications Commission Website: http://www.kcc.go.kr

1997 Article 37~44 of the Telecommunications Basic Act

Legislature Yes: 7 MembersGovernment appropriation: 100%




WORLD



the authority

Country



KyrgyzstanState Communications Agency Website: http://www.gas.gov.kg

1997 Decree of the President of the Kyrgyz Republic 280, 7 October 1997

Other: Annual report to the Government

No: Director

Contributions from operator turnover: 61.6%Financial income: 2.7%

Spectrum fees: 35.7%

LatviaPublic Utilities Commission Website: http://www.sprk.gov.lv

2001 Law on Regulators of Public Services

Legislature Yes: 5 MembersContributions from operator turnover: 100%

Lesotho1

Lesotho Telecommunications Authority Website: http://www.lta.org.ls2000 Lesotho Telecommunications

Authority Act 2000Sector Ministry (annual report)Other Ministry

No: Chief Executive Officer

Liechtenstein2

Office for Communications Website: http://www.ak.li1999 Telekommunikationsgesetz

(TelG) vom 20. Juni 1996, LGBl. 1996 Nr. 132, Art. 42a (Telecommunications Law of 20 June 1996, Liechtenstein Legal Gazette 1996 No. 132, Art. 42a)


No: Director Government appropriationSpectrum feesNumbering feesRegulatory fees


REGULATORY TABLES


152

WORLD



the authority

Country



LithuaniaCommunications Regulatory Authority of the Republic of Lithuania Website: http://www.rrt.lt

2000 Governmental Resolution No. 617 on setting up NRA

LegislatureOther: Annual report to the Parliament and the Government

Yes: 5 MembersOther: 100% Income for services provided and work performed (administrative charges). The mechanism of financing CRA is through the State budget, whereby all income from administrative charges collected for supervision of use of radio frequencies, numbers, etc. is deposited in the State budget and then reallocated as needed to the CRA

LuxembourgInstitut Luxembourgeois de Régulation Website: www.ilr.lu

1997 Loi modifiée du 21 mars 1997 sur les télécommunications

Other: Annual financial report to minister, who supervises institute

Yes: 3 Members

Contributions from operator turnoverFinancial income

Spectrum feesNumbering feesRegulatory fees

MadagascarOffice Malagasy d'Etudes et de Régulation des Télécommuniations (OMERT) Website: http://www.omert.mg

1997 Décret n° 97-1077 du 28 août 1997



Contributions from operator turnover: 31.41%Other: 1% Testing and verification : 0,21%; Laboratory fees : 0.02%, Other financial products : 0.78%.

Licence feesSpectrum fees: 67.58%




WORLD



the authority

Country



Malawi2

Malawi Communications Regulatory Authority (MACRA) Website: http://www.macra.org.mw 1998 Communications Act 1998 Sector Ministry

(annual report)Yes: 8 Members

Contributions from operator turnover: 38.5%Financial income: 4.5%

Licence fees: 23%Spectrum fees: 24.5%

MalaysiaMalaysian Communications and Multimedia Commission Website: http://www.mcmc.gov.my

1998 Malaysian Communications and Multimedia Commission Act 1998

Other: Report to Sector Minister, Financial Report to Sector Minister, Annual Report and IndustryPerformance Report

Yes: 9 Members

Financial income: 6%

Licence fees: 40%Spectrum fees: 54%

Maldives1

Telecommunications Authority of Maldives Website: www.tam.gov.mv2003 Sector Ministry

(annual report)Yes: 5 MembersLicence fees: 5%

MaliComité de Régulation des Télécommunications (CRT) Website: http://mali-reforme-telecom.mcmtl.com

1999 Ordonnance 99-043 du 30/9/99 (art.43). Décret 00-227 PRM du10 mai 2000 fixant les modalités de fonctionnement du CRT


Yes: 3 MembersSpectrum fees: 51%Regulatory fees: 49%


REGULATORY TABLES


154

WORLD



the authority

Country



MaltaMalta Communications Authority Website: http://www.mca.org.mt

1997 The Telecommunications (Regulation) Act as amended by the Malta Communications Authority Act.


Yes: 5 MembersLicence fees: 14%Government appropriation: 86%

Mauritania1

Autorité de Régulation Website: http://www.are.mr1999 Loi 99 -019 du 11 juillet 1999

et Loi 2000 - 018 du 25 janvier 2001

LegislatureOther: The report is annual and is addressed to the Government and Parliament

Yes: 5 Members

Contributions from operator turnover: 68%Financial income: 1%Provisional budget for financial year 2004

Spectrum fees: 26%Numbering fees: 5%

MauritiusInformation and Communication Technologies Authority Website: http://www.icta.mu

2002 Information and Communication Technologies Act


Yes: 7 Members

MexicoComisión Federal de Telecomunicaciones (COFETEL) Website: http://www.cofetel.gob.mx

1996 Decreto de creación de la Comisión Federal de Telecomunicaciones.Ley Federal de Telecomunicaciones.

Other: Annual report





WORLD



the authority

Country



MoldovaNational Regulatory Agency in Telecommunications and Informatics Website: http://www.anrti.md

2000 Government decision No. 843 of 17.08.2000

Other: Report to Government

Yes: 3 MembersLicence fees: 1.2%Numbering fees: 32.2%Regulatory fees: 66.4%

MongoliaCommunications Regulatory Commission Website: http://www.crc.gov.mn

2002 Communications law of 1995 amended in 2001

Other: Government of Mongolia

Yes: 7 Members

Other: 4.8% bank interest rate etc.

Spectrum fees: 22.4%Numbering fees: 10.9%Regulatory fees: 61.9%

MoroccoAgence Nationale de Réglementation des Télécommunications (ANRT) Website: http://www.anrt.net.ma

1997 Dahir n°1-97-162 du 7 août1997 portant promulgation de la loi 24-96 relative à la poste et aux télécommunications

Other: Director general produces annual report, submitted to prime minister and published in the "Bulletin Officiel"

No: Ag. Director-GeneralContributions from operator

turnover: 53%Other: 2% Various administrative fees related to the declaration of value-added services and equipment approval.

Spectrum fees: 45%

MozambiqueInstituto Nacional das Comunicações de Moçambique (INCM) Website: http://www.incm.gov.mz

1992 Decree 22/92 Sector Ministry (annual report)



REGULATORY TABLES


156

WORLD



the authority

Country



NamibiaNamibian Communications Commission Website: http://www.ncc.org.na

1992 Namibian Communications Commission Act, 1992


No: ChairmanGovernment appropriation: 100%

NepalNepal Telecommunications Authority Website: http://www.nta.gov.np

1998 Telecommunication Act, 1997 Sector Ministry (annual report)

Yes: 5 MembersLicence fees: 100%

NetherlandsOPTA Website: http://www.opta.nl

1997 Wet OPTA Sector Ministry (annual report)

Yes: 3 Members

New ZealandCommerce Commission Website: http://www.comcom.govt.nz

2001 Commerce Act 1986, Telecommunications Act 2001


Yes: 3 Members

Other: 36% recovered from applicants and parties


NicaraguaTELCOR Website: http://www.telcor.gob.ni

1995 Ley General de Telecomunicaciones y Servicios Postales (Ley 200)

LegislatureOther: To the Executive

No: Chief Executive Director

Financial income: 2.26%Other: 1.16% Examination fee, various services

Award/auction of other licence: 2.61%Licence fees: 65.62%Spectrum fees: 28.35%




WORLD



the authority

Country



Niger1

Autorité de Régulation Multisectorielle (ARM) Website: http://niger.arm-niger.org/2004

Nigeria2

Nigerian Communications Commission Website: http://ncc.gov.ng1992 NCC ACT No 75 of 1992 Legislature


Yes: 9 Members

Contributions from operator turnover: 12%Financial income: 13%Other: 4.5% Type Approval (3.5%), administrative fees (1%).

Licence fees: 68%

NorwayNorwegian Post and Telecommunications Authority Website: http://www.npt.no

1987 Sector Ministry (annual report)

No: Director

Other: 3% Postal regulatory fees

Licence fees: 5%Government appropriation: 8%Spectrum fees: 40%Numbering fees: 4%Regulatory fees: 40%

OmanTelecommunication Regulatory Authority Website: http://www.tra.gov.om

2002 TelecommunicationsRegulatory Act 2002

Other: Annual reporting to the Council of Ministers

Yes: 4 Members

Contributions from operator turnover: 13.2%Financial income: 0.8%Other: 0.003%

Award/auction of mobile licence: 6.6%Award/auction of other licence: 2.6%Licence fees: 13.2%Spectrum fees: 65.8%Regulatory fees: 0.6%Fines/penalties: 3.8%


REGULATORY TABLES


158

WORLD



the authority

Country



PakistanPakistan Teleccommunication Authority (PTA) Website: http://www.pta.gov.pk

1996 Pakistan Telecommunication Re-Organization Act 1996

No reporting requirements

Yes: 3 Members

Financial income: 0.09%

Award/auction of mobile licence: 92.5%Spectrum fees: 0.35%Regulatory fees: 3.99%Fines/penalties: 3.07%

Panama1

Ente Regulador de los Servicios Públicos Website: http://www.ersp.gob.pa1996 Ley N° 26 (29 de enero de

1996) por la cual se crea el Ente Regulador de los Servicios Públicos

Other: Assembly of legislators

Yes: 3 MembersContributions from operator turnover: 1%Other: 1% The fee charged, in the case of the telecommunication sector, to service-providing companies for control, monitoring and audit services: 1% of annual gross revenues for type B licensees and 0.25% of annual gross revenues for mobile cellular telephone companies.

Papua New Guinea2

PANGTEL Website: http://www.pangtel.gov.pg/1997 No: Director-GeneralLicence fees

Paraguay1

Comisión Nacional de Telecomunicaciones (CONATEL) Website: http://www.conatel.gov.py1995 Ley 642/1995 de

TelecomunicacionesLegislatureSector Ministry (annual report)Other: The Executive

Yes: 5 Members




WORLD



the authority

Country



PeruOrganismo Supervisor de Inversión Privada en Telecomunicaciones (OSIPTEL)Website: http://www.osiptel.gob.pe

1994 Decreto Legislativo 702 LegislatureOther MinistryOther: Chair of the Council of Ministers,InspectorateGeneral of the Republic, Ministry of Economics and Finance

Yes: 5 MembersContributions from operator turnover: 100%

Philippines2

National Telecommunications Commission Website: http://www.ntc.gov.ph1979 EO546 Sector Ministry

(annual report)Yes: 3 MembersGovernment appropriation: 100%

PolandOffice of Telecommunications and Post Regulation (URTIP) Website: http://www.urtip.gov.pl

2000 Law of July 2000 Sector Ministry (annual report)

No: President Other: 100% State budget in the amount fixed each year in the Annual Budget Act.


REGULATORY TABLES


160

WORLD



the authority

Country



Portugal1

National Communications Authority (ANACOM) Website: http://www.anacom.pt1989 Decree-Law No. 188/81 of 2

July. Decree-law No. 283/89 of 23 August, now revoked by Decree-law 309/2001 of 7 December

LegislatureNo reporting requirementsOther: Annual report to both Government and Parliament. The chairman of the board of administration will respond to requests for hearing addressed by the appropriate committee of the Parliament, to provide information or clarification on its activities.

Yes: 3 Members

Financial income: 0.0007%Other: 2.9% Laboratory tests, postal services, EU subsidies, extraordinary income

Licence fees: 0.0014%Spectrum fees: 97%Fines/penalties: 0.0004%

QatarSupreme Council for Communication and Information TechnologyWebsite: http://www.ict.gov.qa/en/Default.aspx

2004 Ameri Decree 36, year 2004 Other: Report to an independent board chair by His Highness the Heir Apparent of the State of Qatar

No: Secretary GeneralOther: Not determined yet

RomaniaNational Regulatory Authority of Romania (ANRC) Website: http://www.anrc.ro

2002 Government Emergency Ordinance No. 79/ 2002 on the general regulatory framework for communications, approved, with amendments and completions, by Law No. 591/2002, with subsequent amendments and completions

Other: The Government of Romania

No: PresidentContributions from operator turnover: 79%Other: 21% contributions from regulated postal services




WORLD



the authority

Country



Rwanda1

Agence Rwandaise de Régulation des Services d'Utilité Publique Website: http://www.rura.gov.rw/2001 Loi n° 39/2001 du 13/09/2001

portant Création de l'Agence Rwandaise de Régulation des Services d'Utilité Publique


Yes: 7 Members

Contributions from operator turnover: 50,7%Other: 43% World Bank project and State subsidies

Licence fees: 6,3%

SamoaMinstry of Communications and Information Technology (MCIT) (ad interim)

2005 Telecommunications Act 2005 LegislatureOther: Minister

No: Regulator

Financial income

Award/auction of mobile licenceLicence feesSpectrum feesNumbering feesRegulatory feesFines/penalties

Saudi ArabiaCommunication and Information Technology Commission Website: http://www.citc.gov.sa

2002 The Council of ministers decision No. (74) Dated 2001



Senegal1

Agence de Régulation des Télécommunications Website: http://www.art.sn2001 Other: President of

the RepublicNo: Director GeneralSpectrum fees: 100%


REGULATORY TABLES


162

WORLD



the authority

Country



SingaporeInfocomm Development Authority of Singapore Website: http://www.ida.gov.sg

1992 Telecommunication Authority of Singapore Act 1992 and subsequently superseded by the Info-communications Development Authority of Singapore Act 1999

Other: Annual Report to the Minister for Information,Communicationsand the Arts.

No: CEOOther: financed from a combination of licence fees, numbering fees and spectrum fees

Slovak RepublicTelecommunication Office Website: http://www.teleoff.gov.sk

2000 Act. No. 195 / 2000 C.l. on Telecommunications

Other: Annual report to the National Council of the Slovak republic

No: PresidentGovernment appropriation: 100%

SloveniaPost and Electronic Communications Agency (APEK) Website: http://www.apek.si

2001 Government's Decision on establishment of Telecommunication and Broadcasting Agency

Other: Annual Report to the Government and the National Assembly of the Republic of Slovenia

No: M. Sc., Acting Director

Other: 10% Postal fees

Licence fees: 6%Spectrum fees: 37%Numbering fees: 46%Fines/penalties: 1%

South Africa1

ICASA Website: http://www.icasa.org.za2000 ICASA Act 13 of 2000 Legislature






WORLD



the authority

Country



SpainComisión del Mercado de las Telecomunicaciones (CMT) Website: http://www.cmt.es

1996 Real Decreto-Ley 6/1996, de 7 de junio, de liberalización de las Telecomunicaciones. Actualmente se regula en la Ley 32/2003, General de Telecomunicaciones

LegislatureOther: To the Government, for submission to Parliament

Yes: 9 MembersContributions from operator turnover: 0.15%Other: Various telecommunication fees, from the issue of registration certificates, technical opinions and inspection or monitoring activities.

Sri Lanka1

Telecommunications Regulatory Commission of Sri Lanka Website: http://www.trc.gov.lk1991 Sri Lanka Tel. Act No. 25 of

1992 & Tel. (Amendment Act) No. 27 of 1996


Yes: 5 MembersLicence fees: 30%Spectrum fees: 51%Regulatory fees: 19%

St. LuciaNational Telecommunications Regulatory Commission Website: http://www.ntrc.org.lc

2000 Telecommunications Act 2000 Sector Ministry (annual report)

Yes: 5 MembersSpectrum fees: 100%

St. Vincent and the GrenadinesNational Telecommunications Regulatory Commission (NTRC) Website: http://www.ntrc.vc

2001 Telecommunications Act No.1 2001


Yes: 5 Members

Financial income: 1.0%Other: 2.0% Application fees and rental of conference room


SudanNational Telecommunication Corporation (NTC) Website: http://www.ntc.org.sd

1994 Telecommunication Act 2001 (including establishment of Regulatory Authority)


Yes: 10 Members


Licence fees: 33.3%Spectrum fees: 50%


REGULATORY TABLES


164

WORLD



the authority

Country



SurinameTelecommunications Authority Suriname (TAS) Website: http://www.tas.sr

1998 Decree 1998 Establishment of the Telecommunications Authority Suriname

Sector Ministry (annual report)Other: Board of Directors

No: Managing DirectorGovernment appropriation: 100%

SwedenNational Post and Telecom Agency Website: http://www.pts.se/

1992 Förordning med instruktion för telestyrelsen (Decree)



Other: 18%

Licence fees: 41%Government appropriation: 41%

SwitzerlandOffice Fédéral de la Communication (OFCOM) Website: http://www.ofcom.ch

1992 Loi du 30 avril 1997 sur les télécommunications (LTC)


No: DirectorGovernment appropriation: 100%

Tanzania1

Tanzania Communications Regulatory Authority Website: http://www.tcc.go.tz1994 Tanzania Communications

Act No. 18 of 1993, Tanzania Communications Regulatory Authority Act No. 12 of 2003 (merger of TCC and TBC in 2003)


Yes: 7 Members

Contributions from operator turnover: 52.39%Other: 2.08% Applications forms and Type approval fee

Licence fees: 2.37%Spectrum fees: 36.54%Numbering fees: 4.25%Fines/penalties: 0.05%




WORLD



the authority

Country



TFYR MacedoniaAgency for electronic communications Website: http://www.aec.mk

2005 Law on electronic communications

Other: Annual report to the Parliament

Yes: 5 MembersLicence fees: 8,5%Spectrum fees: 58,5%Numbering fees: 11,6%Regulatory fees: 21.4%

Thailand1

National Telecommunications Commission (NTC) Website: http://www.ntc.or.th2004 The Act on Organizations to

Assign Radio Frequency Sprectrum and to Regulate the Sound Broadcasting, Television Broadcasting and Telecommunication Services, B.E.2543

Other: Report to the Cabinet

Yes: 7 Members

TogoAutorité de Réglementation des Secteurs de Postes et Télécommunications (ART&P)Website: http://www.artp.tg

1998 Loi n° 98-005 du 11 février 1998 en son article 57


Yes: 5 Members

Trinidad and TobagoTelecommunications Authority of Trinidad and Tobago Website: http://www.tatt.org.tt

2004 Telecommunications Act 2001 (as amended by Telecommunications(Amendment) Act 2004




REGULATORY TABLES


166

WORLD



the authority

Country



TunisiaInstance Nationale des Télécommunications Website: http://www.intt.tn/

2001 Loi n° 2001-1 du 15 janvier 2001, portant promulgation du code des télécommunications


Yes: 7 MembersNumbering fees: 100%

TurkeyTelecommunications Authority Website: http://www.tk.gov.tr

2000 Amending Law No. 4502 Other: Turkish National Assembly

Yes: 7 Members

Contributions from operator turnover: 9%Other: 3% caution returns; form, book, publication, hologram sales returns, etc.

Spectrum fees: 88%

UgandaUganda Communications Commission Website: http://www.ucc.co.ug

1997 The Communications Act 1997 LegislatureSector Ministry (annual report)

Yes: 7 Members

Contributions from operator turnover: 33%Financial income: 1.1%Other: 14% Rent

Licence fees: 4%Spectrum fees: 47%

United Arab EmiratesTelecommunications Regulatory Authority Website: http://www.tra.ae

2004 Federal Law by decree No. (3) of 2003

Other: UAE Telecom Supreme Committee

Yes: 5 MembersOther: 100% UAE Federal Governmet




WORLD



the authority

Country



United KingdomOffice of Communications (OFCOM) Website: http://www.ofcom.org.uk

1984 Telecommunications Act 1984; 2003 Communications Act

LegislatureSector Ministry (annual report)Other: Select Committees of Legislature (Public AccountsCommittee, Trade & Industry Select Committee). The annual reports to the sector Ministry is also required to be provided to the Legislature by the Minister.

No: Chief Executive

Contributions from operator turnover: 15%Other: 5% Government department (DCMS and DTI), commercial property sub-lets and broadcasting application fees and commercial interference income

Licence fees: 16%Government appropriation: 64%

United States1

Federal Communications Commission Website: http://www.fcc.gov/1934 The Communications Act of

1934LegislatureOther: It is independent with Congressionaloversight and budget control

Yes: 5 Members

Other: Auctions receipts are deposited with Treasury. Funds are allocated from receipts as needed to cover the cost of running the Auctions Program.

Government appropriation: 3.6%Regulatory fees: 96.4%

Uruguay2

Unidad Reguladora de Servicios de Comunicaciones (URSEC) Website: http://www.ursec.gub.uy2001 Ley N° 17.296 del 21 de

febrero de 2001Other: The reports are sent to whoever asks for them. The Executive Authority must receive reports on matters falling within its competence.

Yes: 3 MembersSpectrum fees: 78%Regulatory fees: 21%Fines/penalties: 1%


REGULATORY TABLES


168

WORLD



the authority

Country



VenezuelaComisión Nacional de Telecomunicaciones (CONATEL) Website: http://www.conatel.gov.ve

1991 Decreto n° 1826 del 5 de septiembre de 1991


Yes: 8 Members

Other: 82.2% Special contribution


ZambiaCommunications Authority Website: http://www.caz.gov.zm

1994 Telecommunications Act of 1994 Cap 469


Yes: 8 Members

Contributions from operator turnover: 77.4%Financial income: 1.9%Other: 0.8% Rentals

Licence fees: 6%Spectrum fees: 13.9%

ZimbabwePostal and Telecommunications Regulatory Authority of Zimbabwe (POTRAZ)Website: http://www.potraz.gov.zw/

2000 Postal and Telecommunications Act 2000, Chapter 12:05


Yes: 7 Members

Contributions from operator turnover: 47%Financial income: 13%Other: 3% Postal

Licence fees: 17%Spectrum fees: 5%Regulatory fees: 15%




WORLDName of the operator

2. Status of the main fixed-line operatorsStatus

Afghanistan State-owned - Corporatized The government intends to privatize the operator (Year: 2006)

Afghan Telecom

Albania State-owned - Corporatized In the process of privatizingAlbtelecom sh.a

Algeria State-owned - Corporatized The government intends to privatize the operator(Year: 2006)

Algérie Telecom

Andorra2 State-owned - CorporatizedServei de Telecomunicacions d’Andorra (STA)

Angola1 State-owned In the process of privatizingAngola Telecom

Mercury Partially privatized

Mundo Startel Fully privatized

Nexus Fully privatized

Wezacom Fully privatized


REGULATORY TABLES


170



Antigua and Barbuda State-ownedAntigua Public Utilities Authority Telephones

Argentina2

Telecom Fully privatized Date of privatization (first phase): 1990

Telefónica Fully privatized Date of privatization (first phase): 1990

ArmeniaArmenTel JV Partially privatized: 90% Date of privatization (first

phase): 1998: 90% sold

AustraliaTelstra Partially privatized: 51.8% Date of privatization (first

phase): 1997: 33.3% sold

Austria1

Telekom Austria Partially privatized Date of privatization (first phase): 1998: 25% sold






Azerbaijan State-owned The government intends to privatize the operator (Year: 2008)

AzTelekom

State-owned The government intends to privatize the operator (Year: 2008)

Baktelekom

Bahamas State-owned - Corporatized In the process of privatizingBahamas Telecommunications Company Ltd

BahrainBatelco Partially privatized: 64%

Bangladesh State-owned The government intends to privatize the operator

Bangladesh Telegraph and Telephone Board (DTTB)

BarbadosCable & Wireless (Barbados) Ltd Partially privatized Date of privatization (first



REGULATORY TABLES


172



Belarus State-owned - Corporatized No intention to privatize at present

Republican Unitary Enterprise

Belgium2

Belgacom Partially privatized: 49,9% Date of privatization (first phase): 1996: 49,9% sold

Belize2

Belize Telecommunications Ltd Partially privatized: 96.5% Date of privatization (first phase): 1996: 95% sold

Benin2 State-owned The government intends to privatize the operator (Year: 2002/2003)

Office des postes et télécommunications (OPT)

Bhutan State-owned - Corporatized No intention to privatize at present

Bhutan Telecom

Bolivia2

ENTEL S.A. Fully privatized Date of privatization (first phase): 1995: 100% sold






Bosnia and Herzegovina

2 State-owned - Corporatized The government intends to privatize the operator (Year: 2004)

BH Telecom

HT Mostar Partially privatized: 40%

Telekom Srpske Partially privatized: 20%

Botswana State-owned - Corporatized The government intends to privatize the operator

Botswana Telecommunications Corporation

Brazil2

Empresa Brasileira de Telecomunicações S.A. (Embratel)

Fully privatized Date of privatization (first phase): 1998

Brasil Telecom Fully privatized Date of privatization (first phase): 1998

CTBC - Companhia Telefônica Brasil Central

Fully privatized Date of privatization (first phase): 1998

Telemar Fully privatized Date of privatization (first phase): 1998

Brunei Darussalam1 State-ownedJTB

BulgariaBulgarian Telecomunication Company BTC

Partially privatized: 65% Date of privatization (first phase): 2004: 65% sold


REGULATORY TABLES


174



Burkina Faso1 State-owned - Corporatized In the process of privatizingONATEL

Burundi2 State-owned - Corporatized In the process of privatizingOffice National des Télécommunications

Cambodia2 State-owned The government intends to privatize the operator

Ministry of Posts and Telecommunications

Cameroon1 State-owned - Corporatized In the process of privatizingCAMTEL

Canada State-owned No intention to privatize at present

SaskTel

Bell Canada Fully privatized

TELUS Fully privatized Date of privatization (first phase): 1990: 100% sold

MTS Fully privatized Date of privatization (first phase): 1994: 100% sold

Aliant Fully privatized






Cape VerdeCVTelecom Partially privatized: 96.6% Date of privatization (first

phase): 1995: 53.7% sold

Central African Rep.2

Socatel Partially privatized: 40%

Chad1 State-owned - Corporatized In the process of privatizingSOTEL TCHAD

ChileCTC Fully privatized

VTR Fully privatized

EntelPhone Fully privatized

Telefónica del Sur Fully privatized

ChinaChina Telecom Partially privatized Date of privatization (first

phase): 2002

China Unicom Partially privatized Date of privatization (first phase): 2000

China Netcom Partially privatized Date of privatization (first phase): 2004

China Mobile Partially privatized Date of privatization (first phase): 1997


REGULATORY TABLES


176



Colombia State-owned - Corporatized No intention to privatize at present

Colombia Telecomunicaciones

ETB Partially privatized: 60%

EPM Partially privatized: 95%

UNITEL Fully privatized

ETELL Partially privatized: 93%

Comoros2 State-owned - Corporatized No intention to privatize at present

SNPT

Congo2 State-owned - Corporatized (Year: 1965)

The government intends to privatize the operator

ONPT

Congo1 State-owned - Corporatized In the process of privatizingOCPT

State-owned - Corporatized In the process of privatizingRENATELSAT

Costa Rica State-owned No intention to privatize at present

Instituto Costarricense de Electricidad


(Dem. Rep. of)





Côte d'Ivoire1

Côte d'Ivoire Télécom Partially privatized: 51% Date of privatization (first phase): 1997: 51% sold

Croatia1

Croatian Telecom Inc. (HT) Partially privatized: 51% Date of privatization (first phase): 1999: 35% sold

CubaETECSA Fully privatized

Cyprus State-owned - Corporatized No intention to privatize at present

Cyprus Telecommunications Authority (CYTA)

Czech RepublicCESKÝ TELECOM, a.s. Fully privatized Date of privatization (first


2 State-ownedMinistry of Posts and Telecommunications


D. P. R. Korea

REGULATORY TABLES


178



DenmarkTDC Fully privatized Date of privatization (first


Djibouti2 State-owned - Corporatized (Year: 1957)

The government intends to privatize the operator

OPT

Dominica2

Telecommunications of Dominica Partially privatized: 80%

Dominican Rep.1

VERIZON Dominicana Fully privatized

TRICOM, S.A. Fully privatized

ALL America Cables & Radio, Inc. (AAC&R- Centennial Dominicana)

Fully privatized

Ecuador State-owned - Corporatized No intention to privatize at present

Andinatel S.A.

State-owned - Corporatized No intention to privatize at present

Pacifictel S.A.

Egypt State-owned - Corporatized In the process of privatizingTelecom Egypt






El SalvadorCTE Partially privatized: 95.2% Date of privatization (first


TELEFONICA Partially privatized: 51% Date of privatization (first phase): 1998: 51% sold

TELEMOVIL Fully privatized

GCA Fully privatized

Equatorial Guinea2

Sociedad Anónima de Telecomunicaciones de la República de Guinea Ecuatorial (Getesa)


Eritrea State-owned - Corporatized The government intends to privatize the operator

ERITEL

EstoniaElion Enterprises Partially privatized: 72,83% Date of privatization (first


Ethiopia1 State-owned - CorporatizedEthiopian Telecommunication Corporation

Fiji2

Telecom Fiji Limited Partially privatized


REGULATORY TABLES


180



Finland1

TeliaSonera AB Partially privatized: 19,07% Date of privatization (first phase): 1998

ElisaCom Oy Fully privatized

FranceFrance Télécom Partially privatized: 67% Date of privatization (first


Gabon State-owned In the process of privatizingGABON TELECOM

Gambia State-owned - CorporatizedGAMTEL

Georgia State-owned - Corporatized The government intends to privatize the operator (Year: 2004)

GEC

New Nets Fully privatized

Germany01051 Telecom GmbH Fully privatized

Arcor AG & Co. Partially privatized: ~82%

Colt Telecom GmbH Fully privatized

Deutsche Telekom AG Partially privatized: ~62.5% Date of privatization (first phase): 1996: ~26% sold






Ghana1

Ghana Telecom Partially privatized: 30% Date of privatization (first phase): 1997: 30% sold

WESTEL Fully privatized

GreeceOTE (Hellenic Telecommunications Organisation S.A )

Partially privatized: 62% Date of privatization (first phase): 1997: 7.5% sold

Tellas Partially privatized: 51%

Forthnet Partially privatized: 78.94%

Grenada2

Grenada Telecommunications (GRENTEL)

Partially privatized: 70%

Guatemala2

Empresa Telecomunicaciones de Guatemala S.A.

Fully privatized Date of privatization (first phase): 1997: 100% sold

Guinea2

Sotelgui Partially privatized: 60% Date of privatization (first phase): 1995: 60% sold

Guinea-Bissau1

Guiné Telecom Partially privatized: 51% Date of privatization (first phase): 1989: 51% sold


REGULATORY TABLES


182



Guyana2

Guyana Telephone and Telegraph Ltd (GT & T)

Partially privatized: 80% Date of privatization (first phase): 1991

HaitiTéléco Partially privatized: 3%

Haitel Fully privatized

Honduras2 State-owned - Corporatized In the process of privatizingEmpresa Hondureña de Telecomunicaciones (HONDUTEL)

HungaryT-Com Fully privatized Date of privatization (first

phase): 1993: 33.2% sold

Invitel Fully privatized

Hungarotel Fully privatized

Emitel Fully privatized

Iceland2

Iceland Telecom Ltd Partially privatized: 5.2% Date of privatization (first phase): 2001: 6% sold

IndiaBSNL Partially privatized

MTNL Partially privatized

Reliance Fully privatized

Bharti (Airtel) Fully privatized






Indonesia1

PT. Telkom Partially privatized: 48.8%

PT. Indosat Partially privatized: 85%

State-owned - Corporatized In the process of privatizingTCI

Iraq2 State-ownedIraqi Telecommunications and Posts

IrelandEircom Fully privatized Date of privatization (first


BT Ireland Fully privatized

Israel2

Bezeq Partially privatized: 46% Date of privatization (first phase): 1990: 9% sold

Italy2

Telecom Italia Partially privatized: 96.5% Date of privatization (first phase): 1998: 55% sold


Iran (I.R.)

REGULATORY TABLES


184



JamaicaCable and Wireless Partially privatized: 82%

JapanNTT EAST Fully privatized Date of privatization (first

phase): 1999: 100% sold

NTT WEST Fully privatized Date of privatization (first phase): 1999: 100% sold

JordanJordan Telecom (JT) Partially privatized: 40% Date of privatization (first


Kazakhstan2

Kazakhtelecom Partially privatized: 50% Date of privatization (first phase): 1994

Kenya1 State-owned - Corporatized The government intends to privatize the operator (Year: 2006)

Telkom Kenya Limited

Kiribati2

Telecom Services Kiribati Limited Partially privatized: 49%






Korea (Rep.)Korea Telecom Fully privatized

Kuwait State-owned No intention to privatize at present

Ministry of Communications (MOC)

KyrgyzstanKyrgyztelecom Partially privatized: 11%

Saima Telecom Fully privatized

Winline Fully privatized

Instrumentaalshik Fully privatized

Lao P.D.R. State-ownedEnterprise of Telecom Lao

Lao Telecom Co. Ltd Partially privatized: 49%

Latvia State-owned No intention to privatize at present

Latvian Railway (LDz)

State-owned No intention to privatize at present

Latvenergo

Lattelekom Partially privatized: 49% Date of privatization (first phase): 1994: 49% sold


REGULATORY TABLES


186



Lebanon2 State-owned The government intends to privatize the operator

Ministry of Telecommunications

Lesotho1

Telecom Lesotho Partially privatized: 70% Date of privatization (first phase): 2000

Liberia2 State-owned - CorporatizedLiberia Telecommunications Corporation

2 State-ownedGeneral Post and Telecommunication Company (GPTC)

Liechtenstein2 State-owned No intention to privatize at present

LTN Lie. TeleNet


Telecom FL

LithuaniaAB Lietuvos Telekomas Partially privatized: 60% Date of privatization (first



Libya





Luxembourg State-owned - Corporatized No intention to privatize at present

Enterprises of the P&T (EPT)

See www.ilr.lu

MadagascarTELECOM MALAGASY Partially privatized: 34% Date of privatization (first


Malawi2 State-owned - Corporatized In the process of privatizingMalawi Telecommunications Ltd

MalaysiaTelekom Malaysia Partially privatized Date of privatization (first

phase): 1990

Maldives2

Dhiraagu Partially privatized: 45% Date of privatization (first phase): 1988

Mali State-owned - Corporatized The government intends to privatize the operator (Year: 2006)

SOTELMA

IKATEL Fully privatized


REGULATORY TABLES


188



MaltaMaltacom plc Partially privatized: 40% Date of privatization (first


Marshall Islands2

National Telecommunications Authority

Partially privatized

Mauritania1

Mauritel S.A Partially privatized: 54% Date of privatization (first phase): 1999: 54% sold

MauritiusMauritius Telecom Ltd Partially privatized: 40% Date of privatization (first


MexicoTeléfonos de México Fully privatized Date of privatization (first

phase): 1991: 100% sold

Micronesia2 State-owned - Corporatized (Year: 1981)

FSM Telecommunications Corporation






Moldova State-owned - Corporatized No intention to privatize at present

Moldtelecom S.A.

Monaco2

MONACO TELECOM Fully privatized Date of privatization (first phase): 1999: 51% sold

MongoliaMongolian Telecommunications Company


MoroccoIttisalat Al-Maghrib Partially privatized: 51% Date of privatization (first


MozambiqueTDM Partially privatized: 20% Date of privatization (first


Myanmar State-ownedMyanma Posts and Telecommunications

State-owned - CorporatizedBagan Cybertech


REGULATORY TABLES


190



Namibia State-owned No intention to privatize at present

Telecom Namibia

Nauru2 State-ownedDirectorate of Telecommunications

Nepal State-owned The government intends to privatize the operator (Year: After 2006)

Nepal Telecom

NetherlandsKPN Partially privatized Date of privatization (first


New ZealandTelecom Corporation of New Zealand

Fully privatized Date of privatization (first phase): 1990: 100% sold

NicaraguaENITEL Fully privatized Date of privatization (first







Niger2

SONITEL Partially privatized: 51% Date of privatization (first phase): 2001: 51% sold

Nigeria2 State-owned - Corporatized In the process of privatizingNITEL

NorwayTelenor Partially privatized: 46%

Oman State-owned - Corporatized In the process of privatizingOmantel

PakistanPakistan Telecom Company Limited (PTCL)


Panama1

Cable & Wireless Panama Partially privatized: 49% Date of privatization (first phase): 1996: 49% sold


REGULATORY TABLES


192



Papua New Guinea2 State-owned The government intends to privatize the operator (Year: 2003)

Telikom PNG Limited

Paraguay1 State-owned - CorporatizedCOPACO

PeruTelefónica del Perú S.A. Fully privatized Date of privatization (first

phase): May-1994: 35% sold

Telmex Fully privatized

BellSouth Perú S.A. Fully privatized

Americatel Perú S.A. Fully privatized

Philippines2

All fixed-line operators Fully privatized

PolandTelekomunikacja Poska Partially privatized: 96.13%

Portugal1

PT Comunicações, S.A. Partially privatized: 95.3% Date of privatization (first phase): 1995: 27.26% sold






QatarQ-Tel Partially privatized Date of privatization (first


RomaniaS.C. Romtelecom S.A. Partially privatized: 54% Date of privatization (first


Russia2

Svyazinvest is the government controlled telecommunications holding company (an umbrella organization).


Rwanda1

Rwandatel Partially privatized: 1%

S. Tomé and Principe

2

Companhia Santomense de Telecomunicações, SARL

Partially privatized: 51%

Samoa State-owned - Corporatized The government intends to privatize the operator (Year: 2006)

SamoaTel


REGULATORY TABLES


194



San Marino2

Telecom Italia Partially privatized

Saudi ArabiaSaudi Telecom Partially privatized Date of privatization (first


Senegal1

SONATEL Partially privatized: 72,33% Date of privatization (first phase): 1996: 63,33% sold

Serbia and Montenegro

2

Telecom Serbia Partially privatized: 49% Date of privatization (first phase): 1997

SeychellesCable and Wireless (Seychelles) Ltd Fully privatized

Telecom (Seychelles) Ltd Fully privatized

Sierra Leone2 State-owned - Corporatized The government intends to privatize the operator (Year: 2004)

Sierratel






SingaporeSingapore Telecommunications Ltd Partially privatized: 32.44% Date of privatization (first

phase): 1993: 22.2% sold

Slovak RepublicSlovak Telecom a.s. Partially privatized: 51% Date of privatization (first


SloveniaTelekom Slovenije Partially privatized: 37.47% Date of privatization (first

phase): 1996

Solomon Islands2

Solomon Telekom Company Ltd. Partially privatized: 42%

Somalia2

Somtel Fully privatized

South Africa1

Telkom Partially privatized: 61.7% Date of privatization (first phase): 1997: 30% sold


REGULATORY TABLES


196



SpainTelefónica Fully privatized Date of privatization (first

phase): 1997

Auna Fully privatized Date of privatization (first phase): 1997

Grupo Ono Fully privatized

Euskaltel Partially privatized

Sri Lanka1

Sri Lanka Telecom Ltd Partially privatized: 50.5% Date of privatization (first phase): 1997: 35% sold

Suntel Ltd Fully privatized

Lanka Bell (Pvt) Ltd Fully privatized

St. LuciaCable & Wireless (West Indies) Limited

Fully privatized

St. Vincent and theGrenadines Cable and Wireless (WI) Limited Fully privatized

SudanSUDATEL Partially privatized Date of privatization (first


KANARTEL Fully privatized Date of privatization (first phase): 2005: 100% sold






Suriname State-owned - Corporatized No intention to privatize at present

Telesur

Swaziland2 State-owned - CorporatizedSwaziland Posts & Telecommunications Corporation

SwedenTeliaSonera AB Partially privatized: 45.3% Date of privatization (first

phase): 2000: 29.4% sold

SwitzerlandSwisscom AG Partially privatized: 34% Date of privatization (first



S.T.E.

Tajikistan2

Tajiktelecom Partially privatized: 5%


Syria

REGULATORY TABLES


198



Tanzania1

Tanzania Telecommunications Company Ltd (TTCL)


TFYR MacedoniaA.D. Makedonski telekomunikacii Partially privatized: 53% Date of privatization (first


Thailand1 State-owned - CorporatizedTOT Corporation Plc., Ltd

TRUE Corporation Plc., Ltd

TT&T Plc. Ltd

Togo State-owned - Corporatized In the process of privatizingTogo Télécom

Tonga2 State-owned - Corporatized In the process of privatizingTonga Telecommunications Corporation Limited (TCC)

Trinidad and TobagoCable & Wireless Partially privatized: 49% Date of privatization (first







Tunisia State-owned - Corporatized The government intends to privatize the operator

Tunisie Télécom

Turkey State-owned - Corporatized In the process of privatizingTurk Telekom

Turkmenistan2 State-owned - Corporatized (Year: 1992)

Turkmentelecom

Tuvalu2 State-owned - Corporatized (Year: 1994)

No intention to privatize at present

Tuvalu Telecom Corporation

UgandaUganda Telecom Partially privatized: 51% Date of privatization (first


MTN Fully privatized Date of privatization (first phase): 1998: 100% sold

Ukraine2 State-owned - CorporatizedUkrainian Telecom Corporation


REGULATORY TABLES


200



United Arab EmiratesEtisalat Partially privatized: 40%

United KingdomBritish Telecommunications Plc Fully privatized Date of privatization (first


Kingston Plc Fully privatized

United States1

All fixed-line operators Fully privatized

Uruguay2 State-owned No intention to privatize at present

ANTEL

Uzbekistan2 State-owned - CorporatizedHalkapo and Machalit

Vanuatu2

Telecom Vanuatu Ltd Partially privatized: 67%






VenezuelaCANTV Partially privatized: 93.41% Date of privatization (first


MOVISTAR Fully privatized

DIGITEL Fully privatized

INFONET Fully privatized

Viet Nam State-owned In the process of privatizingVNPT

State-ownedVIETTEL, ETC, VISHIPEL

SPT Partially privatized

HANOI TELECOM Partially privatized

Yemen1 State-owned - Corporatized No intention to privatize at present

Public Telecom Corporation

Zambia State-owned No intention to privatize at present

Zamtel

Zimbabwe State-owned - CorporatizedTel One

TELEACCESS Fully privatized




WORLD3. Level of competitionInt'l GMPCS ISIMT-

2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased

lines Data VSAT Pag-

ingMSSFSSCable

T VCable

modemIGFWB

C P P ... P ... P ... P PP ...PAfghanistan ...P ... ... PP

C P P P PC C C C CC ...CAlbania CP ... ... PC

P P P P PC P ... P C ... ... ...Algeria CP ... ... PC

M M M ... MM ... M ... ...D M ...Andorra2 ... ... ... ... ...M

C C C C PC ... C C CP C ...Angola1 ...C ...C CC

M ... M C CC C C ... CC ... ...Antigua and Barbuda

MC ...C MM

C C C ... CC C C C CC CCArgentina2 CC ... ... ...C

M M M C PC ... M C CC CCArmenia C ... ... ... ...M

C C C C CC C C C CC CCAustralia CC CC CC

C C C C CC C C C CC CCAustria1 CC CC CC

P M P C PC C P P CC ...CAzerbaijan CP ...C MP

P P P ... MP C C ... CM ... ...Bahamas C ... ...M MP

C ... C ... PC C C ... C ... ... ...Bahrain C ... P ... CC

C C M ... CC C M ... C ... ...MBangladesh CP ... ... M ...

M ... P ... C ... C C ... CC ... ...Barbados ...P ... ... PM

C M M ... PC ... C M CC MMBelarus C ... ...C MM

M = Monopoly; D = Duopoly; P = Partial competition; C = Full competition; ... = Not available

Note: This table reflects what is legally permissible; therefore it may not reflect the actual number of operators in the market.

Key: WLL = Wireless local loop; DSL = Digital subscriber line; VSAT = Very small aperture terminal; FWB = Fixed wireless broadband; FSS = Fixed-satellite service; MSS = Mobile-satellite service; GMPCS = Global Mobile Personal Communications System; IS = Internet services; IG = International gateways;


REGULATORY TABLES


204


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

C C C ... PC C P C C ... CCBelgium2 CP PC ...C

M M M ... ...M ... M P MC ... ...Belize2 ...M ... ... ...M

M M M ... C ... ... C ... ... ... ... ...Benin2 M ... ... ... ...M

M M M M MP P M M PC M ...Bhutan ...M ... ... MM

M M M ... CC ... ... ... CC MMBolivia2 ... ... ... ... ...M

M M M ... P ... ... ... ... CC ... ...Bosnia and Herzegovina

2 ... ... ... ... ... ...

M C M C CC C C ... C ... ... ...Botswana CM ... ... MC

C C C ... CC C C C CC CCBrazil2 CC CC ...C

P P P ... MP ... M ... P ... ... ...BruneiDarussalam

1 ... ... ... ... PP

P P P P PC C C ... CC ...CBulgaria CC P ... ...C

M M M C CC P C P CC PPBurkina Faso1 CC PC MP

C C C ... CC C C C CC C ...Burundi2 CC C ... ...C

P P P ... PP P P ... PP ... ...Cambodia2 ...P ... ... ...P

M M M ... CC P C C CC ... ...Cameroon1 ...C ... ... ... ...

C C C C CC C C C C ... CCCanada CC CC CC

M M M C CC P C C CC PCCape Verde CM CC M ...








2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

M M M ... CC ... ... C ...C ... ...CentralAfrican Rep.

2 ... ... ... ... ...C

M C M ... ...C C ... ... C ... CCChad1 ... ... ... ... ... ...

P C C M CC C ... M CM MCChile PM ...M CP

P P P ... PP C C ... C ... MMChina PP ... ... ...P

P C C ... PC ... C ... CC ... ...Colombia ... ... ... ... ... ...

M M M ... ...M ... ... ... M ... ...MComoros2 ...M ... ... ...M

C C C ... C ... ... C ... ... ... C ...Congo2 ... ... ... ... ... ...

C C P ... CC C P C CC CCCongo (Dem. Rep. of)

1 CP CP ...P

M M M M MM M C M MP MMCosta Rica MM MP ...M

P P P P PC C C C C ... ...CCôte d'Ivoire1 MC ... ... ...P

C C C C CC C C C CC CCCroatia1 CC CC CC

M M M ... MM M M M PP MMCuba MM ...M ...M

C C C ... PC ... M ... CP CCCyprus M ... ... ... ...C

C C C C CC C C C CC CCCzechRepublic

CC CC CC

M M M ... ... ... ... ... ... ... ... ... ...D.P.R. Korea2 ... ... ... ... ... ...





REGULATORY TABLES


206


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

C C C P PC C ... C CC CCDenmark CC PC CC

M M M ... ...M ... M ... M ... MMDjibouti2 ... ... ... ... ...M

M M M ... MM ... M ... ...C MMDominica2 ... ... ... ... ...M

C C C C CC C C C CC CCDominicanRep.

1 CC CC CC

P P P C PC C C ... CC CCEcuador CP PC ...C

M M M ... PC C ... C C ... ... ...Egypt CM ... ... MM

C C C ... CC C C C CC CCEl Salvador CC ...C ...C

M M M ... ...M ... ... ... MC MMEquatorialGuinea

2 ... ... ... ... ... ...

M M M M PP P ... ... PM PMEritrea ...M ... ... MM

C C C C CC C C C CC CCEstonia CC CC CC

M M M M MM M ... M M ... MMEthiopia1 ...M ... ... MM

M M M ... MM ... C ... ...M CMFiji2 ... ... ... ... ...M

C C C ... PC C ... C CC CCFinland1 CC PC CC

C C C C CC C C C CC CCFrance CC CC CC

M C C ... CC C ... C C ... CMGabon CC MM ...C

M M M ... PC M ... ... C ... ... ...Gambia MM ... ... MM








2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

P P C ... CC C ... C CC PPGeorgia PP ...C ...M

C C C C CC C C C CC CCGermany CC CC CC

P P P ... PC P C C CC CCGhana1 PP PC ...P

C C C ... PC C P P CC PCGreece CC PC CC

M M M ... MM ... M ... ... ... MMGrenada2 ... ... ... ... ...M

C M C ... CC C C ... CC CCGuatemala2 ...C ... ... ...C

P P P ... P ... ... C C CC CCGuinea2 ... ... ... ... ... ...

M M M C PC C C C CC CCGuinea-Bissau1 CC PC PM

M M M ... CD ... C ... ... ... ... ...Guyana2 ... ... ... ... ...M

P P P ... PC C ... ... CC ... ...Haiti ... ... ... ... ... ...

M M M ... MC C C C ...C CCHonduras2 ... ... ... ... ...P

C C C C PC C ... C ...C CCHungary ...P PC CC

C C C ... CC C M ... C ... ... ...Iceland2 CC ... ... ...C

C C C ... CC C C ... CC ... ...India ...C ...C C ...

P P P C CC C C C CC CCIndonesia1 CP CC CC

P M M P PP M C M P ... MMIran (I.R.) CC ... ... MM

M M M ... ... ... ... ... ... ... ... ... ...Iraq2 ... ... ... ... ... ...





REGULATORY TABLES


208


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

C C C C CC C C C CC CCIreland CC CC CC

M M C ... C ... C C ... ...M CCIsrael2 ...C ... ... ...M

C C C ... CC C C C CC CCItaly2 CC C ... ...C

C C C C CC C ... ... CP ... ...Jamaica CC ...C CC

C C C C C ... C C C CC CCJapan CC CC CC

P P P C PC C C C CC ...CJordan CC ...C CC

C C C ... DC ... C ... ...C ...CKazakhstan2 ... ... ... ... ...C

P P P C PC P C P C ... ... ...Kenya1 CC ... ... P ...

M M M ... ...M ... ... ... ... ... ... ...Kiribati2 ... ... ... ... ...M

C C C ... CC ... C P C ... ... ...Korea (Rep.) CC CC ...C

M ... M M ...P P P M PP ...MKuwait ... ... ... ... MM

C C C C CC C C C CC CCKyrgyzstan CC ...C CC

P P P P PP ... ... ... P ... ... ...Lao P.D.R. PP P ... M ...

C C C ... CC C C ... CC ... ...Latvia C ... ...C ...C

M M M ... ...C ... ... ... ... ... ...MLebanon2 ... ... ... ... ... ...

P P P ... CP P ... ... C ... ... ...Lesotho1 ...P ... ... MM

P P C ... CP P ... ... ... ... ... ...Liberia2 ...M ... ... ... ...








2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

M M M ... M ... ... ... ... ... ... MMLibya2 ... ... ... ... ...M

C ... C ... PC ... M ... CC ... ...Liechtenstein2 C ... PP ...C

C C C C CC C ... C CC CCLithuania CC ...C CC

C C C C CC C C C CC CCLuxembourg CC CC CC

M M C C CC C C C CC CCMadagascar CC CC CC

M M P ... P ... P ... ... P ... ... ...Malawi2 ... ... ... ... ... ...

C C C C CC C C C CC CCMalaysia CC CC CC

... ... ... ... ... ... ... ... ... ... ... ... ...Maldives1 ... ... ... ... ... ...

P P P C PP P ... ... C ... PPMali ...C ...P ...P

C ... C C CC C C C CC CCMalta CC CC CC

M M M ... ...M ... ... ... ...C MMMarshallIslands

2 ... ... ... ... ...M

M C C C CC C C C CC CCMauritania1 CC CC CC

C ... C C CC C C C C ... CCMauritius CC C ... CC

C C C C CC C C C CC CCMexico PC ...C ...C

M M M ... CM ... C ... ...C CCMicronesia2 ... ... ... ... ...M

C C C C CC C C C CC CCMoldova CC ...C CC





REGULATORY TABLES


210


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

M M M ... MM M C ... CM MMMonaco2 MM ... ... ...M

P P C ... PC C ... ... CC CCMongolia MC ... ... CC

M M M ... CC C C C C ... CCMorocco MM ... ... PC

M M M M CC C C C CC ... ...Mozambique M ... ... ... CC

M M M P MP P ... ... P ... ...MMyanmar PP ... ... MP

M M M M MM C C C CC CMNamibia MM ...M M ...

M M M ... ...M ... ... ... ... ... MMNauru2 ... ... ... ... ...M

P P P C PC C C C CC ... ...Nepal CP ...C PC

C C C ... PP P ... ... PP ... ...Netherlands CP ...P ...C

C C C C CC C C C CC CCNew Zealand CC CC CC

C C C C CC C C C CC ... ...Nicaragua C ... ...C ...C

M M M ... CM P ... C MP ... ...Niger2 ... ... ... ... ...M

C P P ... P ... C C C C ... C ...Nigeria2 ...C ... ... ...P

C C C P PC C ... ... CC CCNorway CC PC ...C

M M M M PM M M M M ... MMOman MM ... ... MM

C C C C PC C C C C ... CCPakistan CC ... ... CC

C C C P PC P P ... CP ...PPanama1 PM ...C ...M








2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

M M M ... MM M M M PM MCPapua New Guinea

2 ...M M ... ...M

M M M ... CC C C C CC CCParaguay1 ...C CC ... ...

C C C ... CC ... C ... ...C C ...Peru ... ... ... ... ... ...

C C C ... CC C C C CC CCPhilippines2 CC C ... ...C

C C C C CC C C C CC CCPoland CC CC CC

C C C C CC C ... C CC CCPortugal1 CC CC CC

M M M M MM M M M MM MMQatar MM MM MM

C C C C CC C C C CC CCRomania CC CC CC

D D D ... CC ... C ... ...C DDRussia2 ... ... ... ... ...M

C C ... ... CC C ... C C ... ... ...Rwanda1 ... ... ... ... ...C

M ... M ... ...M M ... ... ... ... ...MS. Tomé and Principe

2 ...M ... ... ...M

M M M M MP M C M CC MMSamoa PP MC MM

M M M ... CD ... D ... ... ... ...DSan Marino2 ... ... ... ... ...D

M P P M PP P M C C ... PMSaudi Arabia PP P ... PP

C C C C CC C ... ... C ... ... ...Senegal1 CC ...C CC





REGULATORY TABLES


212


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

C C ... ... CC ... C ... ...C ...MSerbia and Montenegro

2 ... ... ... ... ...C

P P P P PP ... ... ... PM ... ...Seychelles ... ... ... ... ...P

M M P ... CP P P ... P ... ...MSierra Leone2 ... ... ... ... ...P

C ... C C CC C C C CM CCSingapore CC CC CC

C C C C CC C C C CC CCSlovakRepublic

CC CC CC

C C C C CC C C C CC CCSlovenia CC CC CC

M M M ... ...M ... ... ... ... ... ... ...SolomonIslands

2 ... ... ... ... ... ...

C C C ... ... ... ... ... ... ... ... ... ...Somalia2 ... ... ... ... ... ...

C C C P PC P C ... C ... PPSouth Africa1 MM ...M PM

C C C C CC C C C CC CCSpain CC CC CC

P C P ... CC C C ... C ... ... ...Sri Lanka1 MP ... ... ...P

... ... ... C C ... C ... C ...C CCSt. Lucia ...C CC C ...

C C C C CC C C C CC CCSt. Vincent and the Grenadines

CC CC CC

P P P P PP P P M PP PPSudan PP P ... MP

M M M M MM M M M PP MMSuriname MM MM MM








2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

M M M ... MM C ... C ... ... CMSwaziland2 ...M ... ... ...M

C C C C CC C C C CC CCSweden CC CC CC

C C C P PC ... P ... CC CCSwitzerland CP PC CP

M M M ... PM M M M PM MMSyria MM PM ...M

M M M ... ...D ... D ... ...D DDTajikistan2 ... ... ... ... ...M

M M M ... CC C C ... C ... ... ...Tanzania1 CM ... ... MM

M M M ... CC ... C ... CC ... ...TFYRMacedonia

CM ...M ...M

P ... M ... P ... C C P C ... MMThailand1 P ... ... ... MC

P M P C PC C C ... C ... PPTogo ...C ... ... PM

P P P ... ...M ... M ... P ... ... ...Tonga2 M ... ... ... ...M

M ... P P MC C C ... CM ...PTrinidad and Tobago

PM ...M MP

M M M ... CC C C C C ... C ...Tunisia PM CC ...M

P C C P PC C P C CC CCTurkey CC PC CC

C M M ... CM ... D ... ... ... DDTurkmenistan2 ... ... ... ... ...D

M M M ... ...M M ... ... M ... M ...Tuvalu2 ... ... ... ... ... ...

P P P P PC P C ... ...C PPUganda PP ...C PP





REGULATORY TABLES


214


2000DSLLocal

servicesLong

distanceWLL Mo-

bile Leased


ingMSSFSSCable

T VCable

modemIGFWB

D M C ... CC ... C ... ...C CCUkraine2 ... ... ... ... ...M

P P P P PP P M M PM M ...United Arab Emirates

PP PP PP

C C C ... CC C P C CC CCUnitedKingdom

CP CC ...P

C C C ... CC C C C CC CCUnited States1 CC CC ...C

M M P ... CC C C C CP CCUruguay2 MM C ... ...M

D D D ... CD ... C ... ...C DDUzbekistan2 ... ... ... ... ...D

M M M ... ...M ... ... ... ... ... ... ...Vanuatu2 ... ... ... ... ... ...

... ... ... ... ... ... ... ... ... ... ... ... ...Vatican2 ... ... ... ... ... ...

C C C ... CC C C ... CC ... ...Venezuela CC ... ... C ...

C C C ... CC C C ... CC ... ...Viet Nam CC ...C CC

M M M ... CM M M C C ... ...MYemen1 M ... ... ... MM

M M M C PP P P P P ... PPZambia PM ... ... CM

C P P ... CC C ... ... C ... ... ...Zimbabwe C ... ... ... MC






GLOSSARY OF TERMS 215

GLOSSARY OF TERMS*

The following definitions are included to assist the readers of this report. They are adapted from non-definitive reference sources and are not intended to replace or contradict the terms and meanings used by each ITU Member State in its national laws and regulations or in international agreements.

3G: Third-generation mobile network or service. Generic name for mobile network/service based on the IMT-2000 family of global standards.

802.20: Refers to IEEE 802.20, a new standard that is being developed for mobile broad band wireless access.

AAS: Adaptive Array System. A system that is designed to enhance the detection and reception of certain desired radio signals.

Access BPL: A system that uses electrical distribution lines, overhead or underground, to provide broadband Internet access to homes and businesses.

Active optical network:

A network in which the passive splitting point is replaced with Optical Line Distribu-tion unit which is a powered unit making it possible to have higher bit rate on individual routes over longer distances than passive optical network.

ADSL: Asymmetric digital subscriber line. A technol-ogy that enables high-speed data services to be delivered over twisted pair copper cable, typically with a download speed in excess of 256 kbit/s, but with a lower upload speed. Corresponds to ITU Recommendation (standard) ITU-T G.992.1.

ADSL2: Asymmetric Digital Subscriber Line 2 (ITU-T G.992.3 and ITU-T G.992.4). A sequel to the original ITU Recommendation. It allows increased line speeds, new power-saving elements, and extends the reach of the original ADSL specification.

ADSL2+: Asymmetric digital subscriber line 2 plus (ITU-T G.992.5). This revised version of ADSL2 enables increased speeds by increasing the frequencies used on the copper line.

Analogue: Transmission of voice and images using electrical signals. Analogue mobile cellular systems include AMPS, NMT and TACS.

Analogue network:

A tele com mu ni cation network in which information is conveyed as a continuously varying electronic signal (see also Digital network).

ASP: Application service provider. Provider of a service that allows users to run applications remotely from a server rather than having the actual programs installed on their com-puters. This allows for higher power appli-cations to run on small or basic terminals.

ATM: Asynchronous transfer mode. A transmission mode in which the information is organized into cells; it is asynchronous in the sense that the recurrence of cells from an indi-vidual user is not necessarily periodic.

Bandwidth: The range of frequencies available to be occupied by signals. In analogue systems it is measured in terms of Hertz (Hz) and in digital systems in bits per second (bit/s). The higher the bandwidth, the greater the amount of information that can be transmit-ted in a given time. High bandwidth chan-nels are referred to as “broadband” which typically means 1.5-2.0 Mbit/s or higher.

Baseband: Refers to transmitting on only a single chan-nel at any one time.

Base station: A radio transmitter/receiver and antenna used in the mobile cellular network. It maintains communications with cellular telephones within a given cell and transfers mobile traffic to other base stations and the fixed telephone network.

GLOSSARY OF TERMS


216

Bent pipe star topology:

A satellite system topology characterized by a large gateway earth station that transmits one or more high-data-rate, forward-link broadcasts to a large number of small user terminals. These broadcasts contain address information that allows each user terminal to select those transmissions intended for it. In the return direction, the remote user ter-minals transmit in bursts at low-to-medium data rates to the gateway.

Bent pipe point-to-point:

This topology calls for a dedicated duplex connection, set up between a large gateway earth station and a single user terminal.

Bit (binary digit):

A bit is the primary unit of electronic, digital data. Written in base-2, binary language as a

“1” or a “0”.

Bit/s: Bits per second. Measurement of the trans-mission speed of units of data (bits) over a network. Also kbit/s: kilobits (1 000) per second; Mbit/s: megabits (1 000 000) per second, and Gbit/s: Gigabits (1 000 000 000) per second.

Blog: Blog is short for weblog. A weblog is a jour-nal (or newsletter) that is frequently updated and intended for general public consump-tion.

Bluetooth: A radio technology that enables the trans-mission of signals over short distances between mobile phones, computers and other devices. It is typically used to replace cable.

Botnets: A jargon term for a collection of software robots, or bots, which run autonomously. A botnet’s originator can control the group remotely, usually through a means such as IRC, and usually for nefarious purposes.

Broadband: Although there exist various definitions of broadband that have assigned a minimum data rate to the term, it may be defined as transmission capacity with sufficient bandwidth to permit combined provision of voice, data and video, with no lower limit. Effectively, broadband is implemented mainly through ADSL, cable modem or wireless LAN (WLAN) services.

Broadband corDECT:

A broadband wireless local loop standard developed in India by Indian Institute of Technology.

Broadband Over Power Line (BPL):

A wireline technology that is able to use the current electricity networks for data and voice transmission.

Broadband Wireless Access (BWA):

Encompasses either mobile or fixed access technologies that provide connections at speeds higher than the primary rate (for example, 2 Mbit/s).

Browser: Application that retrieves WWW documents specified by URLs from an HTTP server on the Internet. Displays the retrieved docu-ments according to the Hyptertext Markup Language (HTML).

Burstiness: Technical jargon used to describe a high peak-to-average rate of packets as they are received over the network. There is no unique mathematical definition of “bursti-ness”, but a traffic stream is considered to be more “bursty” than another if its packets are more clumped together.

Byte: (1) A set of bits that represent a single char-acter. A byte is composed of 8 bits. (2) A bit string that is operated upon as a unit and the site of which is independent of redundancy or framing techniques.

Cable modem: A technology that allows high-speed interac-tive services, including Internet access, to be delivered over a cable TV network.

Cable Television (CATV):

A system for delivery of television video and audio content via a wired network, employ-ing either co-axial cable or fibre.

Calling Party Pays (CPP):

Billing option whereby the person making the call is charged. By contrast, in a “receiv-ing party pays” (RPP) system, the individual that receives the call pays all charges for that call.

Cellular: A mobile telephone service provided by a network of base stations, each of which covers one geographic cell within the total cellular system service area.

CAGR: Compound annual growth rate. See the Techni-cal Notes.

CDMA: Code division multiple access. A technology for digital transmission of radio signals based on spread spectrum techniques where each voice or data call uses the whole radio band and is assigned a unique code.

CDMA2000: Code division multiple access 2000. A third-generation digital cellular standard based on Qualcomm technology. Includes CDMA2000 1x, 1xEV-DO (Evolution, Data Optimized) and 1xEV-DV (Evolution, Data and Voice). One of the IMT-2000 “family” of standards.



Cellular: A mobile telephone service provided by a network of base stations, each of which covers one geographic cell within the total cellular system service area.

Channel: One of a number of discrete frequency ranges utilized by a base station to transmit and receive information from cellular termi-nals (such as mobile handsets).

Circuit-switched connection:

A temporary connection that is established on request between two or more stations in order to allow the exclusive use of that con-nection until it is released. At present, most voice networks are based on circuit-switch-ing, whereas the Internet is packet-based. See also Packet-based.

CMTS: Cable Modem Termination System. Equipment typically found in a cable company’s head-end and is used to provide high speed data services, such as Cable Internet or Voice over IP, to cable subscribers.

Collocation: Facility-sharing in which the incumbent operator houses communications equipment of competitive operators to facilitate connec-tivity to end users.

Competition: Refers to introducing competition among national service suppliers and/or foreign suppliers without any limitations. In the case of mobile cellular, the number of licen-sees is dependent on spectrum availability. Therefore, all countries allowing more than one operator have been listed in this report as “competitive”.

CompetitiveLocal ExchangeCarrier (CLEC):

A network operator or carrier–often a new market entrant--that provides local teleph-ony in competition with the incumbent carrier.

Condominium Fibre Build:

A network model where a group of end-users band together to install strands of fibre optic cable to an ISP at the same time. At completion, the end-users are each given separate strands of fibre for their own usage.

Connectivity: The capability to provide, to end-users, con-nections to the Internet or other communi-cation networks.

Convergence: A term used to describe a variety of tech-nological and market trends involving the blurring of previously distinct lines between market segments such as cable television, telephony and Internet access, all of which can now be provided through a variety of different network platforms.

Corporatiza-tion:

Corporatization involves legal changes to grant the tele com mu ni cation operator ad min is trative and financial autonomy from central government.

Coverage: Refers to the range of a mobile cellular network, measured in terms of geographic coverage (the percentage of the territorial area covered by mobile cellular) or population coverage (the percentage of the population within range of a mobile cellular network).

CSMA: Carrier Sense Multiple Access. A network pro-tocol in which a node verifies the absence of other traffic before transmitting on a shared physical medium, such as an electrical bus, or a band of electromagnetic spectrum.

Distributed Denial of Service (DDoS):

An attack on a computer system or network that causes a loss of service to users, typically the loss of network connectivity and services by consuming the bandwidth of the victim network or overloading the computational resources of the victim system through a system of computers, which are usually zombie computers compromised by viruses or Trojan horse programs.

Digital: Representation of voice or other informa-tion using digits 0 and 1. The digits are transmitted as a series of pulses. Digital networks allow for higher capacity, greater functionality and improved quality.

Digital network:

A tele com mu ni cation network in which information is converted into a series of distinct electronic pulses and then transmit-ted as a digital bit stream (see also Analogue network).

DOCSIS: Data over cable systems interface specifications (ITU-T J.112). An ITU Recom mendation for cable modems. It specifies modulation schemes and the protocol for exchanging bi-directional signals over cable.

DOCSIS2: Data over cable systems interface specifications 2 (ITU-T J.122). The newest, revised version of DOCSIS, approved at the end of 2002.

Domain name: The registered name of an individual or organization eligible to use the Internet. Domain names have at least two parts and each part is separated by a dot (point). The name to the left of the dot is unique for each top-level domain name, which is the name that appears to the right of the dot. For instance, The International Telecommu-nication Union’s domain name is itu.int.

“ITU” is a unique name within the gTLD “int”.

GLOSSARY OF TERMS


218

DSL: Digital subscriber line. See also ADSL, ADSL2, ADSL2+, SHDSL, SDSL, VDSL and xDSL.

DSLAM: Digital subscriber line access multiplexer. A device, located at the central office of a DSL provider, that separates and routes the voice-frequency signals and data traffic on a DSL line.

DSP: Digital signal processing. The study of signals in a digital representation and the processing methods of these signals

DVB-RCS: Digital Video Broadcasting, with a Return Channel via Satellite. An open standard for digital television maintained by the DVB Project, an industry consortium with more than 270 members, and published by a Joint Technical Committee (JTC) of European Tele-communications Standards Institute (ETSI), European Committee for Electrotechnical Standardization (CENELEC) and European Broadcasting Union (EBU).

Dynamic frequency selection (DFS):

It detects the presence of a primary service and switches the WLAN to a clear frequency.

E.164: An ITU-T recommendation which defines the international public tele com mu ni cation numbering plan used in the PSTN and some other data networks.

E-commerce: Electronic commerce. Term used to describe transactions that take place online where the buyer and seller are remote from each other.

ECS: Electronic Communication Service. Services provided for remuneration and consist-ing wholly or mainly in the conveyance of signals on Electronic Communications Networks

EDGE: Enhanced Data rates for GSM Evolution. It acts as an enhancement to 2G and 2.5G General Packet Radio Service (GPRS) networks. This technology works in TDMA and GSM net-works. EDGE (also known as EGPRS) is a superset to GPRS and can function on any network with GPRS deployed on it, pro-vided the carrier implements the necessary upgrades. EDGE provides Enhanced GPRS (EGPRS), which can be used for any packet switched applications such as an Internet connection. High-speed data applications such as video services and other multimedia benefit from EGPRS’ increased data capacity.

E-mail: Electronic mail. The exchange of electronic messages between geographically dispersed locations.

End-user: The individual or organization that origi-nates or is the final recipient of information carried over a network (i.e. the consumer).

Endrun: A fibre optic infra struc ture that provides a dedicated fibre optic cable directly to each user’s premise rather than several premises optically splitting off one line. See PON.

ENUM: Standard adopted by Internet Engineering Task Force (IETF), which uses the domain name system (DNS) to map telephone numbers to Web addresses or uniform resource locators (URL). The long-term goal of the ENUM standard is to provide a single number to replace the multiple numbers and addresses for users’ fixed lines, mobile lines, and e-mail addresses.

EPOP: Expanding point of profitability. A network topography where the network expands incrementally to unserved areas as they become profitable to operators. Newly con-nected areas can then be used as backbones to more remote areas as they eventually become profitable to providers.

Ethernet: A protocol for inter con necting comput-ers and peripheral devices at high speed. Recently Gigabit Ethernet has become avail-able which enables speeds up to 1 Gbit/s. Ethernet can run on several types of wiring including: twisted pair, coaxial, and even fibre optic cable.

Ex-ante and ex-post regulation:

Ex-ante regulation involves setting specific rules and restrictions to prevent anti-com-petitive or otherwise undesirable market activity by carriers before it occurs; ex-post regulation, by contrast, calls for setting few or no specific rules in advance, but applying corrective measures and punishments if and when transgressions do occur.

FDMA: Frequency division multiple access. A cellular technology that has been used in the first-generation analogue systems (i.e. NMT, AMPS, and TACS).

Fixed line: A physical line connecting the subscriber to the telephone exchange. Typically, fixed-line network is used to refer to the PSTN (see below) to distinguish it from mobile net-works.



FPGA: Field programmable gate array. A semiconduc-tor device containing programmable logic components and programmable inter con-nects. The programmable logic compo-nents can be programmed to duplicate the functionality of basic logic gates (such as AND, OR, XOR, NOT) or more complex combinatorial functions such as decoders or simple math functions.

FTTH: Fibre to the home. A high-speed fibre optic, Internet connection that terminates at a resi-dence. See FTTx.

FTTx: Fibre to the x, where x is a home (FTTH), building (FTTB), curb (FTTC), or neigh-bourhood (FTTN). These terms are used to describe the reach of an optical fibre network.

Firewall: Software or hardware that controls access in and out of a network. Firewalls can be dedi-cated computers that act as the intermediary between a business network and the Inter-net, or can be software tools that help indi-vidual computers control which programs are allowed access to the Internet.

First Mile: A topology in which the user or a local service provider – or perhaps even an apart-ment building company – owns the access network and connects to service providers using its own upstream links.

Fixed line: A physical line connecting the subscriber to the telephone exchange. Typically, fixed-line network is used to refer to the PSTN (see below) to distinguish it from mobile net-works.

Frequency: The rate at which an electrical current alter-nates, usually measured in Hertz (see Hz). It is also used to refer to a location on the radio frequency spectrum, such as 800, 900 or 1 800 MHz.

Frequency-Division Duplexing (FDD):

A transmission mode that requires the allocation of two frequency bands, one for the uplink and another for the downlink. It enables one to transmit and receive at the same time.

Frequency-Division Multiplexing (FDM):

A form of signal multiplexing where mul-tiple baseband signals are modulated on different frequency carrier waves and added together to create a composite signal.

FSO: Free space optics. A system of lasers used to transmit data optically through the atmos-phere at very high speeds. Similar to optical fibre without the physical cable.

Gateway: Any mechanism for providing access to another network. This function may or may not include protocol conversion.

GATS: General Agreement on Trade In Services.

GDP: Gross domestic product. The market value of all final goods and services produced within a nation in a given time period.

GEO: Geostationary earth orbit. A satellite in orbit 35 650 km above the Earth in a rotation that mimics that of the Earth, thus appearing sta-tionary in the sky.

GMPCS: Global mobile personal communications by satellite. Non-geostationary satellite systems that are intended to provide global communication coverage to small handheld devices.

Global System for Mobile communica-tions (GSM):

European-developed digital mobile cellular standard. For more information, see the GSM Association website at: http://www.gsmworld.com/index.html.

GNP: Gross national product. The market value of all final goods and services produced in a nation’s economy, including goods and serv-ices produced abroad.

GNI: Gross national income. The market value of all final goods and services produced in a nation’s economy, including goods and services produced abroad. GNI in constant prices, differs from GNP in that it also includes a terms of trade adjustment; and gross capital formation which includes a third category of capital formation: net acquisition of valuables.

GPRS: General Packet Radio Service. It is a mobile data service available to users of GSM mobile phones. It is often described as

“2.5G”, that is, a technology between the second (2G) and third (3G) generations of mobile telephony. It provides moderate speed data transfer, by using unused TDMA channels in the GSM network.

GPS: Global positioning system. Refers to a “constel-lation” of 24 “Navstar” satellites launched initially by the United States Department of Defense, that orbit the Earth and make it possible for people with ground receivers to pinpoint their geographic location. The loca-tion accuracy ranges from 10 to 100 metres for most equipment. A Russian system, GLONASS, is also available, and a European system, Galileo, is under development.

H.323: An umbrella recommendation from the ITU-T, that defines the protocols to provide audio-visual communication sessions on any packet network.

GLOSSARY OF TERMS


220

Half duplex: Half duplex refers to a communication channel that can only handle one-way traf-fic at a time. In essence, each side of the communication must wait until the other is finished transmitting to start sending infor-mation. By contrast, full duplex communi-cation allows for both parties to broadcast and receive at the same time.

Hotspot: An access point to a wireless local area net-work (WLAN). Hotspots are areas where wireless data can be sent and received, and Internet access is provided to wireless devices. For example, a laptop computer can be used to access the Internet in a hotspot provided in an airport or hotel.

HAPS: High altitude platform station. A term referring to balloons and high altitude aircraft that can be used to provide communication services. See LAPS.

HDTV: High-definition television. A new format for television that offers far superior quality to current NTSC, PAL, or SECAM systems. The resolution of the picture is roughly double previous television signals and the pictures are displayed with a screen ratio of 16:9 as compared with most of today’s TV screens, which have a screen ratio of 4:3.

Hertz (Hz): The frequency measurement unit equal to one cycle per second.

HFC: Hybrid fibre copper. A broadband network that utilizes fibre optic cabling to the vicinity and then copper lines to individual users.

HiperLAN: High-performance radio local area network. An ETSI standard that operates at up to 54 Mbit/s in the 5 GHz RF band.

HiperLAN2: High-performance radio LAN Type 2. Wireless LAN (specified by ETSI/BRAN) in the 5 GHz IMS Band with a bandwidth up to 50 Mbit/s. HiperLAN2 is compatible with 3G WLAN systems for sending and receiving data, images, and voice communications.

HIPERMAN: High performance radio metropolitan area net-work. This is a European standard aimed at providing a broadband wireless solution for Metropolitan Area Networks.

Hotspot: An access point to a wireless local area net-work (WLAN). Hotspots are areas where wireless data can be sent and received, and Internet access is provided to wireless devices. For example, a laptop computer can be used to access the Internet in a hotspot provided in an airport or hotel.

HSDPA: High-Speed Downlink Packet Access. This is a new mobile telephony protocol. Also called 3.5G (or “3½G”). High Speed Downlink Packet Access is a packet-based data service with data transmission up to 8-10 Mbit/s (and 20 Mbit/s for MIMO systems) over a 5 MHz bandwidth in W-CDMA downlink. HSDPA im ple men ta tion includes Adaptive Modulation and Coding (AMC), Multiple-Input Multiple-Output (MIMO), Hybrid Automatic Repeat Request (HARQ), fast scheduling, fast cell search, and advanced receiver design.

Hybrid Fibre/Coaxial (HFC):

A tele com mu ni cations industry term for a network that incorporates both optical fibre along with coaxial cable to create a broad-band network.

Hz: Hertz. The frequency measurement unit equal to one cycle per second.

IMS: IP Multimedia Subsystem. A stand ardized Next Generation Networking (NGN) archi-tecture for telecom operators that want to provide mobile and fixed multimedia serv-ices. It uses a Voice-over-IP (VoIP) im ple-men ta tion based on a 3GPP stand ardized im ple men ta tion of SIP, and runs over the standard Internet Protocol (IP). Existing phone systems (both packet-switched and circuit-switched) are supported.

IMT-2000: International Mobile Telecommunications-2000. Third-generation (3G) “family” of mobile cellular standards approved by ITU. For more information see the website at:http://www.itu.int/imt.

Incumbent: The major network provider in a particu-lar country, often a former State-owned monopoly.

In-house BPL: A home networking technology that uses the transmission standards developed by the HomePlug Alliance. Products for in-home networking use the electric outlets in your home (or office).

Instant messaging (IM):

Refers to programs such as AOL Instant Messenger and ICQ that allow users to exchange messages with other users over the Internet with a maximum delay of one or two seconds at peak times.



Inter-connection:

The physical connection of separate tele-phone networks to allow users of those networks to communicate with each other. Interconnection ensures inter op erability of services and increases end users’ choice of network operators and service providers.

Inter-connection Charge:

The charge–typically including a per-minute fee–that network operators levy on one another to provide inter con nection.

Internet: Interconnected global networks that use the Internet protocol (see IP).

Internet backbone:

The high-speed, high capacity lines or series of connections that form a major pathway and carry aggregated traffic within the Inter-net.

Internet content provider:

A person or organization that provides infor-mation via the Internet, either with a price or free of charge.

IP: Internet protocol. The dominant network layer protocol used with the TCP/IP protocol suite.

IP telephony: Internet protocol telephony. IP telephony is used as a generic term for the conveyance of voice, fax and related services, partially or wholly over packet-based, IP-based net-works. See also VoIP and Voice over broadband.

IPR: Intellectual property rights. Copyrights, patents and trademarks giving creators the right to prevent others from using their inventions, designs or other creations. The ultimate aim is to act as an incentive to encourage the development of new technology and crea-tions which will eventually be available to all. The main international agreements are the World Intellectual Property Organization’s (WIPO) Paris Convention for the Protection of Industrial Property (patents, industrial designs, etc.), the Berne Convention for the Protection of Literary and Artistic Works (copyright), and the World Trade Organization’s (WTO) Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS).

ISDN: Integrated services digital network. A digital switched network, supporting transmission of voice, data and images over conventional telephone lines.

ISP: Internet service provider. ISPs provide end-users access to the Internet. Internet Access Providers (IAPs) may also provide access to other ISPs. ISPs may offer their own propri-etary content and access to online services such as e-mail.

ITU: International Telecommunication Union. The United Nations specialized agency for tele-com mu ni cations. See http://www.itu.int/.

Internet Exchange Point (IXP):

A central location where multiple Internet Service Providers can inter con nect their net-works and exchange IP traffic.

JPEG: Joint photographic expert group compression standard. Standard for the compression and coding of still images.

LAN: Local area network. A computer network that spans a relatively small area. Most LANs are confined to a single building or group of buildings. However, one LAN can be con-nected to other LANs over any distance via telephone lines and radio waves. A system of LANs connected in this way is called a wide-area network (WAN). See also WLAN.

LAPS: Low altitude platform station. A system usually consisting of balloons that provides wireless communication services over a wide area. Similar to HAPS but the altitudes are lower.

Last Mile: The topology denotes the operator’s owner-ship of the access network.

Layered Architecture:

The concept of layered network architec-ture divides a network at any specific point into layers, each of which adds value to the physical medium of communication.

LBS: Location-based services. LBS make use of information on the location of a mobile device and user, and can exploit a number of technologies for the geographic location of a user. Some of these technologies are embedded in the networks and others in the handsets themselves. Location capability is already available to some level of accuracy (approx. 150 m) for most users of cellular networks. Increased accuracy can become available through location technologies such as GPS. See GPS.

Leased line: A point-to-point communication channel or circuit that is committed by the network operator to the exclusive use of an indi-vidual subscriber. Under national law, leased lines may or may not be permitted to inter-con nect with the public switched network.

LEO: Low Earth orbit. A term that refers to satellite orbits between 650 km and 2 600 km above the Earth. A LEO satellite is only in view for a few minutes and rotates the Earth every few hours. See GEO.

Licensing: An ad min is trative procedure for selecting operators and awarding franchises for the operation of particular tele com mu ni cation services, for instance cellular radio.

GLOSSARY OF TERMS


222

Line Sharing: A form of network unbundling that allows a

competitive service provider to offer ADSL using the high-frequency portion of a local loop at the same time that an ILEC contin-ues to offer standard switched voice service over the low-frequency portion of the same loop.

LLU: Local loop unbundling. The process of requir-ing incumbent operators to open the last mile of their legacy networks to competitors. Similar reference to ULL (unbundled local loop).

Local loop: The system used to connect the subscriber to the nearest switch. It generally consists of a pair of copper wires, but may also employ fibre-optic or wireless technologies.

Long Run Incremental Costs (LRIC):

The added or extra cost entailed in provid-ing a service, over the long term.

Long Run Average Incremental Costs (LRAIC):

A costing model based on LRIC analysis, in which the total traffic costs for both inter-con necting carriers are divided by the total demand, rather than assigning unique costs to each operator.

Main telephone line:

Telephone line connecting a subscriber to the telephone exchange equipment. This term is synonymous with the term fixed line used in this report.

Market efficiency gap:

The shortfall between commercially avail-able access and universal access.

Mesh network: A way to route data, voice and instructions between nodes. It allows for continuous connections and reconfiguration around blocked paths by “hopping” from node to node until a connection can be established.

Mobile: As used in this report, the term refers to mobile cellular systems and to mobile phones.

Mobile virtual network operator (MVNO):

A company that does not own a licensed fre-quency spectrum, but resells wireless serv-ices under their own brand name, using the network of another mobile phone operator.

Multimedia: The presentation of more than one medium, typically images (moving or still), sound and text in an interactive environment. Multi-media requires a significant amount of data transfer and bandwidth, and it invariably requires computational facilities.

MP3: MPEG-1 Audio Layer-3 (MPEG stands for Moving Pictures Experts Group). A standard technology and format for compression of a sound sequence into a very small file (about one-twelfth the size of the original file) while preserving the original level of sound quality when it is played.

Multimedia: The presentation of more than one medium, typically images (moving or still), sound and text in an interactive environment. Multi-media requires a significant amount of data transfer and bandwidth, and it invariably requires computational facilities.

National Regulatory Authority (NRA):

The regulatory agency or official at the central or federal government level that is charged with implementing and enforcing tele com mu ni cation rules and regulations.

Network Unbundling:

Providing access to, or making available, some or all of the disaggregated elements and/or functions of a telephone network–usually the local portion of the network–for inter con necting carriers to use in serving their own customers.

Network Topology:

The pattern of links connecting pairs of nodes of a network.

Next genera-tion network (NGN):

These are packet-based networks in which service-related functions are independent from underlying transport-related technolo-gies. They are able to provide telecommu-nication services and make use of multiple broadband transport technologies. For a more detailed definition, see http://www.itu.int/ITU-T/ngn

Node: A point of connection to a network. A switching node is a point at which switching occurs.

Number portability:

The ability of a customer to transfer an account from one service provider to another without requiring a change in number. Other forms of portability allow end users to change residence or subscribe to a new form of service (e.g., ISDN) while retaining the same telephone number for their main telephone line.

OFDM: Orthogonal frequency division multiplexing. A method of digital modulation in which a signal is split into several narrowband chan-nels at different frequencies in order to minimize interference among channels that are close in frequency. OFDM is used in European digital audio broadcast services, and also in wireless LANs.



On-Board Processor (OBP) Switching:

In this topology, the satellite rather than the gateway is the central node in a star network. The satellite is connected to the gateway by one or more high-data-rate trunks. The on-board processor de-multiplexes the uplink trunk into several downlinks for different geographical areas, usually determined by the footprint pattern. The forward down-links contain messages for large numbers of user terminals, and the destinations are identified by message headers. In the return channel, the uplink transmissions from user terminals in one or more cells are multi-plexed onto a downlink trunk to the gateway.

One DSL for Universal Service (See UDSL/UniDSL)

Operator-neutral model:

A network access model in which the passive access network infra struc ture is often owned by housing companies, condominiums or tenant organizations and, in some cases, by municipalities. In this model, access networks are connected to a shared access network backbone. Any service provider can then connect its network gateway and offer services using the access network.

OPGW: Optical power ground wire. A special ground wire with a fibre cable in the core.

Optical network terminal (ONT):

The equipment used to terminate the fiber.

P2P: Peer to peer. P2P refers to networks that facilitate direct connections among indi-vidual nodes rather than through a central-ized server. However, many famous P2P networks, such as “Napster”, actually relied on a central server to connect users. Other networks (such as “Gnutella”) offer true peer-to-peer, decentralized connections.

PABX: Private Automatic Branch eXchange. A tel-ephone exchange that is owned by a private business, as opposed to one owned by a common carrier or by a telephone company.

Packet: Block or grouping of data that is treated as a single unit within a communication net-work.

Packet-based: Message-delivery technique in which pack-ets are relayed through stations in a network. See also Circuit-switched connection.

PAN: Personal area network. For the purposes of this report, a PAN is referred to as the inter con-nection of information technology devices within the range of an individual person, typically within a radius of 10 metres. For example, a person travelling with a laptop, a personal digital assistant (PDA), and a port-able printer could inter con nect these devices through a wireless connection, without the need for physical wiring. Conceptually, the difference between a PAN and a wireless LAN is that the former tends to be centered around one person while the latter has a greater range of wireless connectivity, typi-cally serving multiple users.

PDA: Personal digital assistant. A generic term for handheld devices that combine computing and possibly communication functions.

Peering: The exchange of routing announcements between two Internet Service Providers for the purpose of ensuring that traffic from the first can reach customers of the second, and vice-versa. Peering takes place predomi-nantly at IXPs and usually is offered either without charge or subject to mutually agreed commercial arrangements.

Penetration: A measurement of access to tele com mu ni-cations, normally calculated by dividing the number of subscribers to a particular service by the population and multiplying by 100. Also referred to as teledensity (for fixed-line networks) or mobile density (for cellular ones), or total teledensity (fixed and mobile combined).

Personal Communica-tion Services (PCS):

In the United States, this refers to digital mobile networks using the 1 900 MHz fre-quency. In other countries, it refers to digital mobile networks using the 1 800 MHz fre-quency (See DCS-1800). The term Personal Communications Network (PCN) is also used.

Pervasive computing:

A concept which describes a situation in which computing capability is embedded into numerous different devices around the home or office (e.g. fridges, washing machines, cars, etc.). Also referred to as ubiquitous computing. Pervasive communications implies that the microchips in these devices are also able to communicate, for instance their location and status.

Phishing: The fraudulent practice of disguising spam as legitimate email in an attempt to coax recipients into revealing private financial data.

GLOSSARY OF TERMS


224

PLC: Power line communications. A communication

network that uses existing power lines to send a receive data by using electrical signals as the carrier. Power flows on the line at 50-60 Hz while data is sent in the 1 MHz range.

PON: Passive optical network. A type of full passive wave division multiplexing (WDM) network that allows multiple locations to connect to one optical fibre strand (or wavelength) by using optical splitters to break up the wave-length of light into allocated time slots for each user. See Endrun and WDM.

Port 25: The traditional TCP port used by the Simple Mail Transfer Protocol.

Portal: Although an evolving concept, the term portal commonly refers to the starting point, or a gateway through which users navigate the World Wide Web, gaining access to a wide range of resources and services, such as e-mail, forums, search engines, and shop-ping malls.

PPP: Purchasing power parity. An exchange rate that reflects how many goods and services can be purchased within a country taking into account different price levels and cost of living across countries.

Protocol: A set of formal rules and specifications describing how to transmit data, especially across a network.

Primary rate connection (PRI):

One kind of service for ISDN. It consists of 23 B-channels and one 64 kbit/s D-channel in the United States or 30 B-channels and 1 D-channel in Europe.

Private network:

A network based on leased lines or other facilities, which are used to provide tele com-mu ni cation services within an organization or within a closed user group as a comple-ment or a substitute to the public network.

Private ownership/ Privatization:

The transfer of control of ownership of a state enterprise to private parties, generally by organizing the enterprise as a share com-pany and selling shares to investors. More generally, the term is sometimes used to refer to a wide range of modalities whereby business is opened to private enterprise and investment.

PSTN: Public switched telephone network. The public telephone network that delivers fixed tele-phone service.

PTO: Public tele com mu ni cation operator. A provider of tele com mu ni cations infra struc ture and serv-ices to the general public (“public” refers to the customer base). Also referred to as an operator, service provider, carrier or “telco”.

QoS: Quality of service. A measure of network performance that reflects the quality and reliability of a connection. QoS can indicate a data traffic policy that guarantees certain amounts of bandwidth at any given time, or can involve traffic shaping that assigns vary-ing bandwidth to different applications.

RLAN: Radio local area network. See WLAN.

RFID: Radio-frequency identification. A system of radio tagging that provides identification data for goods in order to make them traceable. Typ-ically used by manufacturers to make goods such as clothing items traceable without having to read bar code data for individual items.

SCADA: Supervision, control and data acquisition. The term refers to a large-scale, distributed measurement (and control) system.

SDR: Software defined radio. A radio communication system which uses software for the modula-tion and demodulation of radio signals.

SDSL: Symmetrical DSL. A proprietary North American DSL standard. However, the term SDSL is often also used to describe SHDSL.

Second-generation mobile (2G):

A general term for second generation mobile networks, for example GSM.

Server: (1) A host computer on a network that sends stored information in response to requests or queries. (2) The term server is also used to refer to the software that makes the proc-ess of serving information possible.

Session Initiation Protocol (SIP):

A protocol developed by the IETF MMUSIC Working Group and proposed standard for initiating, modifying, and terminating an interactive user session that involves multimedia elements such as video, voice, instant messaging, online games, and virtual reality. In November 2000, SIP was accepted as a 3GPP signalling protocol and permanent element of the IMS architecture. It is one of the leading signalling protocols for Voice over IP, along with H.323.

SHDSL: Single pair high-speed DSL. The informal name for ITU-T Recommendation G.991.2 that offers high-speed, symmetrical connec-tivity over a twisted copper pair.



Short Message Service (SMS):

A service available on digital mobile cellular networks and even landline telephones, typi-cally enabling end users to send and receive messages with up to 160 characters.

SkyWrap: A process, where there already are power lines, that wraps fibre around the transmis-sion lines.

SMTP: Simple Mail Transfer Protocol. The de facto standard for email transmission across the Internet.

Softswitch: A type of telephone switch that uses soft-ware running on a computer system to carry out the work that used to be carried out by hardware.

Spam: Unsolicited commercial email, some of which may contain computer viruses or worms, fraudulent consumer scams or offensive content.

Spam Zombie: (see Zombie)

Spectral Efficiency:

A measure of the performance of encoding methods that code information as variations in an analogue signal.

Spectrum: The radio frequency spectrum of hertzian waves used as a transmission medium for cellular radio, radiopaging, satellite commu-nication, over-the air broadcasting and other services.

Spread-spectrum technology:

A radio technique that continuously alters its transmission pattern either by constantly changing carrier frequencies or by con-stantly changing the data pattern.

SPIM: Spam over Instant Messenger. An unsolicited message made using instant messenger over the internet.

SPIT: Spam over Internet Telephony. An unsolicited telephone call made using IP telephony over the internet.

Switch: Part of a mobile or fixed telephone system that routes telephone calls or data to their destination.

Synchronous Digital Hierarchy (SDH):

A standard developed by ITU (G.707 and its extension G.708) that is built on experi-ence in the development of SONET. Both SDH and SONET are widely used today; SONET in the U.S. and Canada, SDH in the rest of the world. SDH is growing in popularity and is currently the main concern with SONET now being considered as the variation.

Synchronous Optical Network (SONET):

A standard for communicating digital information using lasers or light emitting diodes (LEDs) over optical fiber as defined by GR-253-CORE from Telcordia. It was developed to replace the PDH system for transporting large amounts of telephone and data traffic and to allow for inter op erability between equipment from different vendors.

TCP: Transmission control protocol. A transport layer protocol that offers connection-oriented, reliable stream services between two hosts. This is the primary transport protocol used by TCP/IP applications.

TDMA: Time Division Multiple Access. It is a technol-ogy for shared medium (usually radio) networks. It allows several users to share the same frequency by dividing it into different time slots. The users transmit in rapid suc-cession, one after the other, each using their own timeslot. This allows multiple users to share the same transmission medium (e.g. radio frequency) whilst using only the part of its bandwidth they require.

TD-SCDMA: Time Division Synchronous Code-Division Mul-tiple Access. A 3G mobile tele com mu ni cations standard, being pursued in China by the Chinese Academy of Telecommunications Technology (CATT), Datang and Siemens AG, in an attempt to develop home-grown technology and not be “dependent on West-ern technology.” TD-SCDMA uses TDD, in contrast to the FDD scheme used by W-CDMA.

Time-Division Duplexing (TDD):

Enables use of the same frequency for uplink and downlink of a transmission by allocating discrete, short-duration time slots to the two links.

Teledensity: Number of main telephone lines per 100 inhabitants. See Penetration.

Third-Generation Mobile (3G):

A general term for the next generation of broadband digital mobile cellular systems, which will have expanded broadband capa-bilities for mobile data applications. See IMT-2000.

Total teledensity:

Sum of the number of fixed lines and mobile phone subscribers per 100 inhabit-ants. See Penetration.

TPC: Transmit power control. A technical mechanism used within some networking devices in order to prevent too much unwanted inter-ference between different wireless networks.

Traffic Exchange Point:

Traffic exchange points are used by operators to exchange traffic through peering directly between service networks rather than indi-rectly, via transit through their upstream providers.

GLOSSARY OF TERMS


226

Triple Play: A term refers to the bundling of voice, video and data services.

True access gap:

The shortfall between market-based regula-tory measures and universal access.

UASL: Unified Access Services Licensing. A licensing framework in India that gives the licensee freedom to offer both fixed and mobile serv-ices using any technology.

Ubiquitous computing:

A term that reflects the view that future communication networks will allow seam-less access to data, regardless of where the user is. See Pervasive computing.

UDSL/UniDSL:

A new variant of DSL, integrating all ear-lier DSL variants. It promises aggregated bit rates of up to 200 Mbit/s, including 100 Mbit/s symmetrical connections. Uni-DSL would require a fibre backbone infra-struc ture and would use only the part of the existing subscriber line closest to the user premises.

ULL: Unbundled local loop. See LLU.

UMTS: Universal mobile tele com mu ni cations system. The European term for third-generation mobile cellular systems or IMT-2000 based on the W-CDMA standard. For more information see the UMTS Forum website at: http://www.umts-forum.org/.

Universal access:

Refers to reasonable tele com mu ni cation access for all. Includes universal service for those that can afford individual telephone service and widespread provision of public telephones within a reasonable distance of others.

USO: Universal service obligations. Requirements that governments place on operators to offer service in all areas, regardless of economic feasibility.

UTP: Unshielded twisted pair. A cable with one or more twisted copper wires bound in a plastic sheath. It is used extensively for high-speed connections because it allows the release of radiation that would interfere if kept in the line with a shielded cable.

Value-added network serv-ices (VANS):

Telecommunication services provided over public or private networks which, in some way, add value to the basic carriage, usually through the application of com put erized intelligence. Examples of VANs include reservation systems, bulletin boards, and information services. Also known as enhanced services.

VDSL: Very-high-data-rate digital subscriber line. (ITU-T G.993.1). The fastest version of DSL that can handle speeds up to 52 Mbit/s over very short distances. Often used to branch out from fibre connections inside apartment buildings.

Voice over broadband:

A method of making voice calls over a broadband connection. The calls can be either made via a computer or through traditional phones connected to voice over broadband equipment. See also IP telephony and VoIP.

VoIP: Voice over IP. A generic term used to describe the techniques used to carry voice traffic over IP (see also IP telephony and Voice over broadband).

VPN: Virtual private network. A method of encrypt-ing a connection over the Internet. VPNs are used extensively in business to allow employees to access private networks at the office from remote locations. VPNs are especially useful for sending sensitive data.

VSAT: Very Small Aperture Terminal. A 2-way satel-lite ground station with a dish antenna that is smaller than 3 meters, as compared to around 10 meters for other types of satellite dishes.

WAN: Wide area network. WAN refers to a network that connects computers over long distances.

Wavelength division multiplexing:

A technology which multiplexes multiple optical carrier signals on a single optical fibre by using different wavelengths (col-ours) of laser light to carry different signals.

W-CDMA: Wideband code division multiple access. A third-generation mobile standard under the IMT-2000 banner, first deployed in Japan. Known as UMTS in Europe. See also CDMA.

WDM: Wave division multiplexing. Technology that allows multiple data streams to travel simul-taneously over the same fibre optic cable by separating each stream into its own wave-length of light.

Website/Webpage:

A website (also known as an Internet site) generally refers to the entire collection of HTML files that are accessible through a domain name. Within a website, a webpage refers to a single HTML file, which when viewed by a browser on the World Wide Web could be several screen dimensions long. A “home page” is the webpage located at the root of an organization’s URL.



Wi-Bro: A wireless networking technology (IEE 802.16x) that will enable portable Inter-net access. The Republic of Korea has announced plans to commercialize it in 2005-2006.

Wi-Fi: Wireless fidelity. A mark of inter op erability among devices adhering to the 802.11b specification for Wireless LANs from the Institute of Electrical and Electronics Engi-neers (IEEE). However, the term Wi-Fi is sometimes mistakenly used as a generic term for wireless LAN.

Wi-Fi5: Wireless fidelity 5. A mark of inter op erability among devices adhering to the 802.11a standard at 5 MHz.

WiMAX: Fixed wireless standard IEEE 802.16 that allows for long-range wireless communica-tion at 70 Mbit/s over 50 kilometres. It can be used as a backbone Internet connection to rural areas.

Wireless: Generic term for mobile communication services which do not use fixed-line net-works for direct access to the subscriber.

WISP: An ISP that employs a wireless access plat-form, or “wireless ISP”.

WLAN: Wireless local area network. Also known as Wireless LAN. A wireless network whereby a user can connect to a local area network (LAN) through a wireless (radio) connec-tion, as an alternative to a wired local area network. The most popular standard for wireless LANs is the IEEE 802.11 series.

WLL: Wireless local loop. Typically a phone network that relies on wireless technologies to pro-vide the last kilometre connection between the tele com mu ni cation central office and the end-user.

Worm: A self-contained program (usually malicious) that can automatically propagate throughout a network. In addition to damage caused by the program on a user’s machine, the programs can slow down network traffic as all infected machines scan simultaneously to find new hosts.

WSIS: The United Nations World Summit on the Infor-mation Society. The first phase of WSIS took place in Geneva (hosted by the Government of Switzerland) from 10 to 12 December 2003, and the second phase in Tunis (hosted by the Government of Tunisia), from 16 to 18 November 2005. For more information see: http://www.itu.int/wsis.

WTO Agree-ment:

Informal terminology for the Fourth Proto-col to the General Agreement on Trade in Services (GATS). The agreement, concluded in early 1997, included commitments by more than 70 countries to open their mar-kets for basic tele com mu ni cation services. The accompanying Reference Paper, spelled out principles for regulatory treatment of basic tele com mu ni cation service providers, including “major suppliers.”

xDSL: While DSL stands for digital subscriber line, xDSL is the general representation for vari-ous types of digital subscriber line technol-ogy, such as ADSL, SHDSL, and VDSL. See ADSL, SHDSL, VDSL.

Zombie: A computer attached to the Internet that has been compromised by a cracker, a computer virus, or a Trojan horse program.

* Source: ITU, World Telecommunication Development Report, 1999. ITU, Challenges to the Network, Internet Development, 1999.

LIST OF FIGURES


228

LIST OF FIGURES

Figure 1.1: The Number of ICT Users Worldwide, 1994-2004 .........................................................2Figure 1.2: Fixed vs. Mobile Teledensity by Region (per 100 subscribers) ......................................3Figure 1.3: Broadband ..................................................................................................................3Figure 1.4: Global Distribution of Internet and Broadband Subscribers, 2004 ..................................4Figure 1.5: Top 25 Broadband Subscribers, Non-OECD Countries (2004) .......................................4Figure 1.6: Subscriber Growth in the Top 10 Broadband Countries (2000-2004) .............................5Figure 1.7: Broadband Platform Distribution (by Region) 2004 .......................................................6Figure 1.8: Top 10 3G Mobile Markets Worldwide, 2005 ................................................................7Figure 1.9: 2G Standards Usage, by subscribers, world and by region (2004-5) .............................8Figure 1.10: Privatizations, 1991-2005, World.................................................................................9Figure 1.11: Status of Competition Worldwide, 2005 .....................................................................10Figure 1.12: Status of Public Interconnection Agreements and Pricing Information, 2005 ................11Figure 1.13: Countries Requiring Local Loop Unbundling, 2005 .....................................................12Figure 1.14: Number of Countries with National and SubRegional IXPs, by region, 2005 .................13Figure 1.15: Growth of Regulators Worldwide, 2005.......................................................................13Figure 1.16: Separate Regulators, by Region, 2005 ........................................................................14Figure 1.17: Growth of International Traffic .....................................................................................16Figure 1.18: Are Licences Required for WLANS, 2005? ..................................................................17Figure 1.19: Spam as Percentage of Emails Worldwide, 2003-05 ...................................................17Figure 1.20: Spam Regulation, 2005 .............................................................................................18Figure 2.1: WiMAX Growth, Forecast Sales, 2004-5 .....................................................................25Figure 3.1: Comparative Prices for Mobile Data Services .............................................................34Figure 3.2: Migration Paths from 2G to 3G ...................................................................................35Figure 3.3: A mesh network topology versus a traditional network .................................................38Figure 3.4: The Convergence of WLAN and 3G Technologies – A Matrix........................................39Figure 3.5: Empirical Data for Tower, Installation, and Grounding Costs ........................................40Figure 3.6: Linear Model of Tower, Installation, and Grounding costs (USA/Ghana) .......................40Figure 3.7: Radio Tower Types and Heights ..................................................................................41Figure 3.8: Mobility v. Data Rate for Popular BWA Systems ..........................................................44Figure 3.9: Power Consumption of Some Wireless Radio Products ...............................................45Figure 3.10: UNHCR Refugee Camps in Tanzania ...........................................................................46Figure 3.11: The KTTC Computer Laboratory ..................................................................................47Figure 3.12: KTTC’s Eco-Friendly Power Plant ...............................................................................47Figure 3.13: The Mtabila Camp VSAT and PV Systems ..................................................................48Figure 4.1: A Simplified Broadband Value Chain ..........................................................................55Figure 5.1: From Silos to Layers ..................................................................................................76Figure 5.2: From VLF to EHF .......................................................................................................79Figure 5.3: Changing the Technology Paradigm ...........................................................................79Figure 5.4: Globally Harmonized Spectrum: IMT-2000 ................................................................85Figure 5.5: BWA Coverage Areas (Circles) in Ireland ....................................................................87Figure 5.6: The 2.5-2.7 GHz BWA Allocation in Mauritius ............................................................89Figure 6.1: Shades of Grey ..........................................................................................................93

229


LIST OF TABLES

LIST OF TABLES

Table 1.1: Newly Created Regulatory Authorities, 2005 .................................................................... 15Table 3.1: Relative Functionalities of Broadband Wireless Access Technologies ............................... 33


230

Box 1.1: Broadband Wireless Nations ...............................................................................................7Box 2.1: Wi-Fi Beyond Hotspots ................................................................................................... 25Box 3.1: A Radio Transmission Primer ........................................................................................... 34Box 3.2: TD-SCDMA – A Chinese Standard ................................................................................... 35Box 3.3: 802.16 Extensions in the Works ....................................................................................... 36Box 3.4: The Evolution of a “Southern” Solution ............................................................................ 36Box 3.5: Mesh Networks ...............................................................................................................37Box 3.6 Convergence of WLAN and 3G Networks .......................................................................... 39Box 4.1: GSR 2004 Best Practice Guidelines ............................................................................52-53Box 4.2: The Impact of Broadband Access – Examples in Developing Countries ............................. 55Box 4.3: Regulatory Methods To Boost Deployment ........................................................................57Box 4.4: Licensing Incentives for Network Deployment, selected examples .................................... 59Box 4.5: Unified Licensing Frameworks, selected examples ............................................................61Box 4.6: Satellite Broadband in Developing Countries, selected examples ...................................... 62Box 4.7: Local Loop Wholesale Options ......................................................................................... 63Box 4.8: Local Loop Unbundling in Poland ..................................................................................... 63Box 4.9: Removing Network Bottlenecks in India ............................................................................ 65Box 4.10: Using the Indian Rail and Gas Facilities for Backbone Connectivity .................................. 65Box 4.11: GSR 2003 Universal Access Best Practice Guidelines ................................................. 66-67Box 4.12: Korea’s KII Project ........................................................................................................... 68Box 4.13: Use of Universal Access Funds, selected examples .......................................................... 68Box 4.14: Limiting Cross-Ownership in the EU ................................................................................ 70Box 4.15: E-Government in Vietnam ................................................................................................ 70Box 4.16: The Rural-Enlaces Project in Chile ....................................................................................71Box 4.17: Installing Internet Centres in Southern Brazil ....................................................................71Box 4.18: Encouraging PC penetration, selected examples .............................................................. 72Box 5.1: Defining Broadband Wireless Access ...............................................................................77Box 5.2: Spectrum Bands for BWA ................................................................................................77Box 5.3: Software Defined Radios, Adaptive Array Systems and Mesh Networks ............................. 80Box 5.4: Eire’s Response to BWA .................................................................................................. 82Box 5.5: Eire’s Response to BWA, Part Two .....................................................................................87Box 5.6: OFTA’s Consultation on Broadband Wireless Access Licensing ........................................ 88Box 6.1: A VoIP Primer ................................................................................................................. 92Box 6.2: The VoIP Transition ...........................................................................................................97Box 6.3: Classification of VoIP Services ....................................................................................... 100Box 6.4: Emergency Services under the North American Numbering Plan ..................................... 103Box 6.5: The FCC’s Enhanced 911 Service Order1 ....................................................................... 103Box 6.6: SPIT: A Looming Issue ....................................................................................................107Box 6.7: Threats to VoIP Networks and Publicly Available Services ................................................107Box 6.8: Defense Mechanisms against Security Attacks ................................................................107Box 7.1: Australia Telecommunications Act 1997 – SECT 117 ......................................................121

LIST OF BOXES

LIST OF BOXES

© ITU 2006

International Telecommunication UnionPlace des Nations

CH-1211 Geneva, Switzerland

First printing 2006

All rights reserved. No par t of this publication may be reproduced, by any means whatsoever, without the prior writtenpermission of the International Telecommunication Union.

Denominations and classifications employed in this publication do not imply any opinion on the par t of the InternationalTelecommunication Union concerning the legal or other status of any territory or any endorsement or acceptance of any boundary.

Where the designation “country” appears in this publication, it covers countries and territories.

ISBN 92-61-11431-8

ALSO AVAILABLE FROM ITU

PUBLICATIONS

Trends in Telecommunication Reform: Licensing in an Era of Convergence, 2004/05 (6th Edition) ..............................95 CHFTrends in Telecommunication Reform: Promoting Universal Access to ICTs, 2003 (5th Edition) ..................................90 CHFTrends in Telecommunication Reform: Effective Regulation, 2002 (4th Edition) ...........................................................90 CHFTrends in Telecommunication Reform: Interconnection Regulation, 2000-2001 (3rd Edition) .......................................90 CHFTrends in Telecommunication Reform: Convergence and Regulation, 1999 (2nd Edition) .............................................75 CHFGeneral Trends in Telecommunication Reform 1998: World Volume I ...........................................................................75 CHFGeneral Trends in Telecommunication Reform 1998: Africa Volume II ..........................................................................65 CHFGeneral Trends in Telecommunication Reform 1998: Americas Volume III ...................................................................55 CHFGeneral Trends in Telecommunication Reform 1998: Arab States Volume IV ...............................................................45 CHFGeneral Trends in Telecommunication Reform 1998: Asia Pacific Volume V ................................................................60 CHFGeneral Trends in Telecommunication Reform 1998: Europe Volume VI .......................................................................72 CHFCollection of five Regional reports (Volumes II-VI) ......................................................................................................297 CHFCollection of Regional and World reports (Volumes I-VI) .............................................................................................372 CHFThe Arab Book: Telecommunication Policies for the Arab Region, 2002 ......................................................................50 CHFThe Blue Book: Telecommunication Policies for the Americas, 2000 (2nd Edition) ......................................................50 CHF

Please contact the ITU Sales Service:Tel.: +41 22 730 5111Fax: +41 22 730 5194E-mail: [email protected]: www.itu.int

Note: Discounts are available for all ITU Member States, Sector Members and Least Developed Countries.

20

06






W O R L D

Printed in SwitzerlandGeneva, 2006

ISBN 92-61-11431-8

*28388*

TR

EN

DS

IN

TE

LE

CO

MM

UN

IC

AT

IO

N R

EF

OR

M

20

06

ITU-Trends in Telecommunication Reform 2006

Documents

Transcript of ITU-Trends in Telecommunication Reform 2006