Sign In

Communications of the ACM

Communications of the ACM

The Wireless Local Loop in Developing Regions

Wireless networks are often cited as the most economically feasible solution to the severe dearth of communications infrastructure in developing countries. In the short term, deployment of wireless networks is indeed the optimal solution to the information infrastructure gap in these countries. Many of them have embarked on this road, and a loose formula has emerged for employing fixed cellular networks for local telephone loops (in the form of a wireless local loop) and satellite transmission for long-distance and international communications. Yet, with remarkable consistency, many countries have focused solely on the installation of voice-centric networks, seemingly oblivious to the growing pervasiveness of the Internet and other forms of data communication. Predictably, most of these countries are poor and have economies in which the most pressing concern is voice communication.

We hypothesize that a suboptimal solution is for countries to hold out for their poor economies to grow before installing data infrastructure. But a better economic strategy is to incorporate data communications as an integral part of economic growth policies. The research results we outline here show that data communications should be as pressing a concern as voice communications and that this concern should inform the legal, regulatory, market, and spectrum policies of developing countries. We discuss wireless data-centric access options as they pertain to developing countries in general, referring to Kenya and Thailand—representative of Africa and Asia, respectively—as case studies.

We do not argue with the fact that voice communications is still the main driver of communications technology in many countries today, especially in the developing world. However, it is increasingly evident that even the most remote regions will—sooner rather than later—be forced to engage the rest of the world through multiple forms of communications if they are to pull themselves out of poverty and develop their economies. Increasingly, the proportion of communications that is data-oriented dominates over that devoted to voice communications (see Figure 1). The time when data becomes the driver of communications networks deployed is closer than many government officials and phone company executives imagine. That the developed world is leading this march should not obscure the fact that the Internet, for instance, has diffused to the rest of the world much faster than did the fax, which itself spread around the world quicker than previous technologies.

Ultimately, however, our goal is not to help make the case for data services in Kenya, Thailand, or anywhere else, but to demonstrate the economic and functional efficiency of incorporating data capabilities into the construction of voice networks.

Back to Top

Telecommunications in Kenya and Thailand

Although both Kenya and Thailand are developing countries, they are characterized by notable differences in the states of their telecommunications sectors, particularly their wireless networks. Mobile cellular services in Kenya cater to only a few major cities. In fact, even wired infrastructure hardly extends beyond the confines of major urban centers, leaving much of the rest of the country—the rural areas, where a majority of the population lives—bereft of adequate communications infrastructure (see Petrazzini and Kibati's "The Internet in Developing Countries" in this issue). Delivering universal service is, therefore, an overriding objective in Kenya. On the other hand, in Thailand, where wireline-based service is near universal, the overriding objective—especially with regard to deployment of wireless services—has shifted from mere basic service to mobility functionality. Mobile telephone networks (analog and digital) serve all of Thailand, where transmission towers can be seen in most parts of the country. In both countries, however, little attention has been given to the installation of data-centric infrastructure.

The Kenya Post and Telecommunications Corp. (KPTC) is the government-owned public telephone company in Kenya, the sole legal operator and regulator of telecommunications services. With the exception of Internet services (because it is the sole provider of telephone lines) and the manufacture and sale of equipment to be used on customers' premises, the KPTC monopolizes every aspect of the Kenyan telecommunications market.

In 1998, the Kenyan parliament enacted several new laws and regulations mandating the restructuring of the KPTC into three separate entities—a telecommunications company (Telkom Kenya), a postal corporation (Posta), and a regulatory authority (Communications Commission of Kenya, or CCK). These laws and regulations mandated that the CCK regulate licensing, interconnection, "public service obligations," fair competition, and operator obligations. Telkom Kenya will initially be wholly owned by the government. Eventually, through a series of incremental steps, it is expected to go public on the Nairobi Stock Exchange. While the new laws and regulations stipulate that the introduction and fostering of competition is a major objective, the implementation plan is riddled with nondeterministic, or undefined, conditions and a vague timing schedule [1].

Neglect of data requirements is an insidious side effect of the multiplicity of issues facing the country—inadequate infrastructure, a poor economy, lack of competition, and low literacy rates. The natural impetus now, as in the past, has been to work toward installing voice networks, since they seem to be needed more immediately than data services.

In Thailand, four major laws govern the provision of telecommunications services, including the 1955 Radio Communications Act, which authorized the government's Post and Telegraph Department to allocate and administer the country's frequency resources. Two other agencies—the Telephone Organization of Thailand and the Communications Authority of Thailand—are responsible for actually providing telephone services (wired and wireless).

Thailand has consistently emphasized industrial and economic development in order to migrate to the status of a newly industrialized country. As a result of the urgent need to develop a telecommunications infrastructure to support economic development, the government has relaxed some of its control. Moreover, the government is committed to liberalizing the industry by granting numerous concessions in order to motivate private Thai telecommunications companies to bring in modern technology and higher-quality service. Yet even these private operators are not permitted to provide telecommunications services outside joint ventures with government entities or build transfer-and-operate arrangements [8].

Back to Top

Wireless in Africa and Asia

There is no obvious or simple answer to the data infrastructure question. The telecommunications industry is dynamic and the din quite loud from competing vendors of the related technologies. Given this and the overwhelming variety of factors that must be considered, the resulting risk as perceived by potential investors involved in relatively unknown regions frightens them into inaction. The temptation for any investor is to always depend on the tried and tested, sometimes blatantly disregarding the particular priorities of a given undeveloped region. We chose to model wireless technologies, of all the possible local access options, because they represent the quickest and least expensive way to circumvent the dire lack of infrastructure in places like Africa and Southeast Asia.

There are several types of wireless technologies; the main ones are satellite, wireless cable, mobile and fixed cellular, and personal communications systems. In the wireless local-access arena, mobile cellular networks have been most widely deployed thus far, and several standards (analog and digital) have been deployed around the world. On the other hand, relatively few wireless local loop (WLL) networks are in service today. WLL connects subscribers to the public switched telephone network, using radio signals as a substitute for copper wire for all or part of the connection between the subscriber and the switch. It is a modification of the cellular system and may be based on mobile cellular, cordless, or proprietary technologies. All digital WLL networks are based on either time division multiple access (TDMA) or code division multiple access (CDMA) technologies. Unlike mobile cellular, WLL is a relatively new technology—in part because developed regions have more-than-adequate wired local access, rendering wireless networks commercially useful only for mobility purposes. Only with recent deregulation in the U.S. and other industrialized countries has WLL gained popularity as a viable competitor for the local loop against existing local telecommunications operators. In developing regions where wired infrastructure is scarce, WLL promises to be a very viable alternative. WLL networks have already been deployed in Eastern Europe, Latin America, and Southeast Asia.

Back to Top

Wireless in Kenya and Thailand

Thailand has up to nine cellular systems today—four analog, the rest digital. Only six of them have strong market positions; the other three are still under trial or construction. Kenya's major cities have two mobile cellular systems in service. Cordless and cellular WLL networks have recently been introduced on a trial basis but are not yet commercially available. All these systems—mobile as well as fixed—are provided by the KPTC (see Table 1).

Our objective has been to investigate the best routes leading to installation of networks that in the future will meet data as well as voice requirements. The cost model scenarios we calculated for Kenya and Thailand are different, in keeping with the two countries' differing priorities—universal access for Kenya, mobility for Thailand. Although voice services are the main motivation for developing technology today, our models anticipate a future in which data services, such as full Internet access, will be a crucial communications component. Our models were based on the fact that voice-only networks are in place throughout Thailand and in the urban parts of Kenya; there are no wireless networks of any kind—voice or data—in place in most rural parts of Kenya, though plans have been drawn to construct voice-only networks in these parts.

The central issue facing both Kenya and Thailand is whether to continue upgrading existing voice-centric networks (or in rural Kenya, build completely new ones) that may not meet future demand for high-speed data without costly upgrades, or immediately deploy multiservice (voice and full Internet access data compatible) installations.

Accordingly, we developed several cost models:

For Kenya:

  • The global system for mobile communications (GSM), a TDMA-based mobile cellular technology, the current practice in much of Africa (scenario A).
  • WLL (proprietary in urban settings and CDMA in rural settings) implementations (scenario B).

Should they continue upgrading voice-centric networks that may not meet future demand for high-speed data without costly upgrades, Or deploy multiservice installations immediately?

For Thailand:

  • The mobile cellular (GSM) implementation (scenario A) in Bangkok and surrounding metropolitan areas.
  • The mobile cellular (CDMA) implementation (scenario B) in Bangkok and surrounding metropolitan areas, facilitating a relatively inexpensive transition to high-speed data in the future.

For both countries, we modeled GSM mobile networks to represent scenario A, since they are the most widely deployed digital networks. Advanced mobile phone service and GSM are the most widespread cellular communication systems in developing countries, including Kenya and Thailand. Almost 89% of digital cellular networks in Africa are GSM (for details, see While GSM systems ideally support data applications up to 14.4Kbps, this speed is rarely achieved. In today's telecommunications environment, and certainly in tomorrow's, 14.4Kbps should be viewed as the bare minimum data speed for sending and receiving email without large attachments or graphics. However, for major information exchanges, including those involving large amounts of data, graphics, and charts, and for browsing the World-Wide Web, faster data speeds are essential.

In scenario B, we modeled different technologies for each country. For urban Kenya, we proposed deployment of a proprietary technology using, for example, DSC Airspan, a CDMA-based technology that can handle data speeds up to 128Kbps; it was developed by the former DSC Communications Corp., which is now part of Alcatel. For rural Kenya, we modeled a fixed CDMA (IS-95B standard) cellular system. For Thailand, we considered Bangkok and surrounding metropolitan areas and proposed a CDMA mobile cellular system.

In selecting the technologies to model, we took into consideration not only the data capabilities and ranges of the networks (see Table 2), but also the likely costs of a transition to the next generation of wireless networks. A protracted debate has developed between GSM vendors backing a third-generation mobile standard known as wideband-CDMA (WCDMA) and CDMA vendors backing the CDMA2000 (formerly narrowband-CDMA) standard. WCDMA is based on a layered network-protocol structure, similar to the protocol structure in GSM networks; CDMA2000 is based on CDMAONE, the current CDMA standard. Despite a recent agreement between the principle proponents of the two camps—Ericsson (for WCDMA) and Qualcomm (for CDMA2000), the debate continues as to which promises the easiest transition—GSM to WCDMA or CDMAONE to CDMA2000. Nevertheless, it appears that while CDMA2000 will be backward compatible with CDMAONE, since they share the same air interface, the same backward compatibility will not hold true for GSM and WCDMA. All base-station equipment will have to be replaced to manage a shift from GSM to WCDMA. The counterargument advanced by the WCDMA camp is that third-generation subscriber handsets will be dual mode, so compatibility will not be an issue. Dual-mode handsets will of necessity be more complex and consequently more expensive to construct and maintain. While the jury is still out on this issue, the switch to the next generation of wireless networks appears less certain for GSM than for CDMA.

While our models considered only equipment installation and maintenance costs, we also had to account for technical suitability to various population densities and desired penetration levels ("teledensity"). We modeled only major network cost elements.

Our main finding was that although it is difficult to determine the veracity of capacity and cost figures, given all the marketing hype, our most conservative estimates (for CDMA) still show that CDMA is more cost effective than GSM (see Table 3). The higher capacity of CDMA cell sites more than compensates for the higher costs of CDMA electronics. Given CDMA's superior data handling, it would be prudent to deploy CDMA, rather than GSM networks.

For example, a proprietary CDMA network (such as DSC) would be ideally suited to urban Kenyan WLL implementations, since it would provide adequate data functionality immediately. GSM's wider range would offer no special advantage, because the cost driver in a high-density urban center would be cell-site capacity. The only trade-off would be mobility function. However, since the goal of universal service is far more important than the need for mobility, lack of mobility would appear to be an acceptable trade-off in most parts of urban Kenya.

A CDMA cellular-based network would be ideal for WLL implementations in rural Kenya because of its greater capacity. Despite the higher costs of CDMA equipment, CDMA's greater capacity means it can cover larger areas in low-density regions like rural Kenya, making its costs highly competitive. At 13Kbps, CDMA's data capacity, while marginal, is sufficient for rural Kenya's requirements today and would help prepare for a more seamless shift to future broadband functionality. The goal of universal access to communications service overrides the need for mobility.

For Bangkok, mobile CDMA cellular networks, much like fixed CDMA, would position the city for eventual transition to broadband networks.

Back to Top


Because data is a growing component of communications worldwide, developing regions will eventually have to upgrade their infrastructures to accommodate data communications. For these poorer regions, wireless networks represent a more appealing option than conventional copper wire, as they are relatively inexpensive to install and maintain and can be deployed quickly. Our cost model demonstrates it would be prudent to incorporate data capabilities in any new installations and to adopt technology that will readily support the shift to the next (third) generation of wireless networks—although the pace of this shift depends on a country's particular circumstances. In African countries, of which Kenya is typical, proprietary CDMA-based networks would be suitable for urban centers, while fixed CDMA networks would be ideal for rural areas. For urban Thailand, on the other hand, mobile CDMA networks may be the technology of choice. (Please note the results we derived for Thailand were influenced by the lack of an explicit government policy on universal telephone service; were a more explicit policy to be formulated, we would have to offer different recommendations for rural Thailand.)

So, we recommend that data capabilities be incorporated in any new installations and technology selected carefully in anticipation of the transition to the next generation of wireless networks. Meanwhile, local priorities should prevail and a speedy privatization and liberalization process should be followed.

Our models also showed that CDMA is cost-competitive with TDMA. For the moment, the path from CDMA to third-generation networks appears more technologically feasible than the path from GSM. However, no conclusive evidence of the superiority of any particular technology has yet been produced.

Back to Top


1. Ayah, W. Postal and Telecommunications Sector Policy Statement. Kenya Ministry of Transport and Communications, Jan. 1997.

2. Fafunwa, T. A Failure to Connect: Africa's data communication opportunity. In Proceedings of the Africa Telecom'98 (Johannesburg, South Africa, May). International Telecommunication Union, Geneva, Switzerland, 1998.

3. Krairit, D. Mobile communication industry in Thailand. In Worldwide Wireless Communications, F. Barnes, Ed. IEC Press, Chicago, Ill., 1995.

4. Mutooni, P, and Tennenhouse, D. Modeling the Communication Network's Transition to a Data-centric Model. Harvard Infrastructure Project. Harvard University, Cambridge, Mass., 1997; see

5. Thai telecommunications business. Telcom J. (1997).

7. Webb, W. Introduction to the Wireless Local Loop. Artech House Publishers, Boston, 1998.

8. World Telecommunication Development Report. In Universal Access World Telecommunication Indicators, International Telecommunication Union, Geneva, Switzerland, 1998.

Back to Top


Mugo Kibati ( is a graduate student in the Technology, Management, and Policy Program and a research assistant in the Internet Telephony Consortium, Center for Technology Policy and Industrial Development at the Massachusetts Institute of Technology in Cambridge, Mass.

Donyaprueth Krairit ( is a doctoral candidate in the Technology, Management, and Policy Program and a research assistant at the Internet Telephony Consortium, Center for Technology Policy and Industrial Development, Massachusetts Institute of Technology in Cambridge, Mass.

Back to Top


F1Figure 1. Projected international telecommunications traffic.

Back to Top


T1Table 1. Cellular and WLL networks in Thailand and Kenya.

T2Table 2. Characteristics of the modeled networks.

T3Table 3. Estimated costs per subscriber, Kenya and Bangkok, Thailand (in U.S.$)

Back to top

©1999 ACM  0002-0782/99/0600  $5.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 1999 ACM, Inc.