I recently had the privilege of attending a one-day conference/workshop on the wireless Internet opportunity for developing nations . The conference was co-sponsored by the UN Information and Communication Technologies Task Force  and organized by the Wireless Internet Institute , a Boston "Think Tank."
The conference began with plenary presentations followed by breakouts in which sub-groups worked on various questions. The breakout results were then reported back to the entire group, and recommendations developed.
The conference opened with the reading of a greeting statement from Kofi Annan acknowledging the swift emergence of a global information society, and predicting that wireless technology will have a key role to play in reducing the digital divide. It is his hope that "With considerable speed and without enormous investments, WiFi can facilitate access to knowledge and information, for example by making use of unlicensed radio spectrum to deliver cheap and fast Internet access."
The keynote presentation was by Patrick Gelsinger, Senior Vice President and Chief Technology Officer at Intel. He said the number of WiFi hot spots in hotels, restaurants, airports, convention centers, shopping malls, etc. has grown rapidly because WiFi LAN technology uses unlicensed spectrum, standards and competition have driven cost down, and it is "broadband." He estimated the cost of serving a 50-megabyte/month WiFi user at US$2 per month versus 10-15 times that using 3G cellular data service.
Mr. Gelsinger feels developing nations can cross the digital divide and leapfrog ahead of developed nations using wireless technology. He urged developing nations to say "no" to more copper, and deploy fiber backbone and wireless technology aggressively. He envisions WiFi LANs with forthcoming IEEE 802.16a (WiMAX, Worldwide Interoperability for Microwave Access) for backhaul to fiber and IEEE 802.20 for mobile links (see Figure 1) .
Figure 1: Wireless connectivity architecture for developing nations. Wireless LANs will connect to fiber networks via point-to-point WiMAX wireless links.
IEEE ratified the WiMAX standard in March 2003. It promises non-line-of-sight (NLOS) coverage at speeds up to 70 Mbps over a distance of 50 kilometers, in both licensed and license-exempt bands.
Three architectures are envisioned. The simplest is point-point with directional antennas between nodes. This would be used for backhaul in areas where fiber was not available. The second is point to multipoint, in which a base station communicates with many customer premises radios, which could connect a LAN in an Internet café, organization, home, etc. This would be used in a rural area in a developed nation or an urban area in a developing nation . The third configuration is point-multipoint, plus small neighborhood mesh networks (see Figure 2). This configuration has the potential of extending NLOS coverage, but wireless ISPs may have concern over lost revenue as their customers share connectivity with others.
Figure 2: Point-Multipoint Plus Mesh. The WiMAX standard supports three architectures, but equipment manufacturers have not yet demanded the mesh extensions illustrated here .
Fujistsu and Intel have committed to producing WiMAX silicon  and Nokia is a long-time advocate. They are members of the WiMAX Forum , which will certify interoperable products. Margaret LaBrecque is Marketing Manager, Broadband Wireless Access Business at Intel and President of the WiMAX Forum. She says Intel plans to deliver chips this year, and expects to see interoperable, certified equipment from multiple vendors in the second half of 2004 . Equipment manufacturers have not requested silicon support for the mesh architecture, so Intel will include hooks for software support only. (That sounds like an opportunity for an open source project). Ms. LaBrecque expects non-mesh customer premises equipment to cost roughly US$350, but hopes that competition around an open standard will lead to rapid price decreases. Base station prices will vary with higher performance radios designed for licensed bands costing more than lower-performance models aimed at the license-exempt customer. We will see products geared to large carriers and to small, independent wireless ISPs.
In addition to deploying wireless and fiber technology, Mr. Gelsinger urged nations to liberalize spectrum regulation and to adhere to global standards in order to be able to use mass-market products. The International Telecommunication Union World Radiocommunication Conference just established frequency allocations for wireless access , and Mr. Gelsinger pointed out that developing nations have been slower than developed nations in adopting global recommendations for license-exempt spectrum. This may generate revenue in the short term, but he feels it is short sighted.
Mr. Gelsinger stated that Intel is dedicated to connecting all people of the world to the Internet, and closed his talk with a quote from Intel co-founder Robert Noyce urging us not to be encumbered by history, but to do something wonderful.
Iqbal Quadir also gave a noteworthy plenary talk. Mr. Quadir is currently at Harvard University and was the founder of Grameen Phone . He teamed with the Grameen Bank to bring cellular service to Bangladesh. Grameen Bank had worked for many years in Bangladesh, making very small loans to would-be entrepreneurs. The majority of Grameen borrowers are landless rural women, and a typical micro loan might be used for the purchase of a cow . After approximately a year, the borrower would make sufficient profit selling milk to pay off the loan and own the cow. Mr. Quadir felt the same approach could be used to invest in cellular phones.
Mr. Quadir pointed out that telephones and other communication technology have a large marginal effect in a developing nation, and quipped that Bangladesh was already wireless when he began organizing Grameen Phone in 1993. Today Grameen Phone has nearly US$300 million invested and 2002 net income after tax was US$44 million. It is the largest telephone company in Bangladesh with close to a million urban subscribers and community phones serving 45 million people in 30,000 villages. A village phone generates a profit of US$2 per day on revenue of US$100 per month. This is more than twice the per capita income in Bangladesh.
Having adapted Grameen Bank’s approach to micro credit to telephone service, Mr. Quadir is now looking toward Internet service. He says every village in the world has a hut and wonders why every village could not have a "Hut Spot."
After these and several other plenary talks, the attendees broke into small groups focusing on specific issues like regulation, business applications and the role of incumbent carriers.
Inputs from these sessions were pooled to come up with the following recommendations:1. Increase awareness of the benefits of wireless internet for developing nations and underserved areas by:
a. Ranking nations with W2I readiness index
b. Promoting success stories
c. Comparing cost structure of wired infrastructure vs. wireless internet connectivity per user
d. Evangelization, supporting international conference with speakers engagement
2. Fostering consensus among existing and potential wireless internet stakeholders
a. Regional and country specific workshops
b. Sector workshops (municipalities, transportation, power and utility companies etc.)
c. Brainstorming format based upon local conditions
d. Involving regulators, governments, entrepreneurs, end users, technology vendors
3. Supporting field practitioners by organizing best practice sharing
a. Web-based component
b. Repository of case studies
c. Conference calls
4. Providing on going analysis of field developments and feedback to technology and services providers, regulators, international institutions and other underwriters of the program
a. Quarterly updates
b. Workshops and teleconferences
c. Specific surveys
5. "Adopt a community Hot Spot" program to provide seed funding to innovative applications in developing nations and underserved areas
a. Cooperation between W2i, UN Fund for International Partnerships, UN ICT Task Force and private sector
b. Seed financing mechanism for entrepreneurs and communities to jump start infrastructure deployment
c. High profile involvement and PR opportunity
I began by stating that it was a "privilege" to attend this meeting because it was organized by and held at the United Nations. Yet, at the end of the day, in spite of the significance of the United Nations, the meeting was unremarkable. There were many good people, saying things I mostly agree with, but there was nothing really new. This would have been an eye-opening conference ten years ago, but how many reports have we already read stating that telecommunication regulation should be liberalized and that IT offers great promise in education, health care, government services, entertainment, news, commerce, etc.?
There was an air of fuzzy optimism rather than hard-headedness. For example, much of the discussion focused on WiFi as a leapfrogging technology. When Kofi Annan’s greeting message was read, we were told that as soon as he heard about WiFi, he ordered that it be included in his next speech. While WiFi works fine in my home, it is designed for local area networks, but local connectivity is not a constraint in rural villages. Rural villagers can connect their "hot huts" with cable, and have often done so to distribute television signals (see Figure 3).
Figure 3: Rural LAN. Connectivity for television within villages is often achieved with ad hoc cabling, as in this village in Pondycherry, India.
Backhaul is a more difficult problem. Mr. Gelsinger argued that WiMAX will solve the backhaul problem, and no attendees seemed to question this assertion . I truly hope he is correct, but it is surely too soon to say. As the Shosteck Group states "At this point, its potential is still unexplored, and the possible applications of IEEE 802.16 will depend on a wide range of political, economic, and technology factors yet to be decided" .
Vegetation, mountains, and other obstructions will cut performance in NLOS situations, as will noise in crowded urban areas. Licensing, vandalism and electrical power will be problems in some cases. Grameen’s cell phone coverage in Bangladesh has tended to follow major roads, and reaching remote villages with WiMAX will be difficult (See Figure 4).
Figure 4: Roads and cellular coverage in Bangladesh. Terrain and complementary infrastructure determine the difficulty of deployment of wireless equipment.
Even if WiMAX turns out to be extremely successful for backhaul or point-to-multipoint connectivity, network management and scalability will be difficult. While I was in New York for this conference, I tried to connect to the Internet using both a laptop and a WiFi-equipped PDA at pay-phone hotspots Verizon has deployed throughout Manhattan . In tests at four hot spots, I only succeeded in connecting once, and then I failed the authentication challenge in spite of the fact that I am a Verizon DSL customer. If Verizon is having system difficulties in Manhattan, imagine the management problems in a developing nation.
There was also an air of faith regarding unlicensed spectrum. While interference may not be a large problem in a remote village, dense urban areas are a different story. I hope we one day develop of-the-shelf mesh networks, establish a metric for "interference temperature," and design frequency agile, smart radios capable of identifying unused spectrum, adjusting power levels, etc., but, that has not yet happened. I was able to detect four or five competing WiFi base stations from the apartment I stayed in during the conference in New York . WiFi was designed for a LAN in a controlled environment, not New York or New Delhi apartment buildings.
Applications are another issue. The conference attendees all assume there are important network applications that would improve the quality of life in developing nations. While I share their belief, it must still be considered a hypothesis . To date, most of our evidence is from pilot projects, and there have been apparent failures along with apparent successes.
The conference recommendations listed above are worthwhile, but not novel — many workshops and studies have reached similar conclusions. The recommendations are also timid.
Sarah Chayes recently wrote of "international donors’ obsession with quick-impact projects, known as 'quips,' that didn’t cost much and wouldn’t be much of a loss if they failed" . She is in Afghanistan, and suggests the reconstruction of the Kabul-Kandahar road, the road to Urozgan, and the American-designed Helmand Province irrigation system as examples of major projects that would make a significant difference. For example, she feels "A good road to Kabul would make all the difference to Kandahar’s merchants, and jumpstart a whole region’s economy," but states that donors are "loath to commit their money to big projects like these."
Are the United Nations and other donors being too timid? Might it be time for some large networking projects rather than demonstration projects, e-readiness studies, publication of best-practices databases, etc.? Three years ago the G8 nations pledged to identify and implement means of eliminating the digital divide . The G8 nations and the international community have the resources to pursue grand challenges like universal access to scientific and technical journals and data sets or low-earth satellite networks for backhaul. Such projects cannot be justified in the marketplace, but neither could space travel. After a decade of "quips" and conferences, perhaps it is time to identify and undertake the sort of large project advocated by Sarah Chayes.
About the Author
Larry Press is Professor of Information Systems at California State University, Dominguez Hills.
4. See http://www.ieee802.org/16/ and http://www.ieee802.org/20/.
5. The Broadband Wireless Exchange, http://www.bbwexchange.com, currently lists over 4,000 wireless ISPs serving mostly rural areas in the U.S. (where DSL and cable Internet are not available) and urban areas in developing nations. They are using pre-802.16 equipment from various manufacturers, and presumably many will convert once standards-based equipment is available. This market is developing from the bottom up.
6. This figure is from Nokia Networks and was shown in an IEEE Distinguished Speaker presentation by Roger B. Marks, April 2003, http://www.ieee.li/pdf/viewgraphs_wireless_802.pdf.
7. Patrick Mannion, "Intel stokes fixed-wireless fire," EE Times (14 July 2003), at http://www.eetimes.com/story/OEG20030714S0008.
8. See http://www.wimaxforum.org/.
9. Telephone interview, 24 July 2003.
12. See Larry Press, "Connecting Villages," OnTheInternet, volume 5, number 4 (July-August 1999), pp 32-37, and at http://som.csudh.edu/fac/lpress/articles/villages.htm, for more on Grameen Bank and Phone.
13. Intel’s bullish views were amplified in a presentation on "The Wireless Opportunity," by Sriram Viswanathan, Managing Director, Intel Capital at the WCA Symposium, 13 January 2003. Slides from the presentation are available at http://www.ieee802.org/16/docs/03/C80216-03_02.zip.
14. Shosteck Group Pulse, "Wireless MAN: Another Threat to Licensed Operators?" (May 2003), http://www.shosteck.com/news/jun03.htm.
16. I even sniffed out a Verizon hot spot sporadically at night.
17. For a statement of this hypothesis, see Larry Press, "The Role of Computer Networks in Development," Communications of the ACM, volume 39, number 2 (February 1996), pp 23-29. http://dx.doi.org/10.1145/230798.230800
18. Sarah Chayes, "Afghanistan’s Future, Lost in the Shuffle, " New York Times (1 July 2003), http://www.nytimes.com/2003/07/01/opinion/01CHAY.html.
19. See Larry Press, "Is it Time for the G8 Manhattan Project? " OnTheInternet (January-February 2002), at http://www.isoc.org/oti/articles/1201/g8.html.
Paper received 30 July 2003; accepted 18 August 2003.
Copyright ©2003, First Monday
Copyright ©2003, Larry Press
Wireless Internet connectivity for developing nations by Larry Press
First Monday, volume 8, number 9 (September 2003),