 Civil Society > Too Much Global?
Connectivity model for the developing world builds on NSF Net’s success
Special Feature: Wireless Grand Challenge
By Larry Press
Footnote references in the following article may be viewed in the The Weekly Op-Ed section at W2i.org.
At the July 2000 G8 Summit in Okinawa, leaders of the world’s largest economies focused on information and communication technology, or ICT. They expressed concern that the digital divide might exacerbate existing inequalities between countries, but also held out hope that ICT could play a significant role in development. The heads of state formed the Digital Opportunity Task Force[25] (DOT Force) “to identify ways in which the digital revolution can benefit all the world’s people, especially the poorest and most marginalized groups.” Most notably, the Japanese government pledged a total of $15 billion over five years in working toward that end.
The DOT Force is no more, and the $15 billion did not materialize[26], but the pledge gets one thinking. What could we do with $15 billion? Let us consider one possible “Grand Challenge”—to connect every rural village to the Internet by 2015.
After a decade of evangelism and work, there is virtually no connectivity in rural villages of developing nations[29]. Because a business case cannot be made to attract capital to connect these areas, subsidies are needed. There is ample precedent for subsidizing communication. Telephone franchisees and taxpayers are often required to subsidize connectivity for rural, poor or handicapped people as well as emergency-response services and the ability to provide assistance to law-enforcement authorities.
NSFNet
It may be possible to employ the US National Science Foundation’s (NSF) strategy from the 1990s to subsidize Internet connectivity for research and education institutions. The NSF began by building the NSFNet backbone network in 1984[30], and by 1988, 13 backbone nodes linked supercomputer centers and regional networks in a 1.544 Mbps network, which was later upgraded to 45 Mbps.
In 1990, NSF began offering grants for connecting four–year institutions of higher education. Schools could apply for $20,000 in connection assistance, typically a router and a link to a regional NSFNet point of presence, or POP. NSF also made grants to connect foreign research and education networks to the NSFNet, eventually linking 28 research and education networks in 26 countries.
NSFNet was a high–return investment. It became the first global Internet backbone at a cost of less than $100 million to the US taxpayer[31].
The NSF strategy was to build common infrastructure—a backbone network—and then offer a subsidy for connection to that infrastructure. It was a heavily leveraged investment. The aggregate cost of staff, equipment and installation of university local area networks far exceeded the cost of the NSFNet program. The NSFNet was a terrific investment.
Could a similar approach be applied to connecting villages? Could we build a network that brought a high–speed POP to every village on earth?
There is already fiber connectivity in most large urban areas, but rural backhaul is difficult in developing countries. Villages are in remote locations, and the infrastructure that facilitates the construction of backhaul cables—roads, railroads, pipelines—is poor or nonexistent.
Because global backhaul to villages will require wireless as well as cable links, let’s consider three wireless technologies that could be used to supplement wired links where necessary: terrestrial wireless, high–altitude platforms and satellite.
Wireless technologies
Patrick Gelsinger, Senior Vice President and Chief Technology Officer at Intel, advises developing countries to cross the digital divide and leapfrog developed nations using terrestrial wireless technology[33]. Gelsinger urges developing nations to say “no” to more copper, and deploy fiber backbone and wireless technology aggressively. He envisions Wi-Fi LANs with forthcoming IEEE 802.16 (WiMAX, Worldwide Interoperability for Microwave Access) for backhaul to fiber[34].
IEEE ratified the WiMAX standard in March 2003, and early chips and products are just beginning to appear. 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—point–to–point, point–to–multipoint and point–to–multipoint—plus a neighborhood mesh. Intel and others have formed the WiMAX Forum in the hope that interoperability certification and joint marketing will drive volume up and price down as they have for Wi-Fi LANs[35].
Proprietary point–to–point links are currently available for backhaul to fiber, but a standard, commodity WiMAX should lower cost significantly in the long run. Still, there would be problems. WiMAX promises NLOS coverage, but even Intel’s marketing department must bow to the laws of physics. Vegetation, inclement weather and other obstacles will degrade performance. Networks must be carefully planned to determine tower height and placement.
Planning might be facilitated with a program like Radio Mobile[36], which uses Shuttle Radar Topography Mission[37] from NASA and transmitter characteristics to generate predicted coverage maps. Using GIS data on roads, vegetation and climate data along with a program like this one might facilitate network planning. Mapping and planning using tools like this would be a major activity in this Grand Challenge project.
Even with tools like this and commodity WiMAX equipment, there will be areas were terrestrial wireless connectivity is not practical. Another Grand Challenge activity is the design and investigation of high altitude platforms, or HAPs.
HAPs are one alternative to terrestrial backhaul links. A typical HAP would hover between 17 and 22 km above the ground. This altitude is out of the way of aircraft, and wind speed is relatively low. A radio on a HAP would have a footprint from 60 to 400 km in diameter depending upon platform altitude and ground antenna inclination. Altitude would be chosen to balance factors like wind speeds, attenuation due to free space distance and rain, signal latency time, and coverage. A HAP could have optical or radio links to other HAPs or to satellites in addition to communicating with ground stations.
Research and development on a variety of system architectures is being conducted in Europe, the United States and Japan[38]. The platforms are generally unmanned and solar powered. They might be balloons, airships, planes or tethered aerostats. If you are thinking “Hindenburg,” bear in mind that materials and technology have improved since that disaster. For example, there are new plastic envelope materials that are strong, UV–resistant and leak–proof to helium, which is not volatile.
A practical HAP would have several advantages relative to terrestrial or satellite links or in some cases to either.
• Cost would be lower than satellite;
• Bandwidth and throughput would be greater than satellite;
• Free space attenuation would be less than with satellite;
• Platforms can have multiple uses in addition to Internet connectivity: remote sensing, weather monitoring, non–Internet communication, etc.;
• Latency times would be lower than satellite, eliminating problems with TCP[39] and with isochronous applications like telephony;
• Rain interference would be less than with a terrestrial link;
• Foliage, rough topography and vandalism would not be issues as they are with terrestrial links;
• Faster to deploy than satellite;
• Easier to maintain than satellite;
• Incremental deployment is possible; and,
• If the HAP is tethered, the tethers may be used for power and data.
This all sounds good, but HAPs have not yet been proven practical[40]. Power and power storage systems would have to function well during summer and winter. A HAP would need to be kept in a relatively fixed location and orientation, particularly if it is desired to use fixed–position antennae, so systems to maintain orientation will have to be designed and powered. Modulation schemes, spectrum utilization, technology and standards for inter–HAP links must all be designed. Regardless of these hurdles, this technology should be watched and considered as part of a global connectivity toolbox.
While HAPs are unproven, Internet traffic is traveling over geosynchronous satellite links today. According to the Global VSAT forum, very small aperture terminals (VSATs) have been in use for more than a decade, and there are now more than 500,000 systems operating in more than 120 countries[41]. Many of these are in developing nations, and ground stations are relatively easy to deploy.
However, geostationary satellites are limited in number and expensive, and they take a long time to design and deploy. Their distance above the earth results in high latency times.
To some extent, these problems will be mitigated by new technology and architectures. For example, Lloyd Wood and his colleagues argue in favor of breaking the “bent pipe” system architecture in which the satellite is seen as a relay point between two ground stations[42]. They would add inter–satellite communication links and routers to decouple the up- and downlink locations. Moore’s Law will make this and processing and transmission improvements economically feasible in the future.
But geostationary satellites will always be expensive to launch, and be far away. Low earth orbit (LEO) satellites overcome the latency problem, but introduce their own. A single LEO satellite is only visible during a limited window of time from any point on the earth. That would allow store-and-forward applications such as e–mail, but interactive applications are not possible[43]. The solution to that problem is a satellite constellation in which every point on earth is visible to at least one satellite at all times, and they are equipped with inter–satellite communication links and routers[44].
After several bankruptcies, it is clear that the business case for LEO constellations cannot be made today. For example, Teledesic, which is now bankrupt, had planned to offer 64k to 2 mbps IP service by connectivity by 2005. Had they succeeded, Teledesic would have had a major impact on village connectivity in developing nations and elsewhere.
Such a constellation is not economically viable in today’s market with today’s technology, but both of those conditions are subject to change. The technology change is certain. Advances in electronics will reduce the cost of complex, dynamic routing algorithms and radio signal processing. Vehicle and launch technology will also change with engineering advances and variety in commercial enterprises. For example, private companies are competing for the $10 million X PRIZE offered to encourage development of reusable vehicles for space tourism[45].
Even if technology were not to improve, we are talking about a Grand Challenge, and we can follow the NSFNet model by subsidizing the design, construction and launch of a satellite constellation and ground station acquisition and link costs. As with the NSF, the subsidies might be phased out at some point.
|
|
Download this issue today!
