Dozens of interesting experiments are taking place around the world using WiFi to provide voice and data connectivity in areas of the world where it's difficult to build conventional infrastructure, from Inveneo's work in Uganda with VOIP over WiFi to Nepal Wireless's project to provide connectivity in the Himalayas. Wireless hackers like working with WiFi because the hardware is cheap, the frequency WiFi broadcasts on (2.4Ghz) is usually unlicensed, and because brilliant antenna hackers have been able to send WiFi signals over 100 miles.
From a technical perspective, WiFi probably isn't the best choice to transmit data over long distances. WiFi is ultimately limited by line of sight considerations - barriers like mountains, buildings and forests can block signals. And WiFi signals can be affected by moisture in the atmosphere... which helps explain why the world WiFi record was achieved in the dry, flat Nevada desert.
To send data over longer distances without needing line of sight, engineers use lower freqencies than the microwave frequencies used by WiFi. Generally speaking, you can't send as much data at lower frequencies than at higher, but you can send the signals farther. And while netheads in the US and elsewhere are getting excited about the possibility of "fiber-speed" broadband data over wireless at very high frequencies, there's exciting developments in long-distance, lower-bandwidth "narrowband" communications in the developing world.
Arrow Networks, a Ghanaian company that distributes hardware built by Czech company RACOM, is using hardware originally built to connect ATM machines and lottery terminals, and to enable remote monitoring of oil and water pumps, to build a national wireless data network. While the network isn't nearly as fast as WiFi, it can provide modem-speed or better data connections at low cost, over long distances, without line of sight. Arrow reports that their networks can provide 19.2kbps connectivity (about the speed of a slow telephone modem) over distances of 50 -300km, for about $50 US dollars a month.
While the networks Arrow Networks is building are decidely narrowband, not broadband, they may be a useful solution for schools that are far off the phone grid as an alternative to VSAT (satellite) connections to the Internet. Arrow has designed a cacheing server called Javelin which is designed to let a school or business use a narrowband connection to provide email accounts for employees and students and to download and store web pages for browsing on a local network. Arrow is already experimenting with a network at OLA Training College, a women's college in Cape Coast. CEO Kwaku Boadu is raising money to build a broadband and narrowband network as part of a national wireless infrastructure for the nation, using radios that can transmit at speeds from 128kbps to 8mbps (a speed near that of WiFi).
It's unclear whether the low cost and long distance of these solutions will make them popular in Africa despite the limitations of due to low bandwith - Arrow makes it very clear that one wouldn't want to try to run a cybercafe with the modest bandwidth possible with the technology. Should this prove an effective solution to African bandwidth problems, Ghana will benefit in two ways, as Arrow has announced plans to manufacture hardware designed by RACOM in a Ghanaian export zone for sale domestically and across the continent.
$50 a month doesn't seem that cheap. Also, some of the most useful applications aren't going to work well on this - VoIP, telemedecine, etc.
A more enlightened government could allow companies to use a larger portion of the spectrum- nowadays there is a lot less interference due to better equipment (both sending and receiving signal). IIRC, higher frequencies also do not have as many problems with line of sight issues.
$50/month does seem high, but this may be the least expensive alternative (compared to sat, for ex.). But that bandwidth is brutal. I'm not up on cellular tech any more, but aren't EVDO, EDGE, and GPRS services in the megabit range? I know there are 3G gateways that make a cellular internet connection available to PC's as a regular network service, wouldn't using cellular infrastructure make more sense than the Arrow Networks services? Given the penetration of cellular in Africa, there's probably at least a marginal business case to keep infrastructure current for selling to consumers; then offer net access for schools to free as a nice humanitarian gesture.
The Ghanaian government is pretty progressive, spectrum-wise, Daniel. The willingness to allow widespread commercial, long-distance use of WiFi puts them ahead of a number of other governments on the continent. There's extensive use of long-distance microwave technologies with government licenses, but hardware like Motorola's Canopy system is pretty expensive in local terms. Not sure you're right on the frequency issues - my understanding is that higher frequencies tend to propogate shorter distances... and that very low frequencies propogate great distances by bouncing off the ionosphere, like AM and shortwave radio...
Rod, 3G is a good direction to be thinking in, but it's not reality yet except in central urban areas in much of Africa. I know that one of the Ghanaian phone companies was putting in GPRS, which has pretty good data transfer speeds, though certainly not broadband speeds. And, if I recall correctly, the effective data rate falls off rapidly as you move away from the towers. Most of Ghana has good cell coverage, by African standards, but lots and lots of people aren't within a cell tower footprint. I think this is why folks are excited about UHF technologies, even if the data speeds are far less than we expect in the US/Europe.