African economic growth has been impeded for many decades by government-imposed restrictions on telecommunications. This was crippling when the continent was underpopulated and competing with the U.S. and Europe. Now that the rest of the world is communicating via Internet, the population of Africa has exploded, and the competition is with India and China, the denial of telecommunications to Africans is an even worse drag on the economies of the region. With an increased population and a fixed amount of land, people in African can no longer remain on their farms, but when they crowd into cities such as Lagos or Nairobi, the results are not happy.
Foreign aid hasn’t worked very well. The majority of the aid is siphoned off by officials, bureaucrats, and NGO staff who live in $5,000/month houses and maintain an upper class American lifestyle in the midst of desperate poverty. Such aid that does get through is often counterproductive (source).
What if instead of sending aid to Africa, we sent a network of telecommunications that was priced competitively with Internet service in the rich world? For $25 per month, any African could obtain 3 Mbits/second. The network would be available everywhere on the continent, thus reducing the pressure on people to move into the crowded cities. A space-based network would be the only way to achieve universal service and bypass the telecom monopolies that plague many of these nations.
With a universal network priced at $25 per month, the monopolies would be forced to cut their prices and improve their service (currently South Africa’s telecom prices transmission through their cable to London at 100X the cost of other undersea cables). Presumably the residents of big cities would be served with faster cheaper service from DSL, fiber, or cable modems. That would not render the satellite-based system obsolete, however, as it would be useful to people in remote areas or in countries afflicted by war.
What would it cost to do this? Iridium, an old and slow network, cost approximately $5 billion to build and launch. http://en.wikipedia.org/wiki/Teledesic promised roughly the same kind of performance that would be required to serve Africa. It was set to cost $9 billion at the time that it was scrapped.
Total foreign aid since 1960, from all the rich countries combined, has been approximately $2.6 trillion, or 260 times the cost of building Teledesic. Building a system like this fits in nicely with the traditional First World concept of foreign aid: we give money to a poor country, but insist that it all be spent on products built in the donor country.
Africans would actually be grateful, I think, for such a system. All of Tanzania, 35 million people, currently share about 30 Mbits of Internet connectivity, the same amount of bandwidth that Verizon offers to American families in their homes via their FiOS service.
I don’t know how many African families are going to be willing to spend $25/month when the per capita GDP is $1500, in Kenya for example. At best, reliable internet connectivity may benefit any companies that are formed or choose to set up shop in those countries, which in turn would benefit Africans enormously.
Isaac: 3 Mbits would be enough for an entire village! Africans have no trouble using 802.11! As for this “only benefitting companies”, it is worth pointing out that it is companies that provide jobs and pay taxes. You can’t have a functioning economy where everyone works for the government, a UN agency, or is a subsistence farmer (though lots of African countries have tried).
phil – this is an awesome propsoal! I nominate you more qualified than most other folks for this task.
– You are savvy about the technology – whatever you don’t know about fiber optics/satellite/tcp-ip networks you can learn very fast
– you have a successful background in entrepreneurship
– you have a great network of technologists and successful entrepreneurs to advise you
– you are financially independent so you can afford to spend a few quarters on this
– this will give you ample chances to travel around the world observe people and make pithy blog posts and use your photography skills as well
– you are writing this blog ergo you are bored 😉
Imagine being the man who transforms the future of Africa!
If you do not at least succeed in drawing large scale attention to this problem – then I don’t know who will.
What do we (a.k.a your blog readers) need to do to launch you on this adventure?
The British government imposes a tax on recieving television via an antenna, and US laws say that it is illegal to eavesdrop on other people’s cell phone conversations, so if the problem you’re trying to solve is working around government regulations, I’m not sure to what extent this helps. (It may be possible that the African governments would be too disorganized to enforce laws against communicating with satellites.)
When you look at the bandwidth available on fiber vs copper, you probably want to be serving Africa using fiber, at least in the parts that have no wired infrastructure yet. I believe the current technology will do well over 200 wavelengths on a single strand of fiber, and 40 gigabits per second per wavelength. You may need a second strand for the signal going back in the opposite direction, and you might want another pair of strands on a different path for redundancy. But you can get all that bandwidth for hundreds of miles without powered repeaters, and you can easily get dozens of times that bandwidth with the number of stands in a typical bundle.
Admittedly, the hardware to use 200+ wavelengths is insanely expensive, but once you bury that fiber, it will probably have plenty of capacity to last a good long time, probably at least half a century of Internet growth, and for now, you can use something modest such as an optical gigabit ethernet transciever.
Fiber also doesn’t provide an electrical path, and therefore simplifies lightning/surge supression issues; the downside is that you can’t power a telephone from a fiber optic cable the way you can power it from a copper pair.
A satellite might have to be constructed using expensive first world labor. A good chunk of the costs of laying fiber in the US is labor, and it seems that fiber in Africa could be installed by the less expensive labor available in Africa, while also providing jobs for Africans. Much of the cost of Internet connectivity in the US is probably the labor cost of building that last mile.
Arbitrage: Thanks for the vote of confidence, but most of my experience is with services built on top of a functioning Internet, not with the plumbing. In any case, I don’t think this is an engineering challenge right now; it seems to be only a funding/willpower challenge. That makes it one for the folks who dole out the $billions.
Joel: There are government telecom monopolies in many African countries and yet nobody is stopped from using an Iridium phone down there. Possibly a handful of countries in Africa would take the trouble to ban the service and certainly it would be nice to see a lot of Africa wired up with fiber, but neither eventuality will stop villagers in the rest of the continent from enjoying satellite-based service.
philg – I disagree 😉 Most of your experience has been with problem solving. I never said this was an engineering problem. This is just a problem – a more multi-faceted one. You have the means and the expertise to draw attention and money to it. And perhaps a few dozen million to it. What would it take to approach some of the tech/telecom billionaires out there and launch a satellite or two to make this service available. Then franchise the hardware manufactured in east asia to spec to receive the signal. You could even do it so that initially one or two birds provide this in a few places – maybe the major urban areas. Maybe hardware runs a couple hundred dollars. Subscription is a few tens of dollars. As Grameen phone has shown in Bangladesh – http://www.ted.com/index.php/talks/view/id/79 – a mobile has the power to reduce poverty.
I reiterate – you dont need billions for this – with new launch vehicles – this can be done very cheaply. You can get a bird aloft for less than cx10E7 where c is less than 2-3?