According to The Wall Street Journal, a company called O3b Networks LTD Traditional is planning to launch up to 16 satellites by the end of 2010 to provide Internet access in Africa, the Middle East, and parts of Latin America. This satellite plan, unlike many others, could be good. These are low earth orbit satellites or LEOs so they will be able to avoid the latency problems which are unavoidable with the geostationary satellites used by companies like WildBlue and Hughes to provide “last resort” Internet access in the US.

O3b stands for “other 3 billion”—the people in the developing world. No one should ever underestimate a developing market. O3b’s backers for $60 million of the estimated $650 million they’ll need include Google, HSBC Holdings PLC, Allen & Company, and Liberty Global Inc.

The trouble with geostationary satellites is that they have to be 22,000 miles above the equator in order to appear stationary from earth (law of physics). Even at 186,000 miles/sec (another law of physics), radio signals take so long to get up and down that noticeable latency is introduced. Typically these services are useless for gaming, near useless for VoIP, and increasingly useless for web surfing as web sites get more and more interactive. They can be OK for file downloading, email, and relatively small uploads.

The advantages of geostationary satellites are that very few of them cover lots of the earth, they are high so they stay up and useful for a long time, and, because they appear stationary from earth, an antenna can be pointed precisely at them to save on transmit power.

LEOs are used by services like Iridium and GlobalStar primarily for voice use in remote locations or at sea where cellular networks don’t reach. The services are expensive—partly because of how many LEOs are required to cover the earth and partly because friction with the atmosphere makes low satellites fall relatively soon after they’re launched. So it’s economics and not physics that’ll be a problem for O3b. To succeed, they’ll have to get good penetration (but they do have greenfield markets) and offer high data rates at prices much lower than the primitive data offered by Iridium and GlobalStar. Unlike these companies, which offer service directly, O3b plans to be a wholesaler and leave distribution of Internet access to local providers. This strategy means that they can rely on high bandwidth smart head end antennas which can track satellites as the move across the sky and that the satellite radios don’t have to worry about numerous relatively low-speed connections.

Sounds like it might work.