Rumors have been floating around for the past few weeks about a possible Google plan to launch a very large satellite constellation (1600 satellites) to provide global Internet connectivity. Now it looks like Facebook has similar plans.
According to information released last week, Facebook has been hiring employees from places NASA Ames Research Center, JPL, and the National Optical Astronomy Observatory for its new Facebook Connectivity Lab. As described on the website Internet.org, the Facebook Connectivity Lab seeks to bring the Internet to the major portion of the world (70%) that currently lacks connectivity.
Facebook plans to rely heavily heavily on drones (unmanned air vehicles or UAVs) operating at an altitude around 65,000 feet — high enough to avoid weather and traffic conflicts, but low enough to maximize signal strength. At that altitude, a single drone could cover a medium-sized city. Drones would be powered by solar panels during the day and batteries at night. Facebook is already building the first prototype. The effort will be staffed by key employees from Ascenta, a UK-based drone developer recently acquired by Facebook.
Nevertheless, Facebook believes there will still be places where it is uneconomical or impractical to deploy drones to provide internet connectivity. In these lower-density areas, satellites may provide a cheaper alternative.
Facebook is considering both geosynchronous (GEO) and Low Earth Orbit (LEO) communications satellites.
LEO satellites would have lower power requirements, could be smaller and cheaper to launch, and provide lower signal latency for real-time services like the web or voice calling. A LEO constellation would require more satellites, however, and ground-segment antennas that can track moving satellites.
Instead of ordinary radio links, Facebook plans to used infrared laser-based communications, called FSO (free space optical) technology.
FSO communications enables much faster data speeds. FSO technology is harder to implement for LEO satellites, though, because of satellite movement.
Facebook notes that the cost of space transportation is a problem. “Even if you can build satellites for relatively cheaply, transport to space can cost millions — or in some cases tens or hundreds of millions — of dollars.”
Regulatory issues are a problem also. Facebook notes that the ITU licensing process for the microwave spectrum currently takes 5-7 years. (FSO technology is currently unregulated, however.)