In February 2009 the U.S.'s Iridium 33 satellite collided with the Russian Cosmos 2251, instantly destroying both communications satellites. According to ground-based telescopes tracking Iridium and Cosmos at the time, the two should have missed each other, but onboard instrumentation data from even one of the satellites would have told a different story. Why weren't operators using this positional information?
Orbital data are actually guarded secrets: satellite owners view the locations and trajectories of their on-orbit assets as private. Corporations fear losing competitive advantage—sharing exact positioning could help rivals determine the extent of their capabilities. Meanwhile governments fear that disclosure could weaken national security. But even minor collisions can cause millions of dollars' worth of damage and send debris into the path of other satellites and spacecraft carrying humans, such as the International Space Station, which is why the Iridium-Cosmos crash prompted those in the field to find an immediate fix to the clandestine problem.
In the current working solution, the world's four largest satellite communications providers have teamed up with a trusted third party: Analytical Graphics. The company aggregates their orbital data and alerts participants when satellites are at risk. This arrangement, however, requires that all participants maintain mutual trust of the third party, a situation often difficult or impossible to arrange as more players enter the field and launch more satellites into orbit.
Now experts are thinking cryptography, which can eliminate the need for mutual trust, may be a better option. In the 1980s specialists developed algorithms that allowed many people to jointly compute a function on private data without revealing any number of secrets. In 2010 DARPA tasked teams of cryptographers to apply this technology to develop so-called secure multiparty computation (MPC) protocols for satellite data sharing. In this method, each participant loads proprietary data into its own software, which then sends messages back and forth according to a publicly specified MPC protocol. The design of the protocol guarantees that participants can compute a desired output (for example, the probability of collision) but nothing else. And because the protocol design is public, anyone involved can write their own software client—there would be no need for all parties to trust one another.
One of the current drawbacks of cryptography for orbital data is speed. Calculating the probability of collision between two satellites requires intense calculations: insecure computations take milliseconds, whereas these protocols take 90 seconds when performed on commodity hardware. As computing power improves, however, the MPC protocols will become more practical to use. Now DARPA's efforts are wrapping up, and a proof-of-concept algorithm is ready. At present, no one is using the protocols in practice, but cryptographers are looking for adopters of the technology.