Whether it's a Microsoft commercial telling you to "take it to the cloud" or you happen to work on some of the most cutting-edge computing issues in modern times, it's hard to escape cloud computing.
The uses for cloud computing, in which companies or individuals can run applications on shared data centers, seems limitless. But they're not. One of the constraints of cloud computing is that encrypted information can’t be worked on within the cloud because there’s no way to keep that data safe once it's been opened up to do computations. Enter homomorphic encryption.
Homomorphic encryption allows multiple users to work on encrypted data in an encrypted form, but yields the same results as if the data had been unlocked. This idea had been theoretical in the computing world until Craig Gentry, a researcher with IBM, found an algorithm that makes it possible. The breakthrough was named a top 10 emerging technology of 2011 by MIT's Technology Review, but it's still far from market-ready.
"With fully homomorphic encryption, I can take lots of problems and do lots of calculations and all the owner can see is a few encrypted numbers," said Steven Bellovin, a computer science professor at Columbia University. "The problem is that the performance is far, far, far too slow."
Despite being too slow, many people doubted whether, in theory, it could even be done, making Gentry's work a big step.
If it's possible to make enough advances to bring the technology to market, it will have applications far beyond the cloud. Homomorphic encryption could be used any time you’re doing a calculation on something you don’t own, said Bellovin.
He gave the example of an electronic wallet. If you could use your smartphone like a wallet, depositing checks and taking out money for purchases, the bank would not want you to hack into the system to increase your checking balance. Instead, the bank would want a system where it can make the updates to your checkbook in terms of how much has been deposited or withdrawn, but you can't see or interfere with those calculations from your smartphone.
Bellovin said the operations are likely to get quicker, but he couldn't say by how much.
"Computers have gotten so much faster," Bellovin told InnovationNewsDaily. "But this is the sort of thing that if we don't get better algorithms, Moore's Law (a theory of exponentially increasing computing power) will never save us."
This story is part of a series covering MIT Technology Review's Top 10 Emerging Technologies of 2011 list. You can read the previous parts of the series here.