Forbes: IBM’s Blindfolded Calculator
Forbes reports on new encryption technology from IBM, which the company says will improve privacy and security for users. It would allow computers to analyze encrypted data (for example, to filter e-mail spam) without seeing any of the private information.
The premise of what computer scientists call “fully homomorphic encryption,” like many long-unsolved mathematical puzzles, sounds both simple and impossible: Can data be encrypted in a way that allows any calculation to be performed on the scrambled information without unscrambling it?
Imagine online accounting software that takes in encrypted data about your salary and expenditures, crunches the numbers and spits out an encrypted tax filing ready to be decrypted and sent to the IRS. The accounting software would never know the contents of what it had analyzed, and you might be less fretful about sending sensitive financial data to that online tax service. [...]
[T]he trend in computing today is moving data and applications like tax accounting and e-mail filtering off the desktop to independently owned data centers–the so-called “cloud.” That shift makes better use of scarce resources and also means that your files are always at your fingertips, no matter where you are or what device you’re using. For cloud computing to fulfill its promises, users need more confidence that their secrets will stay secret. That’s where [IBM researcher Craig Gentry's] discovery fits in, though it could take another decade to put his theory into practice. [...]
Gentry and others argue that encryption needs to be more adaptable. “If you have to decrypt your data to actually use it, no one’s going to encrypt in the first place,” he says. “Inflexible encryption is basically just an obstacle.”
Possibly related posts: