IBM scientists have patented data encryption technology that enables processing of encrypted data without decrypting it.

The technology solves one of the biggest problems with encryption. While a service provider can store a user's data in encrypted form, for them to perform operations on the data, it has to be decrypted, exposing potentially sensitive information to an outside company.

IBM now holds a patent for an implementation technology for fully homomorphic encryption, which enables mathematical operations on data in encrypted state. The US Patent Office has awarded the patent to IBM cryptography researchers Craig Gentry and Shai Halevi.

Explained simply, homomorphic encryption is when data is encrypted but relationships between its different components are maintained in the ciphertext (data in encrypted state). Those relationships are exploited when analytics is performed on the encrypted data, and the result is returned in encrypted form, which can be decrypted with the user's original key.

In today's world, where data storage and processing is increasingly outsourced, and where there is less and less trust that it is safe from hackers or governments, this technology may become hugely beneficial. Companies will potentially be able to safely store financial data with a cloud provider and use that provider's analytics services when they need to retrieve specific elements of that data without having to download it all first, for example.

“Our patented invention has the potential to pave the way for more secure cloud computing services – without having to decrypt or reveal original data,” Gentry said. “Fully homomorphic encryption will enable companies to confidently share data and more easily and quickly overcome challenges or take advantage of emerging opportunities.”

IBM is not the only company researching homomorphic encryption. Earlier this year, Fujitsu announced a fully homomorphic encryption technology that enables sophisticated calculations to be performed on encrypted data.

Fujitsu's technology also enables these calculations to be performed at high speeds, according to the Japanese company. Instead of encrypting data at the bit level (typical approach to homomorphic encryption), Fujitsu's method orders bits into strings and encrypts the data in batches, substantially speeding up the process.