Homomorphic encryption

Craig Gentry created an encryption system that can solve the problem that prevents many organizations from using cloud computing to analyze and mine information: granting access to unencrypted information to cloud computing providers like Google or Amazon is too big a risk for many organizations.

Link to image:
http://www.technologyreview.com/files/60900/0511-TR-F_x582.jpg

Translated from the Technology Review
Jason Fontin, the editor of the technology magazine Technology Review from MIT, will arrive in Israel this week and participate in the Back to the Future conference of the MIT Forum on Thursday. In honor of the occasion, we have chosen to present to you the Technology review's choice for the ten world-changing technologies of 2011. One of the technologies chosen is homomorphic encryption.
Craig Gentry created an encryption system that can solve the problem that prevents many organizations from using cloud computing to analyze and mine information: granting access to unencrypted information to cloud computing providers like Google or Amazon is too big a risk for many organizations.
The technical problem is that although the information can be sent and received encrypted by a cloud service provider, the servers underlying the cloud cannot process the information when it is in an encrypted state. Gantry a researcher at IBM showed that it is possible to analyze information without deciphering it. The key is to encrypt the information in such a way that performing a mathematical operation on the encrypted information and then decrypting the information will give the same result as performing a corresponding operation on unencrypted information. The correspondence between the operations performed on the encrypted information and the corresponding operations on the decoded information is known as homomorphism. Basically Gantry says homomorphic mathematics can be used to secure operations on information passing through the Internet.

Gentry started tinkering with homomorphic cryptography in 2008, at first he was only able to perform a few simple operations on encrypted data before his system started spewing nonsense. Unfortunately a task like finding text in an email requires a chain of thousands of basic operations. His solution was to use a second layer of encryption, which was required to protect intermediate results when the system crashed and had to be reset.
"The problem of how to create true homomorphic cryptography has been at the center of controversy for more than thirty years, and Craig was the first to figure out how to make the math work," says Paul Kucher, president of Cryptography Research. Even so, Kucher warns that because Gantry's formula currently requires an enormous amount of calculations, it has a long way to go before it will be useful.
Gantry admits that the way he implemented the homomorphic encryption is a bit "breaking out of the window" and that in his estimation it ran too slowly for practical use, but he claims that he is working on optimizing it for specific applications such as searching in databases. He estimates that such applications will be able to enter the market in five to ten years.
To register for the conference Back to the Future of the MIT Forum where Jason Fontin, the editor of the Technology Review, will present the ten hottest technologies for 2011