You are here : Home > Remote computing just got more secure

News

IoT for healthcare

Remote computing just got more secure


​List, a CEA Tech Institute, implemented homomorphic encryption technologies that enable "blind"—in other words, totally confidential—computation on data stored on remote servers. Several industrial applications have been developed using the technique.

Published on 29 March 2018

Homomorphic encryption allows computation on data without knowing the values of that data. Here's how it works: The data is computed exclusively in encrypted form; the result is also encrypted and requires an encryption key to decipher. In research conducted under the EU HC@Works project, List applied its cryptocomputing technology (or compiler) to actual use cases provided by industrial partners, including Atos, which hopes to use the technology in its future e-healthcare platform, which will offer day-to-day in-home patient monitoring services.

One of the major hurdles to the widespread adoption of the technology is that encrypted data requires much more space and much more time to compute—1,000 times more—than unencrypted data. List researchers overcame this challenge by interfacing two different encryption systems. In layman's terms, a traditional encryption system ensures simplified encryption of the data. The encrypted data does not take up any more space than unencrypted data and can thus be sent to Atos' cloud infrastructure. The data is then "trans-encrypted" by a homomorphic encryption system before being processed. Test results are sent from the cloud server to the patient's doctor, who is the only person to possess the encryption key to unscramble the data. The source data can only be read by its owner.

The project has produced some encouraging results, with latency of less than two minutes. The platform is currently being tested in real-world conditions. An open-source version* of the compiler was made available in early 2018 to support the creation of a European community uniting professionals from industry and academia.

*https://github.com/CEA-LIST/Cingulata.

Top page

Top page