TC: Small: Homomorphic Encryption for Cloud Privacy

Worcester Polytechnic Institute, Worcester MA

Investigators

Abstract

This project applies tools from homomorphic encryption to improve security in cloud services. Cloud computing saves money for companies by allowing them to pay for services only if they need them and just at the time that such services are needed, rather than purchasing and maintaining hardware/software that may see little use. The drive to gain these savings may transform both business and personal computing. This new level of resource sharing and complexity naturally generates new problems related to security and privacy, problems that are both novel and challenging. In the broadest sense, homomorphic encryption allows one to borrow computational power from an untrusted source. If made practical, this has especial implications for cloud computing. The software owner may employ homomorphic encryption in this setting to reduce the threat of reverse engineering and to administer software licenses. When made available over a untrusted cloud, the software is run homomorphically, data is submitted and returned in encrypted form, and the execution of instances is highly robust against malicious users. This project removes the uncertainties in the implementation efficiency of homomorphic encryption schemes and improves it to the point where it becomes practical for use in cloud computing. Specifically, the project centers on three modules: instruction set development for homomorphic computing, processor-specific optimizations for homomorphic schemes, and the investigation of new homomorphic schemes.