Current cloud based systems enable distributed access to both information and computational resources. In this setting, it is imperative to have secure communication, and powerful and expensive cryptographic techniques have been proposed to address this issue. A severely limiting factor, however, is that these methods for securely accessing or processing data between participating parties can result in communication overheads when processing large amounts of data. This project focuses on the question of minimizing the communication complexity of cryptography and explores new approaches for tackling them. This work could lead to wider deployment of ideas in cryptography which will result in better security and privacy for users, as well as reduced costs and increasing security for cloud service providers.
The goal of this project is to investigate computationally lightweight solutions whose security can be analyzed without resorting to unproven mathematical conjectures, and diversifying the set of conjectures that suffice for provable security to be achieved. The project develops new techniques for low communication interactive proofs and arguments in order to improve their efficiency, diversify the underlying intractability assumptions, and minimize the necessary set-up assumptions. The investigators explore the possibility of reducing the communication complexity of secure computation protocols while avoiding the use of expensive general-purpose tools such as fully homomorphic encryption. The project investigates new connections between communication complexity and cryptography by using recent results and techniques for interactive compression towards minimizing information leakage in unconditionally secure protocols. The results from this project could enable new technologies for secure manipulation of big data.
|