Many challenging real world problems, e.g., voting and blind auction, require computation over sensitive data supplied by multiple mutually-distrustful entities. Elegant cryptographic theories have been developed to solve these problems without relying on a mutually-trusted third party. Practitioners also built prototypes capable of securely computing set intersection, AES encryption, Hamming distance, etc. However, many other applications, such as data mining and running universal machines, are far more complex than what can be supported by the state-of-the-art techniques. This work studies new methods and platforms to enable large-scale, complex cryptographic protocols in more efficient ways. Software produced will be released as open-source projects to benefit future research in related areas. One part of this project explores new extensible approaches to accommodate multiple Secure Multiparty Computation (SMC) protocols and enable studying interesting combinations of them, including a new compiler to provide a higher-level abstraction for generating SMC protocols. The other part involves research into several new optimizations to improve the performance and scalability of SMC, through reuse of validated components to amortize validation costs. The educational activities of this project include mentoring both undergraduate and graduate students in applied cryptography research and developing a graduate course in applied cryptography.
CRII: SaTC: Efficient Secure Multiparty Computation of Large-Scale, Complex Protocols