Computer Science Colloquia
Monday, January 9, 2012
Advisor: abhi shelat
Attending Faculty: Jack Davidson, Chair; Dave Evans, and Gabriel Robins.
Rice Hall, Room 230, 10:00 am
Ph.D. Qualifying Exam Presentation
Efficient use of Hybrid Symmetric/Fully Homomorphic Encryption
Fully homomorphic encryption (FHE) systems represent a breakthrough in secure two-party computation, but the high communication and computation costs make these systems impractical. The communication cost can be mitigated by using hybrid symmetric/fully homomorphic encryption, which is similar to the approach of hybrid symmetric/public key encryption. However, the computational cost of this approach may be very high. In this project, computationally efficient homomorphic evaluation of symmetric ciphers is investigated. The block cipher Serpent is compared to previous work on AES, and a provably secure pseuorandom function is also considered. It is shown that although Serpent is more computationally intensive than AES, it can be evaluated more efficiently using state-of-the-art FHE systems.