Abstract
The increasing concern over privacy of consumer, enterprise and government data stored in cloud has led to renewed interest in the 15-year-old topic of searchable encryption. Various cloud service, storage system and database designers are looking for solutions to provide end-to-end encryption on user data, yet not to lose on the value that they offer on stored data. However, the space of searchable encryption is cluttered with an array of options that are easy enough to confuse system designers.
This talk is aimed at demystifying the following cryptographic primitives that can be used to search on encrypted data: multi-party computation, oblivious RAM, searchable symmetric encryption, functional encryption, property-preserving encryption, fully-homomorphic encryption. Technical analyses that compare performance-security tradeoffs, threat/leakage modeling, etc., will also be presented.