The National Science Foundation will support Papamanthou’s research at Yale with a new grant on attacking and defending encrypted databases. The collaborating institutions are Brown University and George Mason University.
Encrypted databases have been proposed as a practical approach to searching encrypted data stored at a server. In an encrypted database, the client encrypts the data before uploading it to the server. The server is not given the decryption key, yet the server can execute searches on the encrypted data requested by the client and return the corresponding encrypted answers. An encrypted database efficiently achieves the above functionality by allowing some leakage about the original (plaintext) data that appears harmless but could be exploited by an attacker who attempts to reconstruct the original data by synthesizing leakage from the encrypted answers and encrypted queries over time. This project aims to develop the next generation of methods for building efficient encrypted databases and for analyzing their resilience to reconstruction attacks. To achieve this goal, the team will use methods from algorithms, statistics, geometry, databases, and computer systems. Specific research activities include cryptanalysis of the leakage from encrypted databases with suppressed leakage as well as cryptanalysis and schemes for high-dimensional encrypted queries.