Anti-Tamper Databases: Processing Aggregate Queries over Encrypted Databases

Abstract

This paper is about anti-tamper databases, where the database contents are encrypted for security in a way to allow efficient query processing directly on the encrypted database without decrypting it. We use a family of openform and closed-form homomorphism encryption/decryption functions and a computing architecture where (a) the database is encrypted a priori, and (b) for SQL queries expressible in relational algebra, there is no extra query processing cost except for the decryption of the final query output. For complex aggregate queries and nested queries that require extra query processing over the encrypted database, we present query execution strategies. We quantify the additional costs incurred when executing aggregate nested SQL queries over encrypted relational databases, and present detailed experimental results. We observe the crossover points as to when processing a query over an encrypted database is still more advantageous than shipping it over the internet to a secure server housing the original, non-encrypted database, evaluating it and returning the query output to the user.