Efficient Keyword Query in Encrypted Cloud Databases

Abstract

Encryption is an efficient way of protecting database security. Traditional keyword based queries cannot be executed directly over encrypted databases. An encrypted keyword search method is presented based on the bloom filter mechanism. Keywords are mapped into a bloom filter array, which are attached to database tables and stored as string fields in the tables. When processing queries, the bit-array and the encrypted index are compared, then the query can be carried out against encrypted fields. The principle of bloom filter is introduced together with its error analysis. We also implement multiple hash functions as the building block of our design. Finally, we make an extension of the current SQL in our system by creating user-defined functions and CRL assemblies. The prototype of our design is implemented on a SQL Server database. Experiment results show that our method is efficient and highly applicable.