An Encryption Algorithm based on Matrix Supporting Fuzzy Retrieval in Cloud Computing

摘要

With the extensive application of cloud computing, privacy has become the key problem. The traditional encryption technology can effectively guarantee the security of sensitive data, but it does not support the operations on ciphertext data directly, so that the security and computability of outsourcing data can't be taken into account together. Aiming at this problem, the paper proposes a retrievable encryption algorithm RESVMC (Retrievable Encryption Scheme based on Vector and Matrix Calculations), which realizes the prefix matching by the scalar product of vectors, and then realizes the fuzzy retrieval based on relevance ranking. The security analysis shows that RESVMC is IND-CCA (Indistinguishability under Chosen Ciphertext Attack) when the attacker only can visit the encryption Oracle and decryption Oracle of outsourced data. Compared with the existing scheme, RESVMC has smaller encryption and decryption computational loads, but the retrieval load is larger, and the storage/communication loads are slightly larger; the values of the performance indexes increase with the increase of vector dimension.