HybridHash: An Efficient Hash Index for Encrypted Databases

This paper presents a new hash index structure called HybridHash for encrypted key-value databases. HybridHash employs two rounds of hashing on the search keys to partition the ciphertext data effectively without leaking plaintext information, thus enhancing the retrieval efficiency. In addition, instead of directly storing encrypted key-value pairs, HybridHash stores the addresses of ciphertext data in the data files within buckets. By employing searchable encryption for intra-bucket queries, HybridHash significantly improves query performance within buckets while reducing the maintenance cost of the index. We conduct experiments on various workloads to compare HybridHash with two baselines, including EEH and SWP, and the results suggest the efficiency of our proposal.