Forward and Backward Private Searchable Symmetric Encryption with Efficient Updates.

Zhiming Yin,Xiong Li,Kim-Kwang Raymond Choo,Yuzhen Liu,Wei Liang

DOI: https://doi.org/10.1109/hpcc-dss-smartcity-dependsys57074.2022.00107

2022-01-01

Abstract:In dynamic searchable symmetric encryption, clients can search and update encrypted data stored on the server (assumed to be semi-honest). In such schemes, forward privacy guarantees that prior queries are not related to subsequent updated entries, while backward privacy ensures that the previously deleted files are not retrievable from the subsequent queries. However, most of the existing schemes have poor update efficiency, for example incurring significant computational and communication overheads since the use of the complex structures and cryptographic primitives used. Hence, we propose a Forward and Backward private searchable symmetric encryption scheme with Efficient uPdates, called FBEP. Specifically, FBEP achieves efficient updates based on Bloom filters and lightweight symmetric cryptographic primitives. It enables the client to remove file identifiers without interacting with the server and insert them efficiently, while maintaining high search efficiency. We precisely define the leakage of FBEP, and demonstrate that FBEP supports forward and Type-II backward privacy with formal security analysis. Besides, the experiments show that FBEP is much better than other related schemes in terms of comprehensive performance. Specially, the update efficiency of FBEP outperforms at least $1.6\times, 25\times, 525\times$ than Mitra, $\text{SD}_{d}$ and $\mathbf{Janus} ++$ , respectively, while the search time is at least $2\times, 702\times$ faster than Mitra and $\mathbf{Janus} ++$ , respectively.