Communication and Client Storage Optimized Dynamic Searchable Symmetric Encryption

Most dynamic searchable symmetric encryption (DSSE) schemes that satisfy forward and backward privacy require maintaining a state table for each keyword. Storing these state tables on the client side incurs additional storage overhead. Furthermore, while some constant client-side storage schemes reduce client-side storage, they increase the computational and communication overhead during updates and searches. Forward and backward privacy security are ensured and communication is optimized in a dynamically designed searchable symmetric encryption scheme with constant client-side storage, which is developed based on a hierarchical index structure and a hash key chain and designated as Communication and Client Storage Optimized DSSE (CCSO). The scheme utilizes a hash key chain and a global counter for key distribution: the one-way property of the hash key chain guarantees forward privacy, while backward security is achieved via lazy deletion. The state information of all keywords is bound to the global counter to reduce client-side storage overhead. During search execution, only the keyword search token under the current state needs to be uploaded to complete the search operation on the entire encrypted database. Finally, a security analysis of the CCSO scheme is conducted, and comparative experiments on update efficiency, search efficiency, and communication performance are performed between CCSO and existing schemes (SDa, CLOSE-FB, and DISCO-h). The results show that: (1) the CCSO scheme achieves forward privacy; (2) CCSO only requires 100 bits of uplink communication for searches, which is much lower than the sublinear communication overhead of 1 200 bits in SDa, improving practicality; (3) The single-search computational latency of the CCSO scheme is only on the order of 10^-4, which is lower than the 10^-2 order computational latency of CLOSE-FB and DISCO-h, achieving sublinear search computational overhead. In summary, CCSO features O(1) client-side storage overhead while reducing the uplink communication complexity during search to a constant level and the computational complexity to a sublinear level, effectively addressing the imbalance between client-side storage overhead, communication overhead, and search computational overhead.