PriWildMatch: Privacy-Preserving Flexible Wildcard Pattern Matching In The Cloud
Yunfeng Zhang, Xiaowen Fan, Haoyan Huang, Qiang Wang · 2025
Privacy-preserving wildcard pattern matching enables data owners to outsource textual data to the cloud while allowing users to perform flexible queries without disclosing sensitive information. Existing solutions typically support either single-character or multi-character wildcards but are predominantly designed for single-server settings. The extension of such functionality to two-server architectures remains underexplored, and naive adaptations of existing schemes often incur excessive communication rounds and considerable computational overhead. In this paper, we propose PriWildMatch, a privacy-preserving and flexible wildcard pattern matching scheme designed for two-server architectures. It supports both single-character and multi-character wildcards and achieves constant-round interaction, allowing clients to perform secure and efficient pattern matching queries over outsourced datasets. We formally define the security model and prove that PriWildMatch achieves privacy guarantees under standard cryptographic assumptions. Extensive experimental results demonstrate that PriWildMatch achieves significantly lower communication rounds and improved computational efficiency while preserving privacy, and validating its practicality in real-world cloud environments.