Privacy-Preserving Polynomial Evaluation over Spatio-Temporal Data on an Untrusted Cloud Server

Wei Song, Mengfei Tang, Qiben Yan, Yuan Shen, Yang Cao, Qian Wang, Zhiyong Peng

Abstract: Nowadays, with the popularity of location-aware devices, multifarious applications based on the spatio-temporal data come forth in our lives. In these applications, a platform (enterprise) collects the users’ spatio-temporal data based on which it recommends the top-k users (passengers) to the registered service providers (drivers). Outsourcing the tremendous scale of spatio-temporal data to cloud provides an economical way for the enterprises to implement their applications. In this paradigm, the third-party cloud server is not completely trustworthy. The collected spatio-temporal data can hold users’ privacy, so it’s a critical challenge to design a secure and efficient query mechanism for this scenario, such as the ride-hailing or the ride-sharing services. However, the existing solutions for the privacy-preserving kNN queries mainly focus on data privacy protection or computation complexity. There still lacks a practical privacy-preserving polynomial evaluation solution over the spatio-temporal data. In this paper, we propose a virtual road network structure to storage and index the spatio-temporal data in the road network and design a novel homomorphic encryption scheme based on Order-Revealing Encryption to enable an untrusted cloud server to execute the polynomial evaluation over the encrypted spatio-temporal data in the road network. We formally prove the security of the proposed scheme under the random oracle model. Extensive experiments on real world data demonstrate the effectiveness and efficiency of the proposed scheme over alternatives.