Secure Spatial Query with Differentially Private Access Pattern

Abstract

The spatial transformation mechanism (e.g., space-filling curves) enables the spatial query in the encypted spatial data stored on the third party service provider. Although this mechanism will not reveal the original spatial data nor the query content to the service provider, the service provider can learn the relationship between the query tokens and the accessed files, which is called the access pattern. Several researches have demonstrated that the access pattern can be used to recover the content of a query. In this paper, we propose a secure spatial query scheme with differential privacy access pattern. The proposed scheme employs the d-privacy to measure the utilities of the candidate spatial indexes, and applies subsampled exponential mechanism to select the obfuscated spatial index. In this differentially private way, the outsource spatial data is able to resist the query recovery attacks. The experimental results demonstrate that our scheme could generate high quality candidate spatial indexes for the subsampled exponential mechanism, and achieve good precision in the spatial range query.