Efficient Privacy-Preserving Query Processing on Outsourced Geographic Databases

Location-based services (LBS) enables a user to query on spatial-temporal data with certain query criteria. With the popularity of mobile devices such as mobile phones and navigators, LBS has shown increasingly importance in various real world applications. However, in a traditional LBS system, the service provider maintains a geographic database (Geo-DB) and acquires the users' real-time locations. This raises the concerns of users privacy. This paper presents a privacy-preserving LBS scheme. In our scheme, the data owner outsources Geo-DB to the server in encrypted form, and the user enjoys the LBS without disclosing its location to the server. We construct a secure index to data objects in the Geo-DB by using the Encrypted Coded Quad-tree (ECQtree), such that user's queries can be efficiently located over the encrypted data. We also propose a Bloom filter based method for users to generate query trapdoors by hashing user's query regions into a Bloom filter vector, which is able to hide users' locations and query regions from the server. We also give security proof and analysis. Extensive simulations have been conducted and the results have shown the superior performance of our proposed scheme.