Query Integrity Assurance of Location-Based Services Accessing Outsourced Spatial Databases

  • Authors:

  • Wei-Shinn Ku;Ling Hu;Cyrus Shahabi;Haixun Wang

  • Affiliations:

  • Dept. of Computer Science and Software Engineering, Auburn University, USA;Computer Science Department, University of Southern California, USA;Computer Science Department, University of Southern California, USA;IBM Thomas J. Watson Research Center, USA

  • Venue:
  • SSTD '09 Proceedings of the 11th International Symposium on Advances in Spatial and Temporal Databases
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Outsourcing data to third party data providers is becoming a common practice for data owners to avoid the cost of managing and maintaining databases. Meanwhile, due to the popularity of location-based-services (LBS), the need for spatial data (e.g., gazetteers, vector data) is increasing exponentially. Consequently, we are witnessing a new trend of outsourcing spatial datasets by data collectors. Two main challenges with outsourcing datasets is to keep the data private (from the data provider) and ensure the integrity of the query result (for the clients). Unfortunately, most of the techniques proposed for privacy and integrity do not extend to spatial data in a straightforward manner. Hence, recent studies proposed various techniques to support either privacy or integrity (but not both) on spatial datasets. In this paper, for the first time, we propose a technique that can ensure both privacy and integrity for outsourced spatial data. In particular, we first use a one-way spatial transformation method based on Hilbert curves, which encrypts the spatial data before outsourcing and hence ensures its privacy. Next, by probabilistically replicating a portion of the data and encrypting it with a different encryption key, we devise a technique for the client to audit the trustworthiness of the query results. We show the applicability of our approach for both k -nearest-neighbor and spatial range queries, the building blocks of any LBS application. Finally, we evaluate the validity and performance of our algorithms with real-world datasets.