Efficient Integrity Verification of Secure Outsourced kNN Computation in Cloud Environments
Hong Rong, Huimei Wang, Jian Liu, Wei Wu, Ming Xian · 2016
With the ever increasing demand on computational and storage resources to deal with tremendous growth of big data, clients tend to outsource their data mining tasks to cloud service providers. Nevertheless, concerns of data integrity, security and privacy are also on the rise: how can the clients with weak computational power verify the integrity of mining results returned by the server while preserving their privacy. In this paper, we focus on the specific task of outsourced k-Nearest Neighbor (kNN) computation. The cloud server is considered to be potentially semi-honest and unscrupulous by offering incorrect answers due to economic incentive or execution failure. We propose an efficient probabilistic verification method called Verifiable Secure kNN (VSkNN), an integrity verification delegate framework which utilizes the algebraic properties of scalar products in encryption schemes and a small quantity of artificial tuples for correctness checking. Both theoretical analysis and experimental results demonstrate that our approach can provide high probabilistic guarantees on the accuracy of kNN query results efficiently in a privacy-preserving manner.