Secure Why-Not Spatial Keyword Top-k Queries in Cloud Environments

Yiping Teng, Miao Li, Shiqing Wang, Huan Wang, Chuanyu Zong, Chunlong Fan

Abstract: With the rapid growing volume of spatial data, answering why-not questions on spatial keyword top-k queries, aiming at refining spatial keyword queries to include the missing objects in query results with minimal costs, has attracted much attention in the field of spatial databases. To alleviate the burden of local storage and computation, when service providers outsource the why-not query services to public cloud, it may raise privacy concerns. To address these issues, in this paper we first present a basic secure why-not spatial keyword top-k query (BSWoSKQ) scheme, featuring a secure weight vector generation method to obtain the best approximate refined query with minimal costs. Furthermore, to improve the query efficiency, we propose an optimized scheme named SWoSKQ that employs a new secure index structure, i.e., SSR-tree, and efficient pruning methods based on such secure index. Comprehensive analysis demonstrates the security and computational complexity of our approach, and extensive experiments on real and synthetic datasets validate the query performance of the proposed methods.