Authenticating moving spatial queries

Abstract

A spatial query is a database query that searches for information from a geometric dataset. k nearest neighbor (kNN) queries and range queries are two types of spatial queries. A k nearest neighbor query searches for the k nearest objects of a query location; a range query searches for objects that are within a specified distance from a query location. A moving spatial query is a continuous submission of a spatial query by a moving client who wants to continuously get the most updated spatial query results. For example, a moving 5NN query can be used to continuously monitor the 5 nearest gas stations of a moving car. As another example, a moving range query can be used to continuously monitor all restaurants that are within 1 kilometer from a walking tourist. A safe region is a technique to optimize mobile query processing. In addition to the query result, a service provider that evaluates moving spatial queries often returns the mobile client a safe region that guarantees that, as long as the mobile client is inside the safe region, the query result is still valid. Hence, the safe region can reduce the communication frequency between the service provider and the mobile clients. However, when a service provider is not trustworthy, it may send inaccurate query results (e.g., sponsored facilities are ranked higher by the service provider) or incorrect safe regions to mobile clients. (e.g., the service provider may return a very large safe region so as to save computational resources). Therefore, it is necessary to devise techniques to authenticate the query results and safe regions in mobile query processing. In this thesis, we present an architecture for carrying out moving spatial query authentication. We develop algorithms to authenticate the results and the safe regions of moving spatial queries. We theoretically prove that our methods can carry out moving query authentication using a minimal amount of communication overhead. Extensive experiments are carried out using both real and synthetic datasets. Experiment results show that our methods can carry out moving spatial query authentication with small communication costs and overheads.