Privacy enhancement in mobile data broadcasting environments

Abstract

In mobile and wireless environments, clients will request for information by submitting queries to the server, which delivers the required data via point-to-point connection or broadcast channels. The broadcast paradigm possesses the advantage of scalability. However, the broadcast data is vulnerable to eavesdropping and the adversary may be able to associate the requested data items and hence the original query with a specific client, a form of privacy threat that we would like to protect against. In this paper, we propose to protect the privacy of queries generated by individual clients by allowing them to hide themselves behind other clients. We assume a possibly untrustworthy server and make use of a trustworthy anonymizer for the queries. At the query anonymizer, user queries are clustered and consolidated into subsuming queries, to obfuscate the queries from being recovered. The consolidated queries are then expanded to improve the obfuscation effect, making it harder for the adversary to deduce the original queries. We define various performance metrics on the query privacy, by studying the ability that the adversary could crack the client queries, and propose algorithms to obfuscate a collection of user queries.