7. Conclusions
Location and identity privacy in Location-Based Services are major concerns for users who want to protect their privacy from a malicious LBSP, as well as from an eavesdropper. While several so- lutions for guaranteeing privacy in LBSs have been proposed in the literature, they are often centralized, or do not take into account the prior knowledge of the attacker about user profiles. In this pa- per we present ABAKA, our collaborative solution that guarantees k-anonymity, as well as p-sensitivity in LBSs, taking into account the issued query semantics. In our approach, users have a set of attributes associated to their profile. Their attributes are bound to a CP-ABE private key. An LBS message is first processed by the is- suer, and then forwarded through a multi-hop route to the LBSP. ABAKA enables each issuer to delimit a cloaked area within which she wants to be anonymous, and to specify a list of k − 1 poli- cies, i.e., attribute combinations, that users in the multi-hop path must satisfy in order to forward the query message to the LBSP. ABAKA provides the possibility of performing a trade-off between the stringency of privacy protection and quality of service for the issuer in her current location, based on the query semantics. We addressed the threat of active and passive adversaries by means of CP-ABE and multi-hop routing approaches. We simulated our pro- tocol on synthetic datasets derived from real population statistics (considering two cities: New York (USA), and Milan (Italy)), and demonstrated that our approach is feasible and efficient.
