Trust no one: a decentralized matching service for privacy in location based services

We propose a new approach to ensure privacy in location based services, without requiring any support from a "trusted" entity. We observe that users of location based services are sensitive about their i) location coordinates and ii) their interests and social relationships, as captured in their queries. We also observe there are entities that naturally have access to at least one of these pieces of information. The user and/or their mobile operator has access to their current location, and the LBS provider needs to know of the interests (in businesses, services and acquaintances) of a user. In this paper we consider whether it is possible for these entities to exchange information such that a user's queries to the LBS can be answered without i) any one entity coming to know of all sensitive information ii) a loss in the quality of service of the query, or an inordinate load on the user. Specifically, we outline the design of a decentralized matching service that takes encoded information from both the participating entities, and creates triggers when a user, and their objects of interest are in the vicinity of each other. Given that each component of the matching service has access to only a limited amount of encoded information, we argue that it will be impossible to recreate any sensitive user-specific information.