Establishing location-privacy in decentralized long-distance geocast services

The ability to communicate over long distances is of central importance for smart traffic applications like cooperative route planning or the discovery and reservation of charging stations for electric vehicles. Established approaches are based on centralized architectures with singular service providers. This setup leads to strong privacy concerns, as great amounts of sensitive location data need to be stored at a non-local, centralized entity. Decentralized approaches like the overlay-based geocast service OverDrive propose to solve this issue by eliminating the central data sink and sharing location information with a small subset of other participants. In this paper, we propose techniques for further improving the location privacy offered by decentralized long-distance geocast services. Through obfuscation of location data and mechanisms for detecting location spoofing attempts, we can ensure that precise location data is only shared with participants in the physical vicinity. Simulation results show that our extensions render both the large scale surveillance and the targeted tracking of OverDrive users unfeasible even for strong adversaries controlling hundreds of overlay nodes. In addition, we discuss practical considerations when deploying decentralized and privacy-sensitive systems that rely on cellular networks and present results from an empirical evaluation of connectivity properties.