Simulation-based analysis of threats to location privacy in fog computing

In fog computing, end devices can benefit from low-latency access to computing capacity provided by nearby fog nodes. However, an adversary controlling some fog nodes may infer information about the location of end devices that engage with these fog nodes. The goal of this paper is to analyze the severity of this threat to location privacy. We analyze how precise the information leaked by fog nodes about the location of end devices may be, and how this depends on the ratio of compromised fog nodes and on the adversary's background knowledge. We present the simulator LocPrivFogSim, which extends the existing fog computing simulator MobFogSim with the constructs necessary to model location privacy threats. The findings from preliminary simulations of various attack scenarios show that an attacker controlling even a modest ratio of the fog nodes may be able to infer precise information about the location and trajectory of end devices.