A novel traffic-analysis back tracing attack for locating source nodes in wireless sensor networks

In habitat monitoring applications, when a sensor node detects an endangered animal, e.g., a panda, it reports the animal´s presence and activities to the sink. However, the adversaries can eavesdrop on the network transmissions and make use of the traffic information to locate pandas to hunt them. In this paper, we first define hotspot phenomenon that causes an obvious inconsistency in the network traffic pattern due to the large volume of packets originated from a small spot. Second, we develop a realistic adversary model assuming that the adversary can monitor the network traffic in multiple areas rather than the entire network or only one area. We then introduce a novel attack called Hotspot-Locating where the adversary uses traffic analysis techniques to locate hotspots. Simulation and analytical results demonstrate that Hotspot-Locating attack is a severe threat to the source nodes´ location privacy and the existing routing-based privacy preserving schemes are vulnerable to this attack because they leak traffic analysis information that can be used to locate the source nodes. For stronger privacy preservation, the traffic analysis information such as packet correlation and the nodes´ packet sending rates should be concealed.