Using data mules to preserve source location privacy in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) have many promising applications for monitoring critical regions, like military surveillance and wildlife monitoring. In such applications, it is critical to protect the location of the source sensor that generates the data, as exposure of this information usually reveals the location of the object being monitored. Traditional security mechanisms, like encryption, have been proven to be ineffective as the location of the source can also be revealed by analyzing the traffic flow in the network. In this paper, we investigate the source-location privacy issue. We first propose a realistic semi-global eavesdropping attack model and show its effectiveness in compromising an existing source-location preserving technique. Furthermore, to measure source location privacy against the semi-global eavesdropper, we define a model for α-angle anonymity. Additionally, we design a new protocol called Mule-Saving-Source (MSS) that preserves α-angle anonymity by adapting the conventional function of data mules. We theoretically analyze the delay incurred by using data mules in MSS, and we examine via extensive simulations the trade-off between the delay and privacy preservation under different data mule mobility patterns. We categorize the delay in MSS as being caused primarily due to the buffering time at the source sensor and the data mules. Motivated by this observation, we propose two modifications to MSS, Mule-Saving-Source-Shortest Path (MSS-SP) and Mule-Saving-Source-Two Level (MSS-TL), both aimed at reducing the total delay by reducing the buffering time at the data mule and source respectively. Through theoretical analysis, we examine the delay in the proposed modifications and evaluate their performance with the MSS protocol using a comprehensive set of simulations. Furthermore, to study the impact of the mobility model of the data mules on the MSS protocol, we compare the performance of the MSS protocol by changing the mobility model of data mules to a Random Waypoint based model.