Source location privacy in wireless sensor networks: What is the right choice of privacy metric?

Today, communication between objects, machines, objects to machines and to humans is possible due to the Internet of Things (IoT). However, their applicability is restricted mostly to areas that are inhabited by humans. Monitoring and tracking in wilderness areas is still a challenging task to date, if not impossible. To bridge this gap, IoT networks are instrumented with Wireless Sensor Networks that are capable of providing remote services through multi-hop communication paradigm. Since these networks are deployed in deserted places, it becomes very crucial to protect the privacy of the location information of critical events or sources that these networks are monitoring. To this end, we propose a new random-walk based routing protocol namely BLS (Backward walk, L-walk, Shortest path walk) to protect the location of critical sources/events. The aim is to break the correlations between the network traffic and render the traffic-analysis efforts of the attacker, in locating the source of information, useless. In addition, we also evaluate the performance of the proposed technique by comparing it with the existing techniques using different privacy metrics such as safety period, entropy and capture ratio. Through this research work, we observed that the performance of source location privacy (SLP) preservation techniques is giving differing results for different privacy metrics. Although the proposed solution outperforms in terms of entropy metric by 104.59-folds improvements compared to Forward Random Walk technique, its performance in terms of safety period and capture ratio metrics are very poor with an improvement of just 0.65-folds and 0.1-fold respectively. Therefore, there is a dire need to come up with a right choice of metric for SLP preservation techniques.