| In recent years, Wireless Sensor Networks (WSNs) have been attracting increased attention from the research and engineering communities. A WSN typically consists of a large population of spatially distributed sensor nodes that cooperatively monitor environmental conditions and report their findings to an in-situ base station (BS). The BS not only collects and analyzes the incoming data, but also interfaces the WSN to a higher authority. The unique role of the BS attracts adversary's attention since it can be a single point of failure for the WSN. An adversary that seeks to diminish the network utility can apply traffic analysis techniques to locate the BS, and target it with denial of service attacks.;In this dissertation, we develop three traffic analysis countermeasures that morph the network's traffic pattern in order to divert the adversary's attention away from the BS and make the traffic analysis inconclusive. The first two countermeasures use fake sinks (FSs) and deceptive relay nodes, aiming to mimic the traffic pattern of the BS vicinity in multiple areas across the network. The first countermeasure, named, traffic morphing through routing to fake sinks (MoRF), probabilistically selects some nodes to act as deceptive relays to de-correlate packet flows from the data routes by creating redundant traffic toward the fake sinks. The second countermeasure, which we call multiple sinks illusion (MSI), deterministically determines the amount and distribution of deceptive traffic so that the anonymity goal is achieved while the overhead is reduced and evenly split across many nodes.;We further develop a novel attack model, named Evidence theory analysis with reduced search space (EARS) that increases the adversary's confidence in localizing the BS while significantly reducing the traffic analysis complexity. Additionally, we introduce two novel anonymity metrics. The new attack model and metrics better gauge the effectiveness of a countermeasure in terms of BS protection. Most of the existing countermeasures, including MoRF and MSI, usually expose only some parts of the WSN to the adversary. Such a strategy is deemed a compromise that trades off location privacy and performance. In order to strike a balance in the aforementioned trade-off, we finally present a countermeasure, named, preserve location anonymity through uniform distribution of t raffic volume (PLAUDIT) that strives to equalize the traffic density across the network in order to make the BS undistinguishable. |
