Formal Methods for Wireless Systems

Wireless systems consist of wireless devices which communicate with each other by means of a radio frequency channel. This networking paradigm offers much convenience, but because of the use of the wireless medium it is inherently vulnerable to many threats. As a consequence, security represents an important issue. Security mechanisms developed for wired systems present many limitations when used in a wireless context. The main problems stem from the fact that they operate in a centralised manner and under the assumption of a \closed world". Formal techniques are therefore needed to establish a mathematically rigorous connection between modelling and security goals. In the present dissertation we apply the well-known formalism of process calculus to model the features of wireless communication. The scientic contributions are primarily theoretical.We propose a timed process calculus modelling the communication features of wireless systems and enjoying some desirable time properties. The presence of time allows us to reason about communication collisions. We also provide behavioural equivalences and we prove a number of algebraic laws. We illustrate the usability of the calculus to model the Carrier Sense Multiple Access scheme, a widely used MAC level protocol in which a device senses the channel before transmitting. We then focus on security aspects, in particular we propose a trust model for mobile ad hoc networks, composed only of mobile nodes that communicate each other without relying on any base station. We model our networks as multilevel systems because trust relations associate security levels to nodes depending on their behaviour. Then we embody this trust model in a process calculus modelling the features of ad hoc networks. Our calculus is equipped with behavioural equivalences allowing us to develop an observational theory. We ensure safety despite compromised nodes and non interference results. We then use this calculus to analyse a secure version of a leader election algorithm for ad hoc networks. We also provide an encoding of the endairA routing protocol for ad hoc networks. Finally, we extend the trust-based calculus with timing aspects to reason about the relationship between trust and time. We then apply our calculus to formalise the routing protocol ARAN for ad hoc networks.