A formal methodology to design and deploy dependable wireless sensor networks

Wireless Sensor Networks (WSNs) are being increasingly adopted in critical applications, where verifying the correct operation of sensor nodes is a major concern. Undesired events may undermine the mission of the WSNs. Hence their effects need to be properly assessed before deployment to obtain a good level of expected performance and during the operation in order to avoid dangerous unexpected results. In this paper we propose amethodology that aims at assessing and improving the dependability level of WSNs by means of an event-based formal verification technique. The methodology includes a process to guide designers towards the realization of dependable WSN and a tool ("ADVISES") to simplify its adoption. The tool is applicable to homogeneous WSNs with static routing topologies. It allows to generate automatically formal specifications used to check correctness properties and evaluate dependability metrics at design time and at runtime for WSNs where an acceptable percentage of faults can be defined. During the runtime we can check the behavior of the WSN accordingly to the results obtained at design time and we can detect sudden and unexpected failures, in order to trigger recovery procedures. The effectiveness of the methodology is shown in the context of two case studies, as proof-of-concept, aiming to illustrate how the tool is helpful to drive design choices and to check the correctness properties of the WSN at runtime. Although the method scales up to very large WSNs, the applicability of the methodology maybe compromised by the state space explosion of the reasoning model, which must be faced partitioning large topologies into sub-topologies