Modeling and counteracting virus diffusion in sensor networks for net-centric surveillance systems

The trend of modern networked systems is toward distributed or decentralized or net-centric configurations, where each agent is equipped with communication, computing, and sensing capabilities, and, by exchanging data with neighboring agents, wishes to gain situation awareness about the overall system of interest without the coordination of a fusion center. While decentralization provides benefits in terms of scalability, flexibility, and fault tolerance (as it does not need any coordination infrastructure), unfortunately net-centric surveillance and monitoring systems are still prone to some deliberate threats like the attack of an informatic virus. The goal of this chapter is to analyze this issue from several points of view. First, it will be shown how to model the virus spread over a sensor network and what the effects of a virus epidemic are on a net-centric data fusion algorithm; secondly, the correlations between network connectivity, distributed data fusion performance, and virus diffusion will be investigated; finally, it will be discussed how the effects of a virus attack can be mitigated by a careful redesign of the sensor network as well as of the distributed information fusion algorithm.