Cyber Physical Systems (CPS) Surveillance Using An Epidemic Model

Vast investments have recently been made worldwide in developing the Cyber-Physical System (CPS) technology with the expectations of improving economical and societal structures. However, great care must be paid to the CPS' complexity, the impact of emerging IoT (Internet of Things) protocols on the CPS infrastructure as well as the impact of information dissemination by these protocols on the safety of these infrastructures. This paper addresses the issue of CPS safety by proposing and evaluating the performance of a CPS management framework and the analysis of the dynamics of the underlining IoT network in the cyber-space. The main contributions of this paper are in threefold. Firstly, a new CPS framework is proposed; that: 1) builds around a layered architecture to compartmentalise the CPS functionalities into different modules for efficiency and scalability and 2) uses an inner feedback loop for the efficient management of CPS infrastructure. Secondly, building upon this framework, a novel diffusion model that uses the epidemic (interference) sets to produce accurate diffusion patterns across the CPS IoT subsystem is proposed. Finally, the proposed diffusion model is numerically analysed to show how it can be used to achieve efficient CPS surveillance in order to trigger reconfiguration to re-optimise the CPS when it is under stress. in IoT settings. The numerical analysis of the diffusion model shows that interference propagates in pairwise disjoint sets, with IoT nodes migrating from "susceptible" to "attacked" statuses and finally reaching the "removed" state at a predictable time. Deployment considerations on some of the current social and public networks are also onsidered