Network Optimization for Safety-Critical Systems Using Software-Defined Networks

Software-Defined Networking allows to separate traffic handling from network management. This - in combination with potential cost savings - makes it interesting for areas for which it has not been originally designed: safety-critical systems such as aeroplanes or power grids. These require resilience against faults and failures as well as predictable timing and availability. Network optimization provides a mean to incorporate these demands during the design stage of critical systems while taking limitations such as capacities into account. This paper focuses on obtaining network configurations that satisfy the demands of safety-critical systems. To this end, this paper studies example topologies of both critical and non-critical systems to investigate the effect of resilient routing on network and traffic parameters and solve a minimum cost linear optimization problem that incorporates constraints of safety-critical traffic. The results thus obtained are then compared with a capacity-constrained and an Earliest-Deadline-First placement heuristic. Hence it can be shown that while heuristics can perform well in some aspects, they violate either capacity or timing constraints, thus making them unsuitable for networks that provide safety-critical services.