Robust invariant sets for software rejuvenation via resampled dynamics

This article introduces a method that enlarges the safe operational space of a system within the context of software rejuvenation (SWR), an active cyber-defense method. The presented approach computes a polyhedral inner safe set obtained through the definition of a SWR cycle that is employed to derive simpler resampled dynamics for which a maximal robust invariant set can be computed. Also, constraint satisfaction in the original system is guaranteed by a convenient mapping onto the resampled system. As a result, the state trajectory in the original system is confined within a tube with time-varying section. This technique is applied to a discrete-time linear system and compared with other approaches, showing that the proposed inner safe set offers broader operational margins and therefore enables the system to perform its mission over an enlarged domain.