Abstract
There is an emerging interest in the Zero Trust Architecture (ZTA) concept for cyber–physical systems, including critical infrastructures, but its enforcement strategies are often judged in terms of information-technology metrics rather than physical system performance. Within communication-dependent control structures, policy enforcement points (PEPs) and policy decision points (PDPs) introduce delay, jitter, and possible loss of control authority that could affect closed-loop dynamics. This study provides a stability-aware case-study evaluation of Zero Trust enforcement semantics in a communication-dependent microgrid cyber–physical system. Applying a physics-based microgrid frequency control method, we contrast two typical enforcement strategies during PDP outage: fail-closed, which removes control authority, and a control-aware degraded mode that retains bounded corrective action. We use outage–delay degradation maps to quantify the worst-case performance over the added authorisation delay and the time of outage. Results show that, at the worst-case operating point within the explored delay-outage grid, fail-closed enforcement can amplify peak frequency deviation by up to 66.8% and increase integral absolute error by 71.4% compared to degraded bounded operation under identical conditions. Degradation arises not solely from latency magnitude but from the temporary structural removal of stabilising feedback. These findings demonstrate that, in the studied microgrid setting, enforcement semantics can significantly influence closed-loop behaviour under PDP impairment and therefore warrant explicit consideration when Zero Trust mechanisms are placed in communication-mediated control paths. The results motivate stability-aware validation of Zero Trust enforcement before deployment in time-critical cyber–physical control environments.