Models of Resilient Systems with Online Verification Considering Changing Requirements and Latent Failures

The paper presents the outcomes of developing and implementing a cybernetic approach to creating and deploying resilient systems. It discusses the concepts and principles of real-time evolution for resilient systems, considering changing requirements, environments, and latent failures. This increased resilience stems from the necessity of integrating additional channels into the control system to address shifts in requirements, environments, or unspecified faults and failures. The general structure of resilient systems is founded on the principle of partitioning and managing channels, as well as system reconfiguration based on functional and non-functional characteristics. A case study of a space resilient system with online verification is presented. Three scenarios of system behavior to ensure resilience are proposed, considering various tolerance and management mechanisms. The paper analyzes the results of developing and researching Markov models for these scenarios, offering insights for enhancing the availability of resilient systems.