Adaptive Hierarchical Meta-Scheduling for Distributed Time-Triggered Systems with Diverse Internal and System Architectures

Time-triggered (TT) systems are widely adopted in safety-critical applications for their predictability and fault containment. With the increasing deployment of TT nodes in distributed environments, adaptation becomes essential to address changing system conditions as they impact nodes differently. This paper presents a hierarchical approach to multi-node/multi-core adaptation in distributed TT systems with diverse internal architectures. Our hierarchical meta-scheduler leverages the system’s hierarchical topology to compute node-level and system-level adapted schedules. We introduce a superset reconvergence technique to minimise the size of the Multi-Schedule Graph (MSG) generated by the traditional meta-scheduler, minimising memory requirements in nodes with limited storage capacity. The scheduler employed by the hierarchical meta-scheduler is evaluated for performance and the superset reconvergence technique’s MSG size reduction rate. The results demonstrate the benefits of hierarchical adaptation and minimised memory consumption.