Safe and Secure Global Resource Management for Real-Time and Mixed-Criticality Systems
Gautam Gala, Gerhard Fohler · 2021
Real-time and safety-critical applications deployed on distributed systems with multicore nodes compete for system resources such as CPU, memory, and network. A resource management service with a global view is needed to monitor and make decisions to orchestrate these resources and provision them to the applications at run-time even in the presence of faults or varying demand and availability of the resources to ensure: (i) the real-time applications produce the desired output under stringent timing constraints, and (ii) the safety-critical applications behave as per their safety assurance levels. Previous approaches for such systems present techniques to monitor and orchestrate resources with centralized global resource orchestration decision-making. But they do not consider safety and security together for global resource management itself. We propose an approach based on distributed decision-making while considering safety and security to complement the existing work. We implemented our framework using the Sawtooth Blockchain platform.