A Flattened-Priority Framework for Mixed-Criticality Systems

The ever-increasing integrated services in industrial control systems lead to diverse performance objectives and multiple traffic characteristics. Such systems are called mixed-criticality systems (MCSs). Priority-based scheduling is a commonly used mechanism to achieve differential QoS in MCS. While MCSs have supported prioritization as a standard feature, traditional priority-based schedulers incur significant complexity, especially due to the widespread adoption of mixed-criticality applications in embedded systems. It is desirable to lower the complexity of MCSs under the limited computing resource of smart devices. This article presents a novel technique to design resource-efficient priority schedulers for MCSs. We propose a flattened-priority framework to transform an unprioritized scheduler into a priority-based one. The framework works iteratively based on feedback loops. For P priorities, the transformed scheduler converges in P iterations in the worst case. With the proposed framework, the design of priority schedulers is simplified into the design of nonpriority schedulers. Such a simplification dramatically lowers the design effort and system complexity. A case study was performed on an FPGA-based Industrial Ethernet switch that is a typical MCS. Compared with the previous priority schedulers, the transformed priority scheduler by the proposed framework achieves a 30%-50% reduction in the resource usage of lookup tables without performance loss of forwarding delay.