A unified approach for fault tolerance and dynamic power management in fixed-priority real-time embedded systems

This paper investigates an integrated approach for achieving fault tolerance and energy savings in real-time embedded systems. Fault tolerance is achieved via checkpointing, and energy is saved using dynamic voltage scaling (DVS). The authors present a feasibility analysis for checkpointing schemes for a constant processor speed as well as for variable processor speeds. DVS is then carried out on the basis of the feasibility analysis. The authors incorporate important practical issues such as faults during checkpointing, rollback recovery time, memory access time, and energy needed for checkpointing, as well as DVS and context switching overhead. Numerical results based on real-life checkpointing data and processor data sheets show that compared to fault-oblivious methods, the proposed approach significantly reduces power consumption and guarantees timely task completion in the presence of faults.