Fault-Tolerance and Scheduling in Embedded Real-Time Systems. Doctorat thesis (2021), Université de Batna 2
No Thumbnail Available
Files
Date
2021-07-29
Authors
Barkahoum KADA
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Recently, fault tolerance and energy consumption have attracted a lot of interest in the design of modern embedded real-time systems. Fault tolerance is fundamental for these systems to satisfy their real-time constraints even in the presence of faults and is needed because it is practically impossible to build a perfect system. Transient faults are the most common, and their number is dramatically increasing due to the high complexity, smaller transistors sizes, higher operational frequency, and lowering voltages. Dynamic voltage and frequency scaling (DVFS) is an energy saving technology enabled on most current processors. This work addresses the issue of fault-tolerant scheduling with energy minimization for hard real-time embedded systems. Our first proposition is an efficient fault tolerance approach that combines two well-known methods: active replication and checkpointing with rollback. Based on this approach we have proposed two algorithms. Static Fault-Tolerant Scheduling algorithm SFTS that explores hardware resources and timing constraints to tolerate multiple transient fault occurrences with respect to hard real-time constraints of precedence-constrained applications. Dynamic Voltage and Frequency Scaling Fault-tolerant Scheduling algorithm DVFS-FTS is proposed to satisfy real-time constraints and to achieve more energy saving even in the presence of faults by adapting the DVFS technique. According to the simulation results, the proposed algorithms have been shown to be very promising for emerging systems and applications where timeliness, fault tolerance, and energy reduction need to be simultaneously addressed.