The issue of reducing energy consumption is one of the most important limitations of the design of embedded systems, and for this researchers introduced the technique of "dynamic voltage and frequency Scaling (DVFS)", which is considered one of the most popular technologies for reducing energy consumption. As this technology works to reduce the frequency of the processor used to execute the task, and thus reduce the energy consumed. However, this downscaling reduced the reliability of the system by increasing the processor's transient errors. This poses a new challenge because reliability is a fundamental issue in real-time safety- critical systems, and implementing unreliable application can lead to dire consequences. This paper aims to implement energy-efficient fault-tolerant scheduling for a reliable parallel application on heterogeneous distributed real-time systems, where the parallel application is described by a directed acyclic graph (DAG). To achieve this, this paper proposes an energy-efficient real-time system tasks scheduling algorithm while meeting reliability requirements (EESARR). To verify the efficiency and capability of the proposed algorithm, The proposed algorithm has been compared with the ESRG algorithm, which seeks the same goal in a different manner. The experimental results confirm that the energy consumption was reduced by the proposed EESARR algorithm more than the studied ESRG algorithm.
Al-Baath University Journal.
2021.
Energy-efficient Fault-tolerant Scheduling of Parallel Applications on Distributed Real-time Systems
Students
Teachers
Graduates
Faculties
