A PSO-Based Energy-Efficient Fault-Tolerant Static Scheduling Algorithm for Real-Time Tasks in Clouds
Pengze Guo, Ming Liu, Zhi Xue · 2018
Cloud computing provides dynamic and elastic computing resources for users. The scalability of clouds brings challenges for traditional distributed scheduling algorithm. Energy efficiency and reliability are two important issues in cloud systems. In this paper, we propose an energy-efficient fault-tolerant static scheduling algorithm based on particle swarm optimization (EFTP) for real-time tasks in clouds. In the EFTP algorithm, primary-backup (PB) model is used to realize fault tolerance. Backup overlapping technique is deployed to decrease the overhead caused by task duplication. An adaptive mechanism is put forward to choose proper computing resources. Particle swarm optimization (PSO) approach is extended for task allocation. Simulation results show that the proposed algorithm can achieve good energy efficiency and deadline guarantee ratio.