Dynamic VMs placement for energy efficiency by PSO in cloud computing

• Published in: Journal of experimental & theoretical artificial intelligence Vol. 28; no. 1-2; pp. 97 - 112
• Main Authors: Dashti, Seyed Ebrahim, Rahmani, Amir Masoud
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 03.03.2016; Taylor & Francis Ltd
• Subjects: Architecture; Artificial intelligence; Cloud computing; Computer simulation; Consumers; dynamic resource allocation; Energy conservation; Energy efficiency; energy efficient; Energy management; live migration; provisioning; Task scheduling; Tasks; virtual machine placement
• ISSN: 0952-813X; 1362-3079
• DOI: 10.1080/0952813X.2015.1020519

Recently, cloud computing is growing fast and helps to realise other high technologies. In this paper, we propose a hieratical architecture to satisfy both providers' and consumers' requirements in these technologies. We design a new service in the PaaS layer for scheduling consumer tasks. In the providers' perspective, incompatibility between specification of physical machine and user requests in cloud leads to problems such as energy-performance trade-off and large power consumption so that profits are decreased. To guarantee Quality of service of users' tasks, and reduce energy efficiency, we proposed to modify Particle Swarm Optimisation to reallocate migrated virtual machines in the overloaded host. We also dynamically consolidate the under-loaded host which provides power saving. Simulation results in CloudSim demonstrated that whatever simulation condition is near to the real environment, our method is able to save as much as 14% more energy and the number of migrations and simulation time significantly reduces compared with the previous works.