Algorithms and Architectures for Parallel Processing

Abstract

Due to the wide applications of IaaS (Infrastructure as a Ser- vice), energy-saving technologies of IaaS clouds has attractedmuch attention. However, it is very difficult for IaaS cloud providers to guarantee both of energy saving and performance under the condition of satisfying SLA (Service Level Agreement). Recently, in researches of Iaas cloud resource scheduling strategies, it is focused that SLA violation or overloaded host can trigger migrations of virtual machines. However, it is a new difficulty to resource scheduling among the physical machines that high variable workloads have to be conducted. Therefore, in order to schedule resource optimally, we propose a novel status-prediction-based framework, which seamlessly integrates the virtual machine migration optimal time theorem and the status prediction model of physical machines based on the hidden Markov process. Further, we address a resource scheduling algorithm based on the status predictionmodel on physical machines. Finally, through real experimental scenarios, we verify the effectiveness of the virtual machine migration timing prediction and the resource scheduling algorithm.