Adaptive thresholds determination for saving cloud energy using three-way decisions

• Published in: Cluster computing Vol. 22; no. Suppl 4; pp. 8475 - 8482
• Main Authors: Jiang, Chunmao, Wu, Junwei, Li, Zhicong
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.07.2019; Springer Nature B.V
• Subjects: Algorithms; Cloud computing; Clusters; Computer Communication Networks; Computer Science; Decisions; Energy consumption; Information theory; Nodes; Operating Systems; Overloading; Power consumption; Processor Architectures; Regression analysis; Thresholds; Virtual environments; Three-way decisions; Cloud computing; Energy consumption; Two thresholds; Adaptive
• ISSN: 1386-7857; 1573-7543
• DOI: 10.1007/s10586-018-1879-7

Determining the two thresholds in three-way decisions is one of hot issues, it needs to be considered in a specific field and targetedly. In the research of cloud computing, on the one hand, it is necessary to pursue service quality and economy, on the other hand, there is a large number of problems such as low utilization of cluster and high energy consumption. In this paper, the adaptive thresholds migration (ATM) algorithm is proposed by introducing three-way decisions into virtual machine migration. First of all, the proposed method evaluate load in cluster according to resource usage, and then dynamically determine two thresholds, thereby the VMs will be divided into three types with the help of ideas of three-way decisions, that is, idle state nodes, the normal nodes and overload nodes. Secondly, for nodes in different regions take different acting. For the normal nodes, we do nothing. For the idle and overloaded nodes in the cluster, ATM determine the virtual machine to be migrated and optimize the choice of target host based on the balance of resource usage and transmission overhead. Experiments based on CloudSim show that the ATM algorithm reduces power consumption by about 20% compared to TCEA and MM.