关键词: 三支决策, 云计算, 虚拟机迁移, 负载均衡, 云能耗

Abstract: As an important way to reduce energy consumption of the data center in cloud computing, virtual machine migration is widely used. By combing the trisecting-acting-outcome model of three-way decision, a Virtual Machine Migration scheduling strategy based on Three-Way Decision(TWD-VMM) was proposed. First, a hierarchical threshold tree was built to search all the possible threshold values to obtain the pair of thresholds with the lowest total energy consumption with the data center energy consumption as the optimization target. Thus, three regions were created:high-load region, medium-load region and low-load region. Second, different migration strategies were used for hosts with different loads. Specifically, for high-load hosts, the multidimensional resource balance and host load reduction after the pre-migration of hosts were adopted as targets; for low-load hosts, the host multidimensional resource balance after pre-placing hosts was mainly considered; for medium-load hosts, the virtual machines migrated from other regions would be accepted if they still met the medium-load characteristics. The experiments were conducted on CloudSim simulator, and TWD-VMM was compare with Threshold-based energy-efficient VM Scheduling in cloud datacenters(TVMS), Virtual machine migration Scheduling method optimising Energy-Efficiency of data center(EEVS) and Virtual Machine migration Scheduling to Reduce Energy consumption in datacenter(REVMS) algorithms respectively in the aspects including host load, balance of host multidimensional resource utilization and total data center energy consumption. The results show that TWD-VMM algorithm effectively improves host resource utilization and balances host load with an average energy consumption reduction of 27%.

Key words: three-way decision, cloud computing, virtual machine migration, load balancing, cloud energy consumption