Multi-objective Optimization for Dynamic Virtual Machine Management in Cloud Data Center

Virtual machine (VM) management in cloud data center is an important problem that remains to be effectively addressed. There has been a considerable amount of work investigating the management of physical-to-virtual resource mappings to improve the efficiencies of resource usage and power consumption in data center. However, these different management objectives are conflicting. One solution can't get the optimal at the same time for each objective. In this paper, a multiobjective optimization approach is proposed to manage the dynamic mapping of VMs to physical resources in cloud data center. The main decisions required to solve this problem are when, which and where to move VMs. The decisions of when to migrate VMs are based on the sliding-window and the thresholds, the decisions of which VMs to be migrated are based on the different VM selection strategies, and the decisions of where to migrate VMs are based on the TOPSIS in order to balance the conflict between different objectives. Experimental results show that compared with other approaches, our multi-objective optimization approach can not only get the lower SLA violation, the smaller resource load and the less power consumption, but also have the least number of VM migration.