A new evolutionary multi-objective algorithm to virtual machine placement in virtualized data center
Chao Liu, Chenyang Shen, Sitian Li, Sinong Wang · 2014
Virtual machine (VM) placement is a key technologyto improve data center efficiency. Most works consider VM placement problem only with respect to physical machine(PM) or network resource optimization. However, efficient VM placement should be implemented by joint optimization of above two aspects. In this paper, a multi-objective VM placement model to minimize the number of active PMs, minimize communication traffic and balance multi-dimensional resource use simultaneously within the data center is proposed. The improved evolutionary multi-objective algorithm: NS-GGA is also designed to tackle this problem, which incorporates the fast non-dominated sorting of NSGA-II into the Grouping Genetic Algorithms. The simulation results show that, in most cases, our model and algorithm gains significantly in all aspects and yields better solutions compared to the existing methods.