A New Multi-objective Virtual Machine Deployment and Reconfiguration Framework for Cloud Systems
Xiaolin Liu, Xiyang Liu, Qingshan Li · International Conference on Control Engineering and Communication Technology · 2013
In this paper, a VM deployment and reconfiguration optimization framework is proposed, which is constituted by two steps: initial VM deployment and runtime VM deployment. In initial VM deployment, a multi-objective model to maximize resource usage and balance multi-dimensional resources use is constructed and a Grouping Genetic Algorithm is introduced. In runtime VM deployment, a single-objective model with the same objectives is presented to shorten the VM migration time, which is simplified by a novel technique: Average valued inequality. The Improved Genetic Algorithm with local heuristic method is also proposed to solve this model. The simulation results show that our framework is applicable in maximizing resource usage, balancing multi-dimensional resources use and reducing the time costs of the runtime reconfiguration.