Resource management policy for cloud testbed of China railway

Cloud provides a shared resource pool for different independent or dependent applications. The virtualization techniques and power management techniques can enhance the flexibility with consolidate multiple virtual machines on the physical server. A key challenge is to automate the management of virtual servers while taking account of application requirements, resource utility and so on. This paper proposes a policy for control virtual machines configuration and resource allocation with the aim of dynamic placement of virtual machines and reducing the power consumption. We use utility function to establish the mathematical model to support decision for resource management and validate our approach through cloud testbed for railway application.