The deployment of online virtual machines is an important part in cloud computing. In proper deployment can cause many problems including share conflict of internal resources(CPU, memory) and external resources(bandwidth), or even unbalanced load and network congestion. A scheme of server selection based on minimum maximum utilization rate of resources has been proposed to tackle the aforementioned problems. The goal of this scheme is to utilize the load balance between internal and external resource utilization rate, and secure a high availability of servers. The main idea is to select qualified servers which meet the virtual machine application resources, and conduct periodic extraction of the maximum resource utilization rate, and deploy the virtual machine on the server with lowest rate. The scheme has been compared to the BFD algorithm and random algorithm in several dimensions to prove it's advantages. |