An Efficient Method for Maximizing Total Weights in Virtual Machines Allocation

When allocating virtual machines (VMs) in data centers, weights such as profits and other benefits are associated with all VMs. This paper considers maximizing the total weight in VMs allocation. As virtualization widely adopted in Cloud computing, requests may only consume part of the total capacity of a single hardware resource (for example a physical machine), this requires a new model for maximizing the total weight. In this paper, for the first time we model this problem as shared interval scheduling for capacity proportional weight and propose an exact efficient algorithm for it with computational complexity O(n2) where n is the number of jobs. The proposed method has good scalability and can be applied to maximize the total weight or related metrics in cloud computing.