A novel energy-aware and resource efficient virtual resource allocation strategy in IaaS cloud

摘要

Virtual Machine Placement (VMP) is critical to the efficiency of the cloud infrastructure. In this paper, a server resource wastage model is established to reduce the server physical resource wastage, and the liner server power consumption model was applied to optimize the power consumption synchronously in IaaS. First and foremost, aimed to achieve the optimization on both energy consumption efficiency and physical resource wastage in IaaS at the same time, a novel virtual resource allocation strategy (VRAS) is established. The whole VRAS is composed of three parts: physical server selection with performance evaluation function, virtual machine list descending sort, and the virtual machine dynamic placement to different physical servers. Second, a multi-objective VMP algorithm based on dynamic programming strategy (DPVMP) was designed and implemented where the VMP with VRAS was transformed into a multi-stage knapsack problem. The experiments and algorithm evaluation were carried out on CloudSim platform. Experimental results show that the proposed algorithm makes less energy consumption and less resource wastage contrasted with other algorithms.