Research On Virtual Machine Scheduling Algorithms For Cloud Data Centers

Cloud computing technology has become a hot issue around the world in recent years.Currently,the cloud computing data centers are faced with a number of key issues,and the management of virtual machine resources is of utmost importance in the industry and academia research.With the rapid development and the growing popularity of cloud data centers,the importance of the virtual machine resource management is showing a gradual upward trend.Virtual machine scheduling problem is an important issue in cloud computing,directly related to the stability,efficiency of resource,users’ satisfaction and operating costs of cloud services.This paper mainly focuses on virtual machine scheduling for cloud computing data centers.The contributions of this paper are as follows:(1)SLA-based virtual machine allocation algorithmCloud computing is a model of business,for cloud providers,the full use of resources,maximizing profit is the main goal of their pursuit.Characteristics and SLA requirements of different application loads both are different.How to consider the properties of the application load characteristics,the number of required CPU,bandwidth,memory,and storage to allocate virtual machines,which will be a challenge.Our algorithm is a real-time scheduling of virtual machines,which uses the different characteristics of application loads for cloud computing data centers,and considers different requirements of the virtual machines and meets the SLA requirements.The algorithm takes into account the characteristics of the application loads,and considers SLA parameters such as the number of required CPU,memory,and storage resources.We allocate the virtual machine for the application from the user’s point of view.The purpose of our algorithm is minimizing the cost to complete the application standing user’s point of view,trying to meet the QoS requirements of the application and maximizing resource utilization at the same time.(2)Multi-objective energy-efficient VM placement algorithm for cloud data centersThe mapping between virtue machine and physical nodes determines how to allocate the virtualized resource to multiple applications for data centers.And the mapping has a significant impact on the performance of cloud systems,power consumption and QoS assurance.The algorithms mainly study initial placement and the dynamic management of virtual machines.Two algorithms are proposed for two basic steps of virtual machine placement problem,which is to initialize static allocation and dynamic migration.We place the virtual machine to the physical machine from the providers’ point of views.We ensure the low energy consumption and take into account the service level objectives,the number of nodes to use,the number of virtual machine migration and other multi-target optimization when we do the initial distribution and the dynamic management of virtual machines.The contribution of this paper is to propose a energy-efficient virtual machine placement algorithm for cloud data center,taking into account the multi-dimensional resource constraints,such as CPU,memory,network bandwidth.Experimental results show that our proposed algorithm can make full use of various resources in cloud data centers,while saving energy as much as possible and meeting the service quality requirements(QoS).Our algorithms realize multiple targets,and have a promising prospect in application.