云计算架构下基于BP神经网络负载预测策略的研究  

Research on Load Forecasting Strategy Based on BP Neural Network Under Cloud Computing Architectures

作　　者：王晨辉[1] 张晓亮[1] 梁晓传 

机构地区：[1]国家电网公司信息通信分公司,北京100761

出　　处：《电力信息与通信技术》2016年第11期46-50,共5页Electric Power Information and Communication Technology

摘　　要：目前,传统的数据中心运行方式暴露出在资源分配、管理灵活度、资源利用率、应用服务质量等方面的问题。随着虚拟化技术的成熟,采取基于云计算架构的资源池方案,可对虚拟化资源进行统一管理、调配,达到自动化、智能化管理信息系统的目的。文章在介绍云计算、虚拟化技术的基础上,着重探究基于反向传播(Back Propagation,BP)神经网络预测模型的虚拟机池负载均衡方案,并用仿真实验验证方案的可行性。仿真实验表明,预测模型能够在误差允许的范围内准确预测物理机的负载情况,具有一定的应用价值。Nowadays, there are some problems on resource allocation, management flexibility, resource utilization and application service quality in data center with the traditional operating mode. With the development of virtualization technology, adopting resource pool scheme based on cloud computing architecture could conduct managing and deploying virtual resources uniformly and achieve the goal of controlling information system automatically and intelligently. On the basis of introduction for cloud computing and virtualization technology, this paper studies the load balance scheme of virtual machine pool based on BP neural network predicting model and verifies its feasibility by simulation. Simulation results show that the prediction model can accurately predict the load of the physical machine in the allowed range of error, and it has a certain application value.

关 键 词：云计算 虚拟化 虚拟机池 BP神经网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]