多媒体云计算环境下基于DCN的虚拟服务器动态协同迁移方法研究

项目名称： 多媒体云计算环境下基于DCN的虚拟服务器动态协同迁移方法研究

项目编号： No.61472317

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 张未展

作者单位： 西安交通大学

项目金额： 82万元

中文摘要： 多媒体云计算技术，已成为部署超大规模多媒体应用最有效的途径之一，具有重要的研究与应用价值。然而，相关研究未能综合考虑多媒体云服务的流量特性与云计算平台的基础架构，难以实现云计算资源的优化利用。本项目拟以应用情境感知的虚拟化云计算资源调度为立足点，在数据中心网络（DCN）环境下，研究虚拟服务器的动态协同迁移方法，全局优化云计算资源利用。研究内容：1）基于DCN流量感知的虚拟服务器集群拓扑动态生成；2）虚拟服务器集群资源消耗评价模型与负载状态动态判别；3）基于负载状态与流量特征的虚拟服务器动态协同迁移；4）基于OpenStack的动态调度系统原型。创新点：1）以多媒体云服务的数据特征 动态流量为研究对象，感知上层服务器集群拓扑；2）以部署环境的实时状态DCN资源消耗为核心依据，动态判别虚拟服务器集群的负载状态；3）以体现服务器集群关联特性的协同迁移为具体手段，全局优化平台系统资源。

中文关键词： 多媒体云计算；数据中心网络；虚拟机协同迁移；流量感知

英文摘要： Multimedia cloud computing has become one of the most effective ways to deploy large scale media streaming applications, with significant research and application value. However, the related studies fail to consider the flow characteristics of multimedia cloud services and the infrastructure of cloud platform together. Therefore, they are difficult to achieve an optimal utilization of cloud computing resources. The project intends to apply application-aware scheduling of virtualized cloud computing resources, and studies the dynamic scheduling method of the VM-based servers in the data center network (DCN) environments, so as to globally optimize the cloud computing resources. To be specific, our research is composed of the following four facets: 1) DCN traffic-aware topology generation of virtual server clusters; 2) Resource consumption evaluation model and load status determination of the virtual server clusters; 3) Dynamic deployment and collaboration migration of the virtual server clusters; 4) Dynamic virtual server scheduling prototype based on OpenStack. The innovation points: 1) to perceive upper server cluster topology by studying the dynamic traffic feature of the DCN; 2) to dynamic determine the state of the virtual load server clusters by considering the DCN resource consumption; 3) to globally optimize the cloud computing resources by employing the collaborative migration means.

英文关键词： Multimedia cloud computing;Data center network;Virtual machine collaborative migration;Traffic-aware