基于OpenFlow的跨域流媒体边缘云联合优化策略  

An Unified Optimization Strategy for Cross Domain Streaming Media Edge Cloud with OpenFlow

作　　者：江坤[1] 奚宏生[1] 

机构地区：[1]中国科学技术大学自动化系,安徽合肥230027

出　　处：《电子技术（上海）》2017年第7期10-16,共7页Electronic Technology

摘　　要：随着云计算技术的成熟,跨域边缘云模式下的流媒体服务正逐渐兴起。但流媒体边缘云的服务能力有限,需要通过合理的云资源调度,保证能够在用户请求波动的情况下提供高质量的流媒体服务。同时,由于各个区域用户请求分布和边缘云服务状态存在差异,需要通过用户请求的跨域转发避免局部拥塞和频繁的影片重部署,从而保证整体服务质量。另外,用户请求到达率随时间波动较大,为避免云资源的闲置和保证请求高峰时的服务质量,需要动态调整需租赁的云资源。因此,本文提出了子云内部资源调度,跨域用户请求分配以及总体云资源动态调整相协作的跨域流媒体边缘云联合优化策略。仿真结果表明本文策略能够保证用户服务质量,同时动态调整云资源租赁方案适应用户请求的变化。With the maturity of cloud computing technology, cross-domain edge cloud mode of streaming media services are gradually rising. As the service capability of the edge of the streaming media is limited, it is necessary to reasonably schedule cloud resource to ensure the provision of high-quality streaming media services in the case of user requests fluctuations. At the same time, due to the differences in the distribution of the users requests in each region and the service state of each cloud, it is required to avoid the local congestion and frequent video redeployment through the user request cross-domain forwarding. In addition, the number of user requests fluctuates over time, in order to avoid cloud resource idle and guarantee service quality at peak times, it is needed to dynamically adjust the cloud resources. Therefore, this paper proposes a cross-domain streaming media edge cloud joint optimization strategy with collaboration of intra-cloud resource scheduling, cross-domain user request allocation, and overall cloud resource dynamic adjustment. The simulation results show that the proposed strategy can guarantee the quality of user service, and adjust the cloud resource leasing scheme to adapt to the change of user request.

关 键 词：流媒体边缘云 资源调度 联合优化 请求分配 资源调整 

分 类 号：TP368.5[自动化与计算机技术—计算机系统结构]