论文部分内容阅读
在大规模在线流媒体分发系统中,服务端需处理来自全球各区域的海量用户请求。现有混合云架构不能很好地满足日益增加的动态流媒体内容分发要求,需结合私有数据中心、云和内容分发网络3类平台,充分挖掘各平台的优势以降低费用并提高服务质量。针对基于3种平台的混合云,该文给出了多资源分配问题的描述,将其转化为Nash议价问题,从几何角度获取问题的高效求解算法,并基于实际商用环境中海量流媒体采样数据进行了模拟实验。实验结果表明:相比传统的混合云架构,该算法可显著提升服务质量,在动态和静态内容混合情况下可降低平均约40%的费用,可在包括大量动态流媒体内容场景中进行快速有效的资源分配。
In a large-scale online streaming media distribution system, the server needs to process massive user requests from all regions of the world. Existing hybrid cloud architectures do not adequately address the growing demand for dynamic streaming media content delivery. They need to leverage the benefits of each platform to reduce costs and improve quality of service, combining three types of platforms, including private data centers, cloud and content delivery networks. Aiming at the hybrid cloud based on three kinds of platforms, this paper gives a description of multi-resource allocation problem, transforms it into Nash bargaining problem and obtains the efficient solution algorithm from the perspective of geometry. Based on the sampling data of massive streaming media in real business environment A simulation experiment was carried out. The experimental results show that the proposed algorithm can significantly improve the quality of service compared with the traditional hybrid cloud architecture, reduce the average cost by about 40% when the dynamic and static content is mixed, and can be quickly and effectively performed in a large number of dynamic streaming media content scenarios Resource allocation.