面向海量流媒体信道资源分配快速Nash议价算法

发布时间：2019-01-27 20:29

【摘要】：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。
[Abstract]:In a large scale online streaming media distribution system, the server needs to deal with a large number of user requests from all regions of the world. The existing hybrid cloud architecture can not meet the increasing demand of dynamic streaming media content distribution. It needs to combine private data center cloud and content distribution network to fully exploit the advantages of each platform in order to reduce the cost and improve the quality of service. For the hybrid cloud based on three platforms, this paper presents a description of the multi-resource allocation problem, which is transformed into a Nash bargaining problem, and an efficient solution algorithm for the problem is obtained from the geometric point of view. Simulation experiments are carried out on the basis of mass streaming media sampling data in commercial environment. The experimental results show that compared with the traditional hybrid cloud architecture, the proposed algorithm can significantly improve the quality of service and reduce the average cost by about 40% in the case of dynamic and static content mixing. Fast and efficient resource allocation can be made in scenarios that include a large number of dynamic streaming media content.
【作者单位】：河南工业大学信息科学与工程学院;
【基金】：国家自然科学基金资助项目(61472460,61702162,U1504607) 河南省高校科技创新团队支持计划(17IRTSTHN011) 河南省教育厅科学技术研究重点项目(17A520004,16A520006) 河南省科技厅科技攻关项目(172102110013) 河南工业大学自然科学基础研究重点培育计划项目(2014JCYJ04,2015XTCX03);河南工业大学博士基金(2012BS011) 粮食信息处理与控制教育部重点实验室开放基金课题(KFJJ-2016-104)
【分类号】：TP393.09

【相似文献】