项目名称: 需求感知的分布式缓存网络的研究
项目编号: No.61271199
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 无线电电子学、电信技术
项目作者: 李纯喜
作者单位: 北京交通大学
项目金额: 70万元
中文摘要: 内容服务商的对核心带宽的迫切需求与高成本的核心网扩容之间的差距越来越大;用户接入速率的大规模升级和3G用户的加入更加剧了核心带宽资源的需求危机;依赖既有的高带宽缓存架构无法从根本上解决这个问题。因此,本项目提出"需求感知的分布式缓存网络的研究"来支持大规模、高质量、具有商业模式的数据分发业务。在该研究中,核心网带宽用于备份内容的搬移,边缘服务器实现本地内容分发,通过在一定规划时间内去除冗余数据传输以及优化设计的缓存网络,将在未来十年内满足我国内容分发的需要。研究主要内容包括:基于大规模网络测量分析需求感知模型;研究从用户需求推断核心网带宽需求的模型和算法;研究从全局高度通过预先安排的办法将可预测的用户需求搬移到网缘去,并通过适当的抽象得到这类系统的理论模型。最终给出具有实际指导意义的基础设计模型。
中文关键词: 需求感知;边缘缓存;缓存网络;数据分发;多路径传输
英文摘要: The gap is growing between the urgent need for bandwidth of the backbone and the expensive expansion of the core network; bandwidth resource crisis is increased by the large-scale upgrade of users' access rate and the flush-crowd of 3G users. Traditional high-bandwidth based caching framework is incapable of dealing with this situation. In this project, we propose the research of demand-aware distributed caching network (DDCN) to deal with the problem. According to DDCN, backbone is employed to transport content, and edge caching nodes account for locally delivering content; by removing the redundancy in data transmission as well as the optimization design of caching network, it is capable of meeting the on-line data delivery needs in our country in next ten years. The main works in this research project include: to study the demand-aware model based on large-scale network measurement, to study the models and algorithms used to deduce the core-bandwidth demand from user requirements, and at last to study the optimization model which can make the whole caching network collaborate efficiently on transporting the predicted users' demands to their edge cache nodes.
英文关键词: demand estimating;edge caching;cache network;data distribution;multipath transmission