一种基于逗留时间的新型内容中心网络缓存策略

发布时间：2018-03-24 10:12

本文选题：内容中心网络(CCN)　切入点：逗留时间　出处：《计算机学报》2015年03期

【摘要】：内容中心网络(Content-Centric Networking,CCN)作为一种以内容为中心进行路由、缓存的新型未来网络体系架构受到了广泛的关注.在CCN中,关键技术问题之一是网络缓存问题,现有方案主要采用ALWAYSLRU缓存策略,然而该策略容易出现相邻节点重复缓存的问题,使得网络整体缓存效率较低.针对这一问题,文中提出了一种基于逗留时间的新型缓存决定策略,设计了一种适用于CCN的合作缓存机制.在请求泊松到达的假设下,通过在单个缓存器对LRU(Least Recently Used)替换策略使用马尔可夫链建模,该文得到了内容在各缓存器平均逗留时间的近似计算公式.数值仿真结果显示,该方案相比传统缓存策略,有效地提升了网络缓存的效率和缓存内容的多样性,进而减少了用户请求服务的总跳数,增加了内容访问的命中率.
[Abstract]:Content-Centric networking CCNs (Content-Centric networking Network), as a kind of content-centric routing and caching architecture, have attracted much attention. In CCN, one of the key technical problems is network caching. The existing scheme mainly adopts ALWAYSLRU cache policy, but this strategy is prone to the problem of repeated caching of adjacent nodes, which makes the overall cache efficiency of the network low. In this paper, a new cache decision strategy based on sojourn time is proposed, and a cooperative caching mechanism for CCN is designed. By using Markov chain to model the LRU(Least Recently used) replacement strategy in a single buffer, an approximate formula for calculating the average time of stay of the content in each buffer is obtained. The numerical simulation results show that the proposed scheme is more efficient than the traditional cache strategy. It can effectively improve the efficiency of network cache and the diversity of cached content, thereby reducing the total number of hops requested by users and increasing the hit rate of content access.
【作者单位】：北京邮电大学网络与交换技术国家重点实验室;北京邮电大学北京市网络体系构建与融合重点实验室;
【基金】：国家“九七三”重点基础研究发展规划项目基金(2012CB315801) 国家自然科学基金(61300184,61302089) 中央高校基本科研业务费专项资金(2013RC0113)资助~~
【分类号】：TP393.02

【参考文献】