北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2015, Vol. 38 ›› Issue (1): 16-20.doi: 10.13190/j.jbupt.2015.01.003

• 论文 • 上一篇    下一篇

基于相关性概率的信息中心网络协作缓存策略

霍如1, 刘江1, 黄韬1, 陈建亚2, 刘韵洁1   

  1. 1. 北京邮电大学 网络与交换技术国家重点实验室, 北京 100876;
    2. 北京邮电大学 网络体系架构与融合北京市重点实验室, 北京 100876
  • 收稿日期:2014-03-26 出版日期:2015-02-28 发布日期:2015-03-30
  • 作者简介:霍如(1988—),女,博士生,E-mail:huoru1988@126.com;刘江(1983—),男,讲师.
  • 基金资助:

    国家重点基础研究发展计划(973计划)项目(2012CB315801);国家自然科学基金项目(61300184,61302089);中央高校基本科研业务费专项资金资助课题项目(2013RC0113)

Cooperative Caching Strategy Based on Correlation Probability in Information Centric Networking

HUO Ru1, LIU Jiang1, HUANG Tao1, CHEN Jian-ya2, LIU Yun-jie1   

  1. 1. State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2014-03-26 Online:2015-02-28 Published:2015-03-30

摘要:

信息中心网络(ICN)的节点缓存功能有助于海量内容的高效分发,缓解链路拥塞并减少流量冗余. 传统的缓存策略不利于提高全网缓存的内容多样性和缓存节点的平均命中率,现有研究能在一定程度上解决这些问题,实现了公平的内容流复用,但没有充分考虑节点之间的协作,导致节点缓存的利用率不均. 为了解决上述问题,从当前节点缓存状态对其他节点的影响入手,提出一种基于相关性概率的ICN协作缓存策略,根据路径及相邻节点信息做出本节点的缓存判断,从而有效控制缓存冗余. 仿真结果表明,该方法可以减轻服务器负载,丰富全网内容多样性,有利于提高交错复杂网络节点的命中率和利用率,减少请求跳数.

关键词: 信息中心网络, 协作缓存, 相关性, 概率

Abstract:

The in-network caching of information centric networking (ICN) can offer help to the massive contents distributions efficiently, the ease link congestion and the traffic redundancy reductions. Traditional caching policies are not conducive to raise content diversity of the whole network and average hit rate of cache nodes. The existing research can solve the problems to some extent, and achieve fair content flow multiplexing, however, almost no consideration of collaboration between nodes is adopted, which directly results in uneven utilization of cache nodes. To solve these problems, study on influence of caching status between neighbor nodes was made, and a cooperative caching strategy based on correlation probability in ICN was presented. Cache node will make caching decision based on the information of path and its neighbor node, so it can control the caching redundancy. Simulation shows that this method can ease the server load, and enrich the content diversity of the whole network. It is of benefit to improve the hit rate of nodes in complex network, and reduce the number of hops.

Key words: information centric networking, cooperative caching, correlation, probability

中图分类号: