J4 ›› 2014, Vol. 41 ›› Issue (6): 131-136+187.doi: 10.3969/j.issn.1001-2400.2014.06.022

• 研究论文 • 上一篇    下一篇

一种用于内容中心网络的缓存随机放置策略

胡骞;武穆清;郭嵩;彭蠡   

  1. (北京邮电大学 网络体系构建与融合北京市重点实验室,北京  100876)
  • 收稿日期:2014-01-02 出版日期:2014-12-20 发布日期:2015-01-19
  • 通讯作者: 胡骞
  • 作者简介:胡骞(1988-),男,北京邮电大学博士研究生,E-mail:huqian@bupt.edu.cn.
  • 基金资助:

    网络体系构建与融合北京市重点实验室主任基金资助项目(2013BKL-NSAC-ZJ-06)

Random cache placement strategy for content-centric networking

HU Qian;WU Muqing;GUO Song;PENG Li   

  1.  (Laboratory of Network System Architecture and Convergence, BUPT, Beijing  100876, China)
  • Received:2014-01-02 Online:2014-12-20 Published:2015-01-19
  • Contact: HU Qian

摘要:

以内容为中心的网络是一种新型网络架构,它打破了传统的“主机—主机”通信模式,将内容本身作为网络的核心.以内容为中心的网络通过分布式的缓存机制对内容业务提供有效支持,缓存的放置策略也成为研究的热点.首先提出一种分布式的缓存随机放置策略,基于节点与内容源之间的跳数随机选择缓存放置节点,实现节点缓存的负载均衡;然后在该策略下推导缓存命中率与内容热度分布、请求到达率、缓存大小的关系.数值计算及仿真结果表明,该策略可以提高缓存命中率,与以内容为中心的网络传统策略相比,获取内容所需的平均跳数可减少5.59%~6.86%.

关键词: 内容中心网络, 缓存, 缓存放置策略, 缓存命中率, 分布式算法

Abstract:

Content-Centric networking (CCN) is a new network architecture, which has broken the traditional end-to-end communication mode and focuses on contents. In CCN, in-network caching can enhance content delivery on a global scale without requiring a special infrastructure. In this paper we propose a distributed random cache placement strategy, in which caching nodes are selected based on the distance to the content provider. Then the expression for the cache hit ratio is derived as functions of the content request rate, content popularity distribution and cache size in each node. Numerical analysis and simulations show that our strategy can improve the cache hit ratio and reduce average hops to get contents by 5.59%~6.86%.

Key words: content-centric networking, cache memory, cache placement, cache miss ratio, distributed algorithms

中图分类号: 

  • TP393