基于内容流行度的ICN缓存策略性能验证与分析

Performance Verification and Analysis of Caching Strategies in ICN Based on Content Popularity

信息中心网络利用内建缓存提高内容分发效率。现有的缓存策略研究中,常常基于统计的方法获取内容流行度,并进行缓存位置的确定与内容的替换。然而,基于统计方法得到的是历史信息,不能直接用于决策,必须对流行度进行预测。本文的主要内容是对已有的基于统计与预测的流行度缓存策略的性能进行实验验证和分析,结果证明只有依靠统计信息预测内容流行度并据此设计的缓存策略,在性能上并不比简单的无先验信息的LRU策略显著优越。

Information-Centric Networking improves content distribution efficiency by using in-network cache.In the research of the existing caching strategies,it is often using content popularity based on statistical methods to make the decision of cache location and replacement.However,the statistical methods can only get historical information,and cannot be directly used for decision.So,popularity prediction is necessary.In this paper,we study existing caching strategies based on both statistical popularity and statistical popularity prediction,and find that caching strategies based on popularity prediction obtained by statistical methods,is not significantly superior to the simple LRU strategy without a priori information in system performance.