Limit-Caching: A Caching Scheme Based on Limited Content Popularity in ICN

Abstract

The explosive growth of network devices and content poses huge challenges to the current host-centric Internet architecture. Information-centric network (ICN) proposed a promising content-oriented network paradigm. The in-network caching in ICN reduces the content retrieval latency and network traffic as well as enhances the overall network performance. Therefore, cache management in ICN has attracted wide attention of researchers recently. But a caching scheme with lower network traffic and better adaptability remains to be studied. In this paper, we propose Limit-Caching, a caching scheme based on limited content popularity in ICN. In this scheme, we use tokens to control the insertion rate of content and encourage content providers to set high popularity levels for popular content, and routers probabilistically insert caches based on popularity levels. At the same time, the content that is repeatedly requested will be gradually pushed to the edge of the network. We also design Least Recently Used with Time (LRU-T) based on LRU replacement policy to adjust the caching probability. We conduct extensive experiments with a tree topology to evaluate our scheme. Simulation results show that Limit-Caching has better performance than several classic ICN caching schemes under multiple metrics.