A Distributed Content Placement Strategy Based on Popularity for ICN

Abstract

In most distributed systems such as ICN, proper content placement can improve network performance, reduce response time, and reduce source server load. Most ICN cache deployment strategies focus on selecting the appropriate node for caching on the content return path, regardless of the popularity of the content. This paper proposes a distributed content placement strategy based on popularity for ICN. The scheme not only considers the distance between the node on the content return path and the request node, but also considers the content popularity trend, predicts the future popularity of the content with the Markov Chain, and pushes the content with the local popular trend to the network in advance. This paper implements the cache placement strategy on the ndnSIM simulator and performs simulation experiments on the real network topology. At the same time, the comparison scheme greedy scheme is also implemented, and two indicators of network delay and hit rate are evaluated. The experimental results show that the network performance of this scheme is higher than that of the greedy scheme in the network stationary phase, with an average delay of 3% and an increase of 8% cache hit rate.