Hop-based Probabilistic Caching for Information-Centric Networks

Abstract

Transmission efficiency of the current Internet routing system has witnessed a continuous degradation with the increasing popularity of content-oriented applications (e.g., YouTube). As a future Internet architecture, Information-Centric Networking (ICN) is proposed to solve this problem partly by in-network caching which means ICN routers are aware of and able to cache the content being delivered. In this paper, we present a Hop-based Probabilistic Caching (HPC) for ICN. When delivering the data back to the requester, each content router caches the data with a probability determined by the number of the hops between the responder and the router itself. First, we design an approach to computing the hop-based caching probability with a small amount of overhead. Second, we provide a caching weight based on the mean residence time of the cached contents, which guarantees that the cached contents remain in a router for an expected duration. The simulation results show that HPC achieves a caching hit ratio of 60% and extensively improves the transmission efficiency, compared with 20%, 43% and 46% of LRU, LFU and ProbCache, a recently proposed scheme for ICN, respectively.