A prediction-based coordination caching scheme for content centric networking

Edge caching is undoubtedly a promising solution to alleviate the core network data traffic. Pre-caching content in a terminal device closer to the user, reduces the peak data rate of the core network and makes the user's content request delay lower. However, most of the existing works lack a comprehensive consideration of users and contents, which lead to the caching performance of the overall system achieving a sub-optimal result. In this paper, we propose a scheme of taking both users content requests and content centric networking (CCN) caching strategies into consideration to improve the caching efficiency of the overall system. In this scheme, we first calculate the popularity of all contents based on the prediction results of users' content request distributions. Then, we use a novel prediction method, namely an echo state networks (ESN) as an machine learning framework. And then the caching contents are selected through different strategies and cached to caching units of small base stations (SBSs). We investigate the proposed scheme, and the simulation results indicate that our scheme has more than 8% improvement in terms of the cache hit rate and 15.2% decrease in terms of the average hop count, respectively, when compared to the existing CCN caching.