On-Demand Fog Caching Service for ICN Using Synthetical Popularity, Cost, and Importance Status

Abstract

Information-centric networking (ICN) is a significant next-generation network architecture with caching ability that greatly affects the network service performance. Because of limited caching space, on-demand cache configuration is an important issue to optimize redundant cache, shorten the time delay and reduce the traffic overhead. Existing works on caching mainly realize cache configuration in a single dimension such as content popularity or node weight, which cannot realize multilevel comprehensive decision-making for efficient on-demand caching services in complex ICN. Moreover, existing works just focus on centralized cache configuration, which causes low efficiency and high latency. To address aforementioned challenges, we propose an on-demand fog caching service for ICN, named FogCache, which is realized by perceiving synthetical content popularity, traffic and time cost, and node importance. Regional fog nodes are adapted to configure caching service at the edge of ICN, in which the additional complexity is low because just minimal byte information is added at the end of the data packets. FogCache achieves multiple dimensional synthesis of awareness unified by building a hierarchical factor structure into four layers: target layer, criteria layer, indicator layer, and observation layer. Simulation results show that the proposed FogCache gets better performances of caching than related schemes.