Service Caching in Multi-Tier Fog Radio Access Networks

Abstract

Fog computing, which performs data processing at the edge of the network, is a promising solution for increasing user satisfaction rates while decreasing latency and energy consumption. Similarly, in a multi-tier Cloud Radio Access Network (C-RAN), the network becomes much closer to the user, thereby improving signal quality and network capacity. The amalgamation of these techniques resulting in a multi-tier Fog Radio Access Network (F-RAN) significantly enhances the performance of fog computing. In this paper, we evaluate four performance metrics of service caching in a multi-tier F-RAN: Cache Hit Ratio (CHR), User Satisfaction Ratio (USR), latency, and energy consumption. For this purpose, a multi-objective problem is designed to maximize CHR and USR while minimizing latency and energy consumption. Clustered Collaborative Filtering (C-CF) is used, which first clusters the fog nodes of each tier/layer and then utilizes CF to predict the future popularity of the services. This approach ensures that the most popular services are cached in the fog nodes of each layer. Moreover, fog nodes cooperate, incorporating both intra-layer and inter-layer cooperation to provide services to users. Accurate prediction of future service popularity and optimal caching of services within the limited memory of the fog nodes help minimize latency and energy consumption. Numerical simulations show that compared to the conventional caching methods, the proposed technique brings significant performance improvements in terms of maximizing CHR and USR while minimizing latency and energy consumption.

Keywords

• F-RAN
• service caching
• cache hit ratio
• user satisfaction ratio
• latency and energy minimization