Benchmarking Classical and AI-based Caching Strategies in Internet of Vehicles

Abstract

Edge caching has emerged as a promising approach to deal with the redundant traffic, to improve the Quality of Service (QoS) and to optimize the energy use in Internet of Vehicles (IoV). However, the intrinsic storage limitations of edge servers pose a critical challenge for IoV edge caching scheme. To solve these issues, several caching policy management techniques have been proposed in literature. In this paper, we perform a systematic comparison among the recent Artificial Intelligence (AI) based caching approaches and the classical caching techniques for IoV. Our objective is to provide a roadmap for choosing the best caching strategy for a given constrained environment. Through a practical scenario, the simulation results show that AI-based edge caching methods achieve high performance in terms of total content access cost and edge hit rate while maintaining a relatively low average delay. On the other hand, hash routing strategies tend to maximize the edge hit rate to the detriment of delivery latency.