Caching and Machine Learning Integration Methods on Named Data Network: a Survey

Abstract

The caching mechanism is an essential part of future network design because it can improve the Quality of Experience (QoE) for users. Therefore, recent studies have examined the most appropriate caching techniques for future networks. Named Data Networks (NDN) is a future data-centric network that uses a cache mechanism to store packets of data in content stores. The main problem of traditional caching techniques cannot transmit large data packets, which high speed and changing depending on customers' requests. Undoubtedly, Machine Learning (ML) and deep learning (DL) algorithms play essential roles in many fields. Recent research adds ML or DL functions to cache decisions, such as cache replacement, content selection based on popularity, and cache placement. This paper performs an in-depth review of integration methods of caching and ML algorithms in future networks. The aim is to understand the goals, contributions, selection of learning algorithms, network topology, caching strategies, and their impact on improving network performance. This paper divides caching techniques into four categories to help readers understand the opportunities of the caching method. Furthermore, we discuss how a joint optimization strategy using ML and DL greatly impacted the network.