Intelligent Caching Mechanism in Edge-Cloud Networks

The explosion of data-heavy applications like video streaming, augmented reality, autonomous driving, and Internet of Things (IoT) has imposed serious loads on conventional cloud- based architectures, resulting in longer latency, bandwidth center, and scaling problems. The Edge-Cloud networks overcome these problems by moving the computation and storage to the end users and caching is central in the reduction of latency and backhaul traffic. Conventional policies such as Least Recently Used (LRU), Least Frequently Used (LFU) and First-In-First-Out (FIFO) are very easy to understand but they do not adapt to the dynamic content popularity, user mobility, and service level demands. This paper introduces a new caching policy, which is Explainable Hybrid Popularity-Rule Caching (X-HPC) of cloud-edge settings. X-HPC implements a lightweight popularity prediction model in the cloud layer, and a rule engine in each edge node. The decision engine uses the popularity predicted, size of the content, Quality of service (QoS) class, and recency-based retention to decide on deterministic caching and eviction, with clear written explanations. We provide a problem formulation, the data sets, architecture, algorithm steps implementation, and a comparative performance analysis with LRU, LFU, FIFO, and popularity- based caching. The simulation outcomes show that X-HPC has high hit ratio (83 Percent) and low latency (54 ms) at the normalized cache size of 0.2, which is better than the current methods. The effectiveness of our solution in intelligent caching in 5G/6G edge-cloud networks is confirmed by experimental studies.