Social-Aware Collaborative Caching in Edge-User Networks: A Joint Multi-Graph Approach
Peng He, Shiguang Li, Chao Wang, Yaping Cui, Dapeng Wu, Ruyan Wang, Jifang Li · IEEE Transactions on Network Science and Engineering · 2024
Collaborative caching between Edge servers (ES) and user equipment (UE) is a key solution for tackling backhaul congestion, high latency, and low cache hit rates in modern communication networks. However, the limited storage capacity of ES and caching devices presents a challenge in devising efficient caching strategies. The main hurdle is to optimize caching strategies within these constraints to minimize average download latency across all nodes. This study proposes a three-tier network architecture for investigating edge-user caching techniques. Specifically, a collaborative caching strategy called multi-graph joint edge-user collaborative caching (MGJ-EUCC) is developed with the aim of reducing the average delay in accessing requested content. In MGJ-EUCC, three different graphs are firstly constructed, including a node-file preference graph (NFPG) to carve the preference of nodes, a social-aware graph (SG) to express the communication capability and social relationship, and a file attribute graph (FAG) to express the file similarity. Then, two preference factors are respectively modeled in spaces of nodes and files based on the constructed NFPG, SG, and FAG. The simulation findings validate the efficacy of the MGJ-EUCC approach in minimizing the average download delay and increasing the cache hit rate when compared to the current caching techniques.