CacheCraft: A Topology-Aware PageRank Centrality Algorithm for Cache Optimization in Named Data Networking

This study introduces CacheCraft, a novel approach for heterogeneous Content Store (CS) capacity allocation in Named Data Networking (NDN). Traditional NDN allocates CS capacity uniformly across routers, assuming equal storage requirements for all nodes. However, user content preferences and traffic patterns vary significantly, necessitating a more tailored allocation strategy. Additionally, the complexity of network topologies exacerbates the challenge, as static and homogeneous CS allocations lead to inefficiencies, increased latency, and reduced cache effectiveness in dynamic and dense networks. CacheCraft addresses these challenges by leveraging the PageRank algorithm to calculate the centrality of each node in the network. This centrality value determines the proportion of CS capacity assigned to each node, optimizing storage for nodes with higher traffic and strategic importance. The use of PageRank ensures scalable and reliable centrality computation, even in complex topologies. The performance of CacheCraft is validated across diverse network scenarios, including topologies of varying complexity, using metrics such as Cache Hit Ratio (CHR), average latency, and time complexity. Experimental results demonstrate that CacheCraft achieves an average improvement of 7.8% in CHR and a 5.6 ms reduction in latency compared to state-of-the-art methods. Moreover, CacheCraft maintains algorithmic computational efficiency, making it suitable for real-world deployment in complex and dynamic NDN environments. These findings highlight CacheCraft as a robust and scalable solution for optimizing NDN performance through adaptive and efficient CS capacity allocation. Doi: 10.28991/ESJ-2025-09-02-09 Full Text: PDF