An Auction-Based Caching Decision Algorithm for IoT Traffic with Popular and Fresh Content
The exponential growth of Internet of Things (IoT) devices introduces significant challenges to data transmission efficiency, including network congestion, data retrieval latency, and maintaining content freshness. Named Data Networking (NDN), a key architecture within Information-Centric Networking (ICN), has emerged as a promising solution by leveraging in-network caching. However, IoT data traffic poses unique challenges due to its highly dynamic content popularity and stringent freshness requirements. Efficient caching strategies must balance these factors, as frequently requested content can quickly lose relevance, and fresh content is often only valid for short periods.
To address these challenges, this research proposes the Auction-Based Caching Decision (ABCD) algorithm, a novel caching and replacement strategy tailored for IoT data. The ABCD algorithm integrates content popularity and freshness into an auction-based decision-making framework, where content dynamically bids for caching resources. A recency-frequency model combined with a freshness-aware metric is used to calculate bid values, enabling the algorithm to prioritize high-demand and time-sensitive content. When the cache is full, ABCD replaces the least competitive cached content based on its bid value, ensuring an optimal balance between retaining popular items and accommodating fresh content.
Extensive simulations using NDNsim validate the performance of the ABCD algorithm. Compared to benchmark strategies such as LCE, LCD, ProbCache, and CFPC, the proposed algorithm achieves significant improvements in cache hit ratio, data retrieval latency, and content freshness. These results highlight the potential of the ABCD algorithm to enhance the efficiency of in-network caching and replacement mechanisms in dynamic, resource-constrained IoT environments.