SANR-CMF: Semantic-Aware Naming Resolution and Cache Management Framework in Information-Centric Network for Internet of Things

초록

In-network caching stands as one of the core pillars of the Named Data Networking (NDN) paradigm, where every internet router has the capability to store the incoming data packets in its cache along the path between data sources and consumers. This capability is particularly vital in Internet of Things (IoT) environments, where efficient and timely data retrieval is critical. Numerous strategies have been designed for caching IoT data and often rely on the Exact Match Policy (EMP), which strictly matches content names for data retrieval. However, semantic considerations of content names in EMP remain underexplored. Additionally, several caching mechanisms are modeled after traditional IP-based communication approaches, such as neighbor or topology-based centrality methods for content placement and data eviction for replacement. These approaches hold limited relevance in NDN, as content names can originate from anywhere in the network. To this end, we propose a novel semantic-aware naming resolution as an alternative to EMP in cache design. Furthermore, the proposed scheme introduces a content-based centrality mechanism, along with a backup strategy for cache management. Integration of semantics into caching strategies offers a more optimal approach for supporting NDN IoT. This combination can boost overall network performance. Simulation results illustrate that the proposed scheme outperformed the advanced existing studies CaDaCa and PoolCache in cache hit ratios, semantic satisfaction rate, latency, and route elongation factor. © 2025 IEEE.

키워드