Semantic-Driven Intelligent Name Distribution & Retrieval for Named Data Networking

초록

Named Data Networking (NDN) adopts a content-centric paradigm for data retrieval, where content is accessed based on names rather than host addresses. While NDN enhances latency, security, and availability, it faces critical challenges, including name ambiguities, inefficient retrieval mechanisms, and issues related to name distribution. Existing frameworks and studies fail to effectively address the retrieval inaccuracies caused by semantically equivalent, however syntactically different FIB prefixes. Moreover, the relationship between name distribution and semantic similarity remains an underexplored aspect of NDN. To address this gap, this paper introduces an innovative approach that leverages Natural Language Processing (NLP) for semantic name matching, combined with a Feedforward Neural Network (FFNN)-based model for efficient name distribution and retrieval. By incorporating these techniques, results prove that the proposed method minimizes FIB memory overhead by 80-90%, decreases FIB retrieval time by 79.5%, and reduces prefix length by 40-45% as compared to vanilla NDN.