Artificial Intelligence Approach for Name Classification in Information-Centric Networking-based Internet of Things

Abstract

Content management has continuously been among the most challenging problems, especially on the Internet of Things (IoT), where devices are set to be "hungry" for content. In this context, Information-Centric Networking (ICN), a promising Future Internet Architecture, can facilitate the IoT requirements of continuous and long-lasting connectivity, which has burdened the IP-based network system. In ICN, a content naming structure is composed to direct content requests to the nearest content provider with the built-in in-network caching function, storing the content. For a successful implementation of ICN-based IoT, an intelligent algorithm approach for IoT-based ICN implementation aiming to improve content management is proposed in this study. We establish a hybrid ICN with the most suitable machine learning algorithms satisfying the requirements to realize a feasible IoT technology. The selected algorithms from Supervised Learning, Unsupervised Learning, and Reinforcement Learning are evaluated before being chosen as the content forwarding process. The numerical findings show the superiority of the Extended Learning Classifier System under Reinforcement Learning’s scheme compared to the other algorithms.