Systematic review methodology for performance analysis optimization in software-defined networking/internet of things

Abstract

When merging Software-Defined Networking and the Internet of Things, performance evaluation becomes challenging due to IoT ecosystems' dynamic and heterogeneous nature. This extensive study examines the inefficiencies in data management, network congestion, and security challenges in Software-Defined Network-enabled Internet of Things devices. This study utilized a thorough literature review methodology, examining recent research implementing machine learning techniques for performance assessment across several domains. A comprehensive literature review methodology was applied to explore recent studies employing machine learning techniques for performance analysis through various application categories. Hybrid machine learning models, which integrate supervised and unsupervised techniques, are advised for effectively adapting to the highly dynamic nature of Internet of Things networks. Also, future research should focus on creating uniform performance evaluation metrics to make studies and applications more comparable.