The emergence of 6G networks introduces a new level of complexity by requiring robust and adaptive solutions for network management. Although Artificial Intelligence (AI) and Machine Learning (ML) approaches can support dynamic network conditions, their dependence on large datasets, lack of transparency, and high computational demands limit their effectiveness in real-world applications. Accordingly, this paper presents knowledge-defined networking (KDN) as a superior approach that combines domain-specific knowledge with AI/ML capabilities to enhance network management performance. The proposed KDN architecture consists of four modular planes—Data, Control, Knowledge, and Management—that interact seamlessly to improve decision-making and management. Through a comparative analysis, this study highlights the benefits of KDN in routing management, including higher packet delivery ratios (up to 21% improvement), reduced latency (up to 32% lower), lower energy consumption (up to 27% savings), and improved adaptability (up to 36% enhancement) in changing network conditions. Empirical results from a simulated 6G environment show that KDN consistently outperforms other AI/ML approaches. These results support KDN as a crucial framework to overcome the limitations of AI/ML for intelligent and reliable network management.
Tuğçe Bilen (Sat,) studied this question.