The Software-Defined Networking (SDN) architecture marked a milestone by decoupling the control plane from the data plane, providing telecommunications infrastructures with unprecedented centralized programmability. However, the explosion of heterogeneous traffic and the demanding latency requirements of applications such as IoT and 5G/6G networks pose dynamic challenges that exceed the capacity of static policies and traditional manual management.
This conference explores the evolution towards the emerging paradigm of Knowledge-Defined Networking (KDN), which introduces a “Knowledge Plane” powered by Artificial Intelligence (AI) based on SDN. We will analyze how advanced techniques, in particular Deep Reinforcement Learning, are enabling autonomous traffic engineering and proactive detection of security threats.
In addition, we will examine the trends that are shaping this future, such as Intent-Based Networking (IBN) and the transformative potential of Generative AI for operations automation. Finally, we will critically address the main challenges that must be overcome to materialize this vision, focusing on the explainability (XAI) of AI models and their generalization capacity, fundamental pillars to build trust and viability in fully autonomous 6G networks.

Eng. Ariel Hernán Roel
Country: ArgentinaElectronics engineer from the National University of La Plata. Senior Consultant at CONSULTEL. Full Professor of the National University of Lanús of Argentina, for the major of Telecommunications Engineering. Professor of the "Diploma Program in Telecommunications Networks and Services 4.0" of the National Technological University of Argentina. He has participated as the project leader of the Online Charging System (OCS) of the Telecommunications Company of Cuba S.A. (ETECSA) during 2013 and has taught courses for ETECSA. He participated in projects for the government of the city of Buenos Aires in Argentina in 2010 and in 2019 he was a professor at the University of Palermo. He participated in the CMI (Communication, Media & Information Technologies) Congress edition 14, of the Aalborg University of Denmark.