A Massive Clustering Algorithm for Fast Radio Resource Allocation to Distributed Units in O-RAN Networks

The progression towards fifth and sixth generation cellular networks (5G and 6G) radically alters network architectures, moving from traditional configurations to more dynamic and flexible structures. A key aspect of this transition is the disaggregation and virtualization of base station functionality into central units (CUs), distributed units (DUs) and radio units (RUs). The O-RAN Alliance is at the forefront of this change, introducing key innovations for radio access networks (RAN) in line with the vision of 3 GPP. This paper introduces a new massive clustering algorithm for fast radio resource allocation in O-RAN networks, focusing on fronthaul topology optimization using advanced graph learning techniques on a very large driving test dataset. Specifically, we leverage the Leiden Algorithm to improve the efficiency and effectiveness of RU/CU clustering, thereby improving optimization and resource management within the RAN Intelligent Controller (RIC). Our results demonstrate the efficiency in allocating virtualized DU/CU capacities, highlighting the potential of integrating graph learning into the O-RAN framework to meet the needs of next-generation cellular networks.