Machine Learning: A new way for material resources orchestration in a large-scale V-RAN

Radio Access Network (RAN) virtualization is one of the key concepts of 5G Networks and beyond. It provides solution to energy cost problems as well as material and radio resource management. Despite the research made for its apprehension, the RAN virtualization remained not a reality. The existing attempts are still embryonic, limited to an extremely low number of virtualized entities (BBU, RRH), user devices and an abstract fronthaul. In a previous paper, we proposed the first large-scale V-RAN implementation composed of a large-scale vBBU Pool, a large number of RRHs and providing a wide mMTC service. In this paper, we implement in this large-scale V-RAN two network slices: the mMTC and the eMBB. Then, we propose the Machine Learning (ML) as a new way of orchestration for the different V-RAN resources through a two-step ML algorithm development. Finally, we get use of real data allocation results from a Tunisian IT operator to show the efficiency and applicability of this ML-based orchestration.