SO-RAN: Dynamic RAN Slicing via Joint Functional Splitting and MEC Placement

Abstract

Next-Generation (NG) mobile networks are expected to revolutionize wireless communication systems, facilitating the deployment of critical applications with stringent Quality of Service (QoS) requirements. Multi-access Edge Computing (MEC) and Radio Access Network (RAN) slicing are considered vital mechanisms of upcoming NG systems as they allow creation of end-to-end isolated RAN slices and increase of serving mobile data traffic on edge. The efficient deployment of RAN functions promises flexible and disaggregated architectures for the diversified needs of NG networks. In this work, we propose a novel optimization approach that minimizes RAN/MEC economic costs and maximizes served traffic. This framework creates isolated RAN slices by optimal placement of RAN and MEC functions per slice, which solves the problem of operating cost-efficient edge networks and increases the served traffic with diverse QoS requirements. We first formulate our framework as a convex and linear problem and then decompose it using a distributed method that accelerates the performance and guarantees the optimal global solution. Our results show that our proposed work better uses the spectrum, and there is a trade-off between the achieved throughput and the cost incurred to the network.