Max-min Fair Resource Allocation in SD-RAN

Abstract

Software-Defined Radio Access Networks (SD-RANs), introduced for the first time in 5G networks, symbolize a paradigm shift in the way the allocation of cellular network resources is performed. The main feature of SD-RAN is the possibility to decouple the control plane from the data plane, and associating the former with a controller that is located away from the Base Stations (BSs). This enables an increased flexibility in the allocation of network resources, resulting in considerable performance improvements compared to the classical pre-5G resource allocation in cellular networks. However, it is not yet clear to what extent this enhancement ranges in terms of different metrics of interest and optimization objectives. One such objective is to allocate resources so that the minimum value of the data rate in the network is maximized, i.e., to provide max-min fairness. Therefore, in this paper, we consider analytically the problem of max-min fairness in an SD-RAN environment by deriving the policy which accomplishes that along with the achievable performance, of interest to users and cellular operators. We do this for two scenarios; one in which we provide max-min fairness across all users in the network, and the other in which the goal is to provide max-min fairness in throughput across BSs, and within users of each BS. We evaluate the performance with input parameters from real data sets. Results show that the introduction of SD-RAN improves the minimum rate by up to 4× compared to the case with no SD-RAN controller.