Method of 5G TDD Midhaul Multiplexing Gain Estimation based on System-Level Traffic Measurements

Abstract

Cloud Radio Access Network (Cloud RAN) was introduced to reduce network cost and increase system flexibility. Due to the split of baseband functions between Distributed (DU) and Centralized Units (CU) the F1 interface has been defined, creating a new domain for transport access networks - midhaul. In this paper, we estimate the statistical multiplexing gain (MG) of traffic aggregation from different DUs on midhaul link. Results of the evaluation can be deployed in Cloud RAN dimensioning system used in the network planning process. Midhaul transport links must provide sufficient capacity and Quality of Service (QoS) to enable required radio performance. To understand the QoS requirement (temporal values of throughputs) of the radio interface and traffic profile patterns, the system level simulator has been used during the study. To enable scaling of simulation results (from 21 to 200 or more cells network) or use other traffic measurements, we have defined a method that is based on a bootstrap methodology. Results show that an optimal point between D U and CU to place an aggregation point is where it could aggregate traffic from 20–40 cells. This method enables reduction of the computational time from several days to several seconds, which is significant for network dimensioning recommendation and in turn for efficiency and elasticity of the service delivered to the telecommunication operators.