Machine-Learning based Traffic Forecasting for Resource Management in C-RAN

The assumption of a fixed computational capacity at the Baseband Unit (BBU) pools in a Cloud Radio Access Network (C-RAN) deployment results in underutilized resources or unsatisfied users depending on traffic requirements. In this paper a new strategy to predict the required resources based on Machine Learning techniques is proposed and analysed. Support Vector Machine (SVM), Time-Delay Neural Network (TDNN), and Long Short-Term Memory (LSTM) have been tested and compared to select the best predicting approach. Instead of using a regular synthetic scenario a realistic dense cell deployment over Vienna city is used to validate the results. Authors show that the proposed solution reduces the unused resources average by 96 %.