Optimizing the Power Consumption of Mobile Networks Based on Traffic Prediction

Abstract

Nowadays, mobile networks approach a steady growth in traffic demand. As a result, mobile network providers continuously expand their network infrastructure mainly by installing more base stations. Currently, there is a huge number of base stations serving mobile users all over the world, and this number is expected to double in the coming few years, which leads to a larger wastage of energy during low demand times. Exploiting the possibility of turning off base stations at low demand times is one of the promising approaches for saving energy and reducing CO2 emissions. Here, an early and accurate estimation of the traffic is crucial for managing resources proactively. Therefore, in this paper, we introduce a Power Management System that applies a global provisioning policy to base stations for enabling network reconfigurations in terms of power efficiency. This system is based on a Hybrid Traffic Prediction Model that forecasts the workload of base stations by utilizing historic traffic traces. A simulator is implemented to evaluate the proposed management system, which is fed with real data provided by the Open Mobile Network project. The experimental results show the possibility of turning off 49% of the base stations at some times of the day without degrading the QoS.