Energy Consumption Estimation of Mobile Networks' Base Stations Due to Traffic Change

Mobile operators are facing strong challenges regarding the environmental impact of their operations. The International Telecommunications Union (ITU) considered energy consumption as the prime source of Carbon dioxide (CO2) emissions for the whole Information and Communications Technology (ICT) sectors. ITU encouraged all ICT sectors to commit their efforts in order to control their energy consumption and consider renewable energy sources. By the year 2050, CO2 emissions should be fully neutralized as per the Paris climate agreement. The energy consumption of the Radio Access Network (RAN) represents almost 80% of the total mobile network energy consumption. RAN mainly consists of a large number of distributed Base Stations (BSs). There are some factors that affect the total amount of energy consumption of those BSs such as adding more Base Stations (BSs), upgrading existing ones, and the amount of carried traffic. The number of BSs is increasing rapidly to satisfy the growing demand for radio capacity and coverage. Moreover, upgrading BSs to the next generations like 5G is causing a considerable shift in energy consumption for existing BSs. However, the impact of the incremental traffic on BSs energy consumption is difficult to estimate or even measure, unlike the other factors. The objective of this study is to build a model that can estimate the amount of BSs energy consumption due to changes in traffic. This model will help mobile operators to predict the expected BSs energy consumption considering their forecasted traffic growth on existing mobile networks. Both voice and data traffic changes were considered in this model. Historical data of a large sample of 1,011 BSs were used to generate a linear model describing the relationship between energy consumption and both data & voice traffic using multiple regression techniques. This model was then adapted to predict energy consumption on a network scale by considering only the forecasted change in traffic along with the latest historical actual energy data. This adaptation was validated over a cluster of 1,600 BSs and showed maximum absolute error of 3%. Moreover, the inclusion of operators' historical energy data in the adapted model would make it more universal for use by different mobile operators with different BSs configurations. © 2023 IEEE.