An Improved BBU/RRU Energy Consumption Predictor for 4G and Legacy Mobile Networks using Mixed Statistical Models

Nowadays, the mobile traffic growth is being driven by both the rising number of smartphone subscriptions and an increasing average data volume per subscription, fueled primarily by the video content access. Consequently, the base stations' energy consumption is also growing at a very fast rate. Thus, the mobile operators have to leverage between pursuing high capacity, spectral efficiency and the energy efficient design of their networks. To accomplish this issue, it is demanding for the operators to find mechanisms to accurately monitor their Radio Access Network (RAN) energy costs, by measuring the specific power consumption in each hardware equipment. In order to reduce the costs, this detailed knowledge allows to analyse the future hardware and software upgrades, network refarming for Fifth Generation (5G) or legacy technologies switch-off. The aim of this paper is to develop a multi-technology energy consumption model for mobile networks. An approach based on linear mixed effects model was used, considering Performance Management (PM) and Configuration Management (CM) data. The outcome is the predicted radio equipment power consumption, including radio and base band, detailed over the time. The model was developed and validated using real energy measurements extracted from monitoring equipment, installed on different base stations.