LESS: A Lightweight Energy Saving System for Cellular Networks

Operational expenditure related to energy consumption of cellular base stations is a primary concern for network operators, as densely deployed next generation base stations will need to be 2000 times more energy efficient in the coming decades. This trend is driving interest in autonomic energy saving features that can switch off under-utilized base stations that are within the coverage region of other base stations. In line with this goal, we propose a Lightweight Energy Saving System (LESS) that is highly scalable and can be used to toggle energy savings features autonomously for operators via a data driven approach. As a use case, we consider a switching off mechanism for pairs of cells in a coverage-capacity relationship,as such a mechanism is currently configured manually by operators.We evaluate LESS using a real coverage-capacity utilization dataset with 80 representative base station pairs (i.e., 160 base stations),and show that up to 27% energy savings can be achieved, but that the savings is highly dependent on the levels of utilization of pairs.Moreover, using our LESS system results in less than a 3% gap compared to an ideal energy saving strategy that has perfect future knowledge of the utilization of the coverage-capacity pair, and outperforms several state-of-art forecasting methods.