Advanced Sleep Modes to comply with delay constraints in energy efficient 5G networks

The staggering growth of mobile traffic fostered by the extensive spreading of 5G technology and massive Internet of Things (IoT) applications is leading to network densification, entailing a boost in network power consumption, with consequent higher operational cost for Mobile Network Operators (MNOs) and raising sustainability issues. To reduce energy consumption when the traffic is low, new BSs feature Advanced Sleep Modes (ASM) that allow to reduce the network energy consumption by gradually deactivating the BSs into progressively deeper sleep modes with lower power consumption. However, the deep sleep modes cause high reactivation delays that may jeopardize the Quality of Service.In this paper, focusing on the periods in which traffic is very low, we extensively investigate the potentiality of ASMs based operation in terms of the trade-off between energy saving and delay under different 5G scenarios and traffic loads. By observing that optimal configuration settings depend on the scenario and on the load, we design a framework based on a stochastic model to perform dynamic tuning of the configuration settings that adapts in real time the parameters to the actual traffic load and scenario.