Traffic-aware Advanced Sleep Modes management in 5G networks

Abstract

Advanced Sleep Modes (ASMs) are defined as a progressive shutdown of the Base Station (BS) depending on the activation and the deactivation times of the different components. This transition duration defines different levels of sleep modes that can be implemented in future 5G networks. We propose in this paper a management strategy based on Q-learning approach which will enable to find the best combination and durations of ASM levels depending on the traffic load and the network operator's policy regarding energy reduction versus latency. Our results show that even in delay-sensitive scenarios, high energy gains can be achieved in low and moderate traffic loads, respectively 55% and 10%, without inducing an extra latency. Starting from a certain traffic load (approximately 30%), ASM should not be implemented in case of stringent latency constraint.

Q-learning based Radio Resource Adaptation for Improved Energy Performance of 5G Base Stations

“…The works [2], [7]- [12] focused specifically on SM management under varying loads. In [8]- [11], the authors proposed a RL based algorithm to manage and determine the appropriate SM depending on the UE requests arrival rate. Their results show that up to 90% ES can be obtained at low loads at the expense of slight degradation in UE throughput and latency.…”

“…P total = P baseband + P P A + P syn + M i P bs + P f ixed (8) where in ( 6), ( 7) and ( 8), R c is the average UE rate in the cell, A is the sum of encoding and decoding PC of the BS, L bs is the BS's computational efficiency measured in (Gflops/W), T c is the channel coherence time in symbols, is a fixed constant which is PA dependent, P P A is the PC of the PAs, p is the load experienced by the PA, η is the efficiency of the PA, P max,P A is the maximum output power of the PA, P f ixed is the baseline or no-load PC of the BS, P syn is the local oscillator PC, P bs is the BS's circuit power, and P total is the BS's total PC. The values for these constants have been obtained from [1].…”

“…The values are given in Table I. While previous studies [8]- [11] have mainly considered management of ASMs, in this work, we take an unconventional approach and focus on adapting the configuration of the BS in accordance to the input traffic demand. In other words, we consider SMs to be a natural consequence of two main factors:…”

A Survey on 5G Radio Access Network Energy Efficiency: Massive MIMO, Lean Carrier Design, Sleep Modes, and Machine Learning

“…• Finally, in sleep mode 4, which has a minimum duration of 1 s, only the wake-up functionalities are maintained [137].…”

“…When the cell load rises and UE traffic appears, such UE traffic is buffered, and the BS has to immediately switch on its functionalities, and serve the required data to satisfy the end-user QoS requirements. However, it should be noted that, since hardware activation and deactivation times are not negligible, and their lengths increase with the number of involved hardware components [137], ASM may increase the UE perceived latency, and this can accordingly affects the UE throughput [140] [141]. Therefore, there is the need for optimising the path between different sleep modes, and proactively activate the BS components in order to limit performance losses.…”