How Much Benefit Can Dynamic Frequency Scaling Bring to WiFi?
Zhimin Wang, Qinglin Zhao, Li Feng, Fangxin Xu · IEEE Transactions on Mobile Computing · 2020
Dynamic frequency scaling (DFS) is a state-of-the-art power-saving technique. Various DFS-based WiFi schemes have been proposed for power saving. These schemes demonstrated the power-saving feasibility of DFS via hardware implementation or simulation. This paper is the first that proposes a general theoretical framework to evaluate the performance of these schemes, where we use Queuing theory to analyze the system throughput and use Semi-Markov theory to quantify the power consumption. In addition, we adopt the energy efficiency (i.e., the throughput per energy cost) to compare the gains of these schemes. This efficiency measure can be used to make a trade-off between system throughput and power consumption and therefore help us choose appropriate parameter settings. Extensive simulations verify that our theoretical model is very accurate and our theoretical results well match with universal software radio peripheral (USRP) experiment results. Our study shows that DFS can greatly improve the energy efficiency of WiFi networks even under low SNR conditions; for example, when SNR = 9.7 dB (i.e., the basic requirement for decoding packets in WiFi), the improvement is around 25 percent for 802.11b at rate 11 Mb/s and 16 percent for 802.11 ac at rate 1300 Mb/s. Our study also shows that DFS can be well integrated with other commonly used power-saving mechanisms to improve energy efficiency further.