Cross-layer Energy Optimization for Dynamic Video Streaming Over Wi-Fi

Dynamic Adaptive Streaming over HTTP (DASH) constitutes a large fraction of traffic in the mobile Internet. Mobile devices often use video delivery over Wi-Fi, which is a significant energy drain. Dynamic Power Saving Mode (DPSM) is one of the most widely-used approaches for Wi-Fi devices to save power and shorten packet delay. However, DPSM uses a fixed timeout before a device goes to sleep, leading to excessive energy use in many cases. In this paper, we propose a crosslayer lightweight dynamic timeout adjustment algorithm. The application layer measures the Round-Trip Time (RTT) of video segments in the current timeout setup, determines an updated timeout, sets it to the medium access control layer and requests the next video quality based on the RTT and RTT change. We evaluate our algorithm via simulations in OMNeT++ and show that compared with the legacy Wi-Fi DPSM for DASH, our algorithm reduces the average power consumption in the radio front end while significantly improving the fairness among users. Moreover, the new algorithm retains video quality while achieving better energy efficiency. The improvement increases in significance with increasing number of users, which means that the new algorithm works well in dense scenarios. Keywords—cross-layer, energy optimization, round-trip time, dynamic adaptive streaming via HTTP, dynamic power saving mode