Towards Energy-Efficient Adaptive Mpeg-Dash Streaming Using Hevc

Abstract

Green networking and communications have become important factors in nowadays energy-hungry video transmission, where users expect to gain high-quality video regardless of time and place. Modern processors, displays and high-speed mobile networks already enable high-definition video streaming to mobile devices that can decode and playback the stream successfully while still maintaining the device capacity for other tasks. However, user mobility and network congestion can affect even the best machinery and introduce the need for HTTP adaptive streaming, which has popularized among content providers. Still, its optimal utilization especially in terms of energy-efficiency is slightly investigated issue. In this paper, we tackle this issue and evaluate both the content production (headend) as well mobile user (terminal) energy consumption against different adaptation methods considering CPU vs. GPU based approaches. The results indicate that frame rate adaptation provides the lowest consumption in both ends and usage of GPUs can lead to further savings.