Towards Energy-Aware Video Streaming in 5G/B5G - Application Perspective

This paper presents an evaluation study on how the characteristics and amount of video traffic affect the energy consumption in different video applications utilising GPU and CPU processing, and in selected components of the 5 G mobile network. The focus is put on the widely used HTTP adaptive streaming, which allows a flexible way for delivering dynamic content over the Internet for a variety of mobile devices with unique characteristics. However, these devices tend to maximise the processing performance with the highest available video quality, which leads to the penalty of higher energy consumption. In this paper, we show how different video quality affects the energy consumption in different parts of the components over 5 G RAN, and evaluate selected video applications on energy usage basis. For this purpose, we use an accurate and system wide mobile test network for the evaluation with advanced energy measurement infrastructure. The extensive set of results with different video player applications verify that the use of smaller bitrates with decent QoS and QoE levels can reduce end-to-end energy consumption towards green networking.