IBC2022: This Technical Paper presents some of the findings from the analysis related to energy consumption in the contnet of video content viewership from a smartphone perspective.

 Abstract

With consumer IP video traffic representing 84% of all consumer IP traffic in 2021 (up from 79% in 2016), the notion of Corporate Social Responsibility (CSR) is garnering increasing attention from video service providers. The energy consumption of their service deployments needs to be, at the very least, mastered and minimised. At best, it has to achieve the carbon footprint neutrality already publicly committed to by some of the industry’s key players. 

This paper will present some of the work conducted as part of the New vidEo STandards for Enhanced Delivery (NESTED) collaborative project with regards to video distribution over 5G networks. A key focus of the paper will be on the device-side contribution of next-generation codecs compared with legacy ones in terms of energy consumption. The paper will also provide a comparison of the simulations performed on unicast and multicast 5G network architectures.

Introduction

Over the past few years video traffic has grown steadily. What was limited only to the consumption of live and video-on-demand (VOD) services on a managed network from a set-top-box at home10 years ago has turned into access to the same services from everywhere and on any type of terminal with the advent of over-the-top (OTT) video streaming and adaptive bit rate (ABR) technology. In this context, where users see their experience continuously enriched by new functionalities such as targeted advertising, multiview, watchparty, orvirtualreality-enhanced/360-degree video, it seems obvious that the share of video traffic as a part of over al linternet usage will continue increasing. As a consequence, the question of optimising the carbon footprint of the end-to-end technical chain enabling such advanced experiences arises. New generations of codecs are emerging, and new broadcast technologies are being implemented in order to cope with this demand of richer video experiences and the related increase in video traffic. But, what is the environmental cost of this unending race for innovation?

In this paper, we present some of the findings from our analysis related to energy consumption in the context of video content viewership from a smartphone (endpoint-side) perspective. Such findings are nurtured by the research conducted in the context of the NESTED collaborative project with the support of France’s Brittany region in relation to video distribution over 5G networks based on state-of-the-art standards such as Common Media Application Format (CMAF) and Dynamic Adaptive Streaming over HTTP (DASH).

Figure 1 provides a simplified view of the various technology layers involved in video content streaming. This paper focuses primarily on the end pointside of the video content, network/transport and physical/radiolayers, which proves critical in the rendering of the end user’s perceived quality of experience (QoE). It notably leverages some of the empirical (when possible) and theoretical measurements performed around Viaccess-Orca’s (media player) and Enensys Technologies’ (multicast gateway) end point components. The results shared in the paper should be considered as work-in-progress and will be further enriched in the context of the two-year long NESTED project but will in any case constitute a partial view of a wider and more complex subject (especially when adding the head-end part into the equation).

With regards to the physical/radio and network/transport layers, in the absence of empirical measurements exploitable at time this paper is written, we present the results of a theoretical study related to the impact of 5G broadcast network solutions on a smartphone’s energy consumption. The second part details the results of empirical measurements centered on the High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC) codecs and the associated profiles (video content layer). Finally, we try to apply these results (and draw some—partial—conclusions) to a concrete casestudy: users concentrated in a stadium watching the highlights of the event they are attending.

Download the paper below

Downloads