IEEE Circuits and Systems Magazine - Q1 2023 - 59

Figure 1. Active components in online video applications comprise data centers and the infrastructure of the online video
providers (VP), transmission networks (NW), and end-user terminals (UT). The overall energy consumption of all components
is denoted by the variable E. The video providers store video data and respond to user requests using servers in data centers.
The transmission networks include different transmission technologies using components such as Internet exchange points
(IXPs), base stations, switches, and (home) routers. The interconnecting lines represent wireless (dashed lines) and wired
(solid lines) connections. The end-user terminals comprise playback and recording devices such as smartphones, tablet
PCs, desktop PCs, TV sets, and cameras. All components contribute to the overall energy consumption caused by online
video applications.
In this respect, we adopt the system engineer's perspective
of an online video service. This means that we
consider the complete infrastructure, i.e., all relevant
devices mentioned above. The proposed system model
allows to estimate the overall power and (yearly) energy
consumption caused by a single online video service,
e.g., a provider of video conferencing systems or
an on-demand streaming platform. We note that since
we consider a large, distributed system consisting of
thousands of independent devices located all over the
world, validation measurements are extremely complex
and costly. Hence, in this initial work, we rely on models
for subsystems which have already been validated in
the literature, such that the obtained energy consumption
estimates are sufficiently accurate.
The proposed global model enables us to reveal the
main reasons for the high energy consumption of online
video services, to point out new research directions,
and to reveal opportunities for improving the energy
efficiency in the future. As such, the proposed global
energy consumption model can be used as a baseline
for future research on online video services, where
FIRST QUARTER 2023
researchers can determine the benefits of their work
with respect to the global energy consumption.
The remainder of this article is organized as follows.
In Section II, a thorough literature review will show the
link between GHG emissions and the energy consumption
of the devices required for online video service
provisioning. Furthermore, we give an overview of the
current research on the energy efficiency of data centers,
transmission networks, and end-user devices. Afterward,
Section III dives into a detailed system-level
analysis to construct an energy consumption model
based on the general system model shown in Fig. 1. The
resulting model describes and covers all major energy
sinks that are present in video streaming. Furthermore,
parameter values will be derived from the literature facilitating
concrete energy estimates. Finally, Section IV
considers use cases for the proposed model, where we
identify the most energy-intensive components of different
online video services. Furthermore, we shortly describe
state-of-the-art techniques and approaches to improve
the energy efficiency of online video technology
and present upcoming trends as well as potential future
IEEE CIRCUITS AND SYSTEMS MAGAZINE
59

IEEE Circuits and Systems Magazine - Q1 2023

Table of Contents for the Digital Edition of IEEE Circuits and Systems Magazine - Q1 2023