IEEE Circuits and Systems Magazine - Q1 2023 - 58

rates close to 10% per year. This article reviews the latest findings
concerning energy consumption of online video from the
system engineer's perspective, where the system engineer is
the designer and operator of a typical online video service. We
discuss all relevant energy sinks, highlight dependencies with
quality-of-service variables as well as video properties, review
energy consumption models for different devices from the literature,
and aggregate these existing models into a global model
for the overall energy consumption of a generic online video
service. Analyzing this model and its implications, we find that
end-user devices and video encoding have the largest potential
for energy savings. Finally, we provide an overview of recent
advances in energy efficiency improvement for video streaming
and propose future research directions for energy-efficient video
streaming services.
I. Introduction
W
ith the advent of portable devices and
powerful video compression techniques,
on-demand video streaming and communication
services have become an integral part of the
daily lives of billions of users all over the world within
the last decade [1]. Moreover, due to substantial
advances in digital data transmission technology, the
demand for classical video services such as analog
television broadcast is declining [2], as the bandwidth
of networks is sufficiently high to send videos
separately to individual users, who are getting used
to being able to watch the content on-demand anytime
and anywhere.
As a downside to this development, it was found
that the energy consumption, which is directly related
to the production of greenhouse gases (GHG)
causing climate change [3], has simultaneously increased
dramatically. A recent study from the year
2019 claims that for the year 2017, more than 1% of
the global GHG emissions can be attributed to online
video services [3], which was equal to a quarter of
the emissions caused by global aviation. These emissions
can be attributed to the production of related
devices, which makes up 45% of the online videorelated
emissions, and the actual use of online video
services, which comprises the remaining 55% and
which we focus on in this work. According to [3], the
demand for online video services, including the overall
data rate required for video delivery, will further
increase over the next years, which is consistent with
the findings of [4], [5], [6].
Already for the year 2019, a substantial and strong
growth in transmitted video data of more than 8% was
expected [3]. With the appearance of the Corona virus,
the demand for online communications grew unexpectedly
[7] such that, in March 2020 alone, the Internet
traffic at Germany's biggest Internet exchange
point (IXP) increased by 10%. At the same time, traffic
related to online video conferencing increased by 50%
[8], showing that the actual growth rate of online video
services will likely be much higher than previously
predicted.
In the literature, one can find a plethora of studies
investigating the overall impact of online video technologies
as well as other information and communication
technologies (ICT) on the global energy consumption
and GHG emissions [1], [3], [9], [10], [11], [12]. Furthermore,
some studies attempt to break down the overall
energy consumption to the end-user level, where the
target is to provide end users with information on the
GHG emissions caused by the streaming of a single video
[1], [3], [13], [14], [15]. In contrast, in this work, we
target the overall energy consumption related to online
video services, where we take the viewpoint of a system
engineer constructing and maintaining a certain online
video service.
As a consequence, we provide a generalized overview
on the energy consumption of all devices and systems
that perform online video tasks, which include
data centers storing and providing videos, data transmission
networks comprised of network nodes, and
end-user devices such as TVs, smartphones or tablet
PCs [1], [15]. We intend to cover all major applications
such as (on-demand) streaming, video conferencing,
social networks, broadcasting, and surveillance. To
this end, we map all applications to a general system
model that comprises the most common online video
services. This model is depicted in Fig. 1. The system
model contains three main components: The first
component comprises terminals, which are end-user
devices operated by users who request video streams
[1], [16]. The second component includes all devices
enabling the network connection for video data transmission.
The networks transport the data using different
kinds of data carriers such as Wi-Fi, 4G, and 5G
for wireless transmission as well as copper or optical
fiber for wired transmission. Typical devices in the
transmission network are IXPs, switches, and routers.
Finally, the third component mainly consists of data
centers (DCs), where video data is transcoded, stored,
and sent to the end users [17], [18].
Christian Herglotz, Matthias Kränzler, and André Kaup are with the Chair of Multimedia Communications and Signal Processing, Friedrich-Alexander
University Erlangen-Nürnberg (FAU), 91054 Erlangen, Germany (e-mail: christian.herglotz@fau.de; matthias.kraenzler@fau.de; andre.kaup@fau.de).
Robert Schober is with the Institute for Digital Communications, Friedrich-Alexander University Erlangen-Nürnberg (FAU), 91054 Erlangen, Germany
(e-mail: robert.schober@fau.de).
58
IEEE CIRCUITS AND SYSTEMS MAGAZINE
FIRST QUARTER 2023
IEEE Circuits and Systems Magazine - Q1 2023

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