Study on CPU and RAM Resource Consumption of Mobile Devices

using Streaming Services

Przemyslaw Falkowski-Gilski

and Michal Wozniak

Faculty of Electronics, Telecommunications and Informatics, Gdansk University of Technology,

Narutowicza 11/12, Gdansk, Poland

Keywords: Mobile Devices, Multimedia Content, Streaming Services, Wireless Communication.

Abstract: Streaming multimedia services have become very popular in recent years, due to the development of wireless

networks. With the growing number of mobile devices worldwide, service providers offer dedicated

applications that allow to deliver on-demand audio and video content anytime and everywhere. The aim of

this study was to compare different streaming services and investigate their impact on the CPU and RAM

resources, with respect to type of Internet connection. The paper consists of two parts: theoretical and research.

The first part provides a description of current means of wireless communication, including transmission of

multimedia in Wi-Fi and cellular systems, as well as principles of operation of popular streaming media

available on the marked, including utilized coding algorithm and available bitrates. The second part describes

the set of utilized consumer devices, including 50 smartphones, as well as tools, laboratory equipment,

and research scenarios. Results of this study may aid both researchers and professionals involved in the

digital mobile market, including content and service providers, as well as network operators.

1 INTRODUCTION

The continuous development of mobile devices and

wireless networks has contributed to the creation of

many services enabling streaming of audio and video

content. Unlike traditional terrestrial radio or

television, they allow each and every individual to

choose the content he or she desires at a given

moment, without being limited by a fixed

broadcasting schedule (Kohli, 2020).

Streaming services allow us to consume content

without having to download the entire file into the

memory of a consumer device in order to play it back.

In this case data are being downloaded from the

server continuously in real-time. This is possible

thanks to the increase in both throughput and network

capacity (Muscat, 2019).

Like all data transmitted over the Internet,

multimedia are divided into packets that are sent to

the recipient, sometimes even via different routes

and/or different access media. Even when a drop in

quality of a connection occurs, smooth playback is

maintained thanks to the existence of the so-called

buffer. The buffer may be viewed as a queue that

https://orcid.org/0000-0001-8920-6969

allows to download and process data in advance

(Bouraqia, Sabir, Sadik and Ladid, 2020).

2 WIRELESS COMMUNICATION

INTERFACES

Wireless networks enable to connect and share

resources among multiple consumer devices located

and operating on a predefined serving area

(Kryvinska and Greguš, 2019). Currently, the most

popular ones include Wi-Fi and cellular systems.

2.1 Wireless-Fidelity

Wireless-Fidelity is a proprietary name of the IEEE

802.11 family of standards, in which WLANs

(Wireless Local Area Networks) are based.

These networks are susceptible to interference, due to

utilized frequencies. In order to minimize this effect,

the allocated frequency range of 2.4 GHz has been

divided into 14 channels in Europe and 13 in the

USA, of 22 MHz width each. Whereas, the 5 GHz

Falkowski-Gilski, P. and Wozniak, M.

Study on CPU and RAM Resource Consumption of Mobile Devices using Streaming Services.

DOI: 10.5220/0010624900003058

In Proceedings of the 17th International Conference on Web Information Systems and Technologies (WEBIST 2021), pages 235-241

ISBN: 978-989-758-536-4; ISSN: 2184-3252

235

band has been divided into 23 channels, each of

20 MHz width, respectively.

Over the years, many versions of the IEEE 802.11

standard have been introduced, subsequent ones were

market with letters of the alphabet. Each new version

brought improvements in the speed and range of data

transmission.

The original version, introduced in 1997, allowed

data transmission with a maximum speed of up to

2 Mbps in the 2.4 GHz frequency band. Its range

(coverage) was equal to approx. 20 m indoors and

approx. 100 m outdoors.

In 1999, two subsequent versions were issued,

labelled as IEEE 802.11a and 802.11b. They both

utilized SISO (Single-Input Single-Output)

technology, however differed in operation

frequencies and modulation techniques.

 802.11a utilized 5 GHz and 3.7 GHz radio

frequencies with channel width of 20 MHz,

enabling data transmission with speeds up to

54 Mbps. The connection area ranged from

35 to 120 m for the 5 GHz variant, whereas for the

3.7 GHz variant it was equal to even 5 km

outdoors. The utilized modulation scheme was

OFDM (Orthogonal Frequency Division

Multiplexing);

 802.11b utilized the 2.4 GHz frequency range

with channel width of 22 MHz.

The transmission speed was up to 11 Mbps,

and its serving area for both indoor and outdoor

environments ranged from 35 to 140 m. In this

case, DSSS (Direct Sequence Speed Spectrum)

technique was used.

In 2003, the IEEE 802.11g standard was released,

enabling transmission of data at speeds up to 54 Mbps

with the 2.4 GHz frequency range. The serving area

ranged from 38 to 140 m with SISO technology.

It was compatible with both OFDM and DSSS

techniques, depending on the user’s choice.

The IEEE 802.11n standard, introduced in 2009,

operated at either 2.4 or 5 GHz, and supported a

maximum bandwidth of 150 up to 600 Mbps.

The serving area ranged up to 75 m indoors and

250 m outdoors. This increase in maximum transfer

speeds and range resulted from the introduction of

the MIMO (Multiple-Input Multiple-Output) antenna

array for both sending and receiving data.

Additionally, this version enabled to increase the

channel bandwidth from 20 to 40 MHz.

In 2013, the IEEE 802.11ac standard, operating

within the 5 GHz radio band, enabled to select the

width of the channel from 20, 40, 80, up to 160 MHz.

It introduced MU (Multi-User) MIMO technology.

This standard also brought a novel 256-QAM

modulation scheme. Therefore, the maximum

transfer speed ranged from 450 Mbps up to 1.3 Gbps,

and the connection range indoors was equal to 35 m

(Gast, 2005).

In 2018, the Wi-Fi Alliance concluded that the

names of IEEE 802.11 standards should be more user-

friendly. For this reason, a new nomenclature of

standards has been introduced, as shown in Table 1,

together with new logos, particularly designed and

implemented for mobile devices (see Figure 1).

Table 1: Nomenclature of IEEE 802.11 standards.

Before 2018 After 2018

802.11b Wi-Fi 1

802.11a Wi-Fi 2

802.11h Wi-Fi 3

802.11n Wi-Fi 4

802.11ac Wi-Fi 5

802.11ax Wi-Fi 6

Figure 1: Updated Wi-Fi logos introduced by Wi-Fi

Alliance.

A comparison of currently available IEEE 802.11

standards, including utilized frequency range as well

as maximum throughput, is shown in Table 2.

Table 2: Comparison of different IEEE 802.11 standards.

Standard

Release

date

Freq. range

[GHz]

Throughput

[Mbps]

802.11 1997 2.4 2

802.11b 1999 2.4 11

802.11a 1999 5 54

802.11g 2003 2.4 54

802.11n 2009 2.4/5 150-600

802.11ac 2013 5 400-1300

802.11ax 2019 2.4/5 1200-14000

However, Wi-Fi networks operate in the so-called

ISM (Industrial, Scientific, Medical) band, which is

open for other communication standards, including

Bluetooth, etc. This issue, related with coexistence of

multiple systems and/or standards, is an important

topic, especially when talking about the IoT (Internet

of Things) concept (Polak and Milos, 2020).

This fact, quite the opposite to standardized cellular

networks, may be a crucial differentiator when it

comes to possible interferences.

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

236

2.2 Cellular Networks

The standardization of cellular networks begun in the

1970s and 1980s. However, the first generation (1G)

standard, offering analog radio transmission,

was focused only on speech and text services.

The second generation (2G) system offered some

type of multimedia transmission, namely MMS

(Multimedia Messaging Service) with still pictures

and audio. The breakpoint came with the third

generation (3G), as the expectations of users started

to grow.

The UMTS (Universal Mobile

Telecommunications System) offered speeds up to

384 kbps, with video calls (aside from traditional

voice calls), file sharing, Internet browsing and other

multimedia services available so far only using fixed

cable connections. The next step was the introduction

of HSDPA (High-Speed Downlink Packet Access)

and HSUPA (High-Speed Uplink Packet Access)

protocols, which complemented each other creating

the HSPA (High-Speed Packet Access). It provided

transfers from 1.8 to 3.6 Mbps in the downlink and

1.4 Mbps in the uplink.

In 2014, LTE (Long-Term Evolution) as the

fourth generation (4G) system was introduced.

This standard increased the peak data rates up to

100 Mbps for downlink and 50 Mbps for uplink,

with significant delay reduction and improved

spectral efficiency related with flexible frequency

allocation. LTE allows 6 different channel

bandwidths, namely 1.4, 3, 5, 10, 15, and 20 MHz.

Theoretically, with a 20 MHz wide channel and

4x4 MIMO antenna equipment, it allows speeds up to

326 Mbps for downloading and 86 Mbps for

uploading data. With further improvements, referred

to as LTE-Advanced, related with the growing

number of active network subscribers, throughput can

be increased even up to 3 Gbps and 1.5 Gbps,

respectively (Meraj and Kumar, 2015; Shen, Lin

and Zhang, 2020).

Currently, each and every network operator is

focused on implementing the fifth generation (5G)

network infrastructure. As the number of active users

and their consumer devices continues to grow,

throughput may be further extended to 10 or even

20 Gbps (Raca, Leahy, Sreenan and Quinlan, 2020).

Yet still, most people own and use 4G-compatible

mobile devices. That is why this cellular standard,

along with Wi-Fi connectivity, was evaluated.

3 MOBILE MULTIMEDIA

DISTRIBUTION

The popularity of multimedia content distribution via

the Internet started in the last two decades (Iwacz,

Jajszczyk and Zajaczkowski, 2008). With the

growing demands for transferring large amounts of

data in a timely manner, the IETF (Internet

Engineering Task Force) has developed the RTP

(Real-Time Transport Protocol).

The RTP standard is dedicated to handle

streaming of multimedia over IP (Internet Protocol)

networks that enable to deliver audio and video

packets with low overhead. It manages the streaming

session between the server and clients with the RTCP

(Real-Time Control Protocol). However, RTP has

several disadvantages, such as: blocking packets by

firewalls, no support for currently operating CDN

(Content-Distribution Networks), difficulties when

handling different receiving devices (e.g. processing

power, resolution, etc.).

In order to overcome this, HTTP (Hypertext

Transfer Protocol) was introduced. Unlike RTP,

HTTP is compatible with CDNs and is not blocked by

firewalls. Additionally, in HTTP the client is

responsible for managing the streaming session,

which eliminates the burden on the server. However,

despite many advantages, HTTP cannot handle

streaming different bandwidths for clients using

diverse consumer devices. Therefore, HAS (HTTP

Adaptive Streaming) was proposed.

HAS allows to adjust the quality of multimedia to

the available network resources and technical

parameters of the receiving device. This is possible

by dividing multimedia files into short segments,

which are then encoded at different data rates.

Multimedia transmitted in such a way may contain

both video and/or audio content, as well as subtitles

in various languages.

The coded segments are available on the web

server so that the client can download them on

demand. Before starting the essential playback,

the client downloads a MPD (Media Presentation

Description) file, containing information about the

streamed content, in the form of an XML (Extensible

Markup Language). It contains information such as:

start and end time of each segment, available

transmission rates, URL (Uniform Resource Locator)

for each segment.

Based on a set of parameters, including Internet

connection, screen resolution of the consumer

device, etc., a schedule for downloading subsequent

segments is prepared. The schedule may be

dynamically changed, based on network quality

Study on CPU and RAM Resource Consumption of Mobile Devices using Streaming Services

237

parameters, in order to provide the highest quality

possible while maintaining smooth playback.

Currently, the most widely-known and utilized

standard is MPEG (Moving Pictures Expert Group)

DASH (Dynamic Adaptive Streaming over HTTP),

utilized by a variety of streaming services, including

Netflix and YouTube (Vetro, 2011; Gazdar and

Alkwai, 2018; Hoßfeld et al., 2015).

4 MOBILE STREAMING

SERVICES

This chapter discusses popular mobile streaming

services (Falkowski-Gilski and Uhl, 2020;

Falkowski-Gilski, 2020), including utilized codecs

and available bitrates that were evaluated during

this study.

4.1 Spotify

Spotify is a streaming service that allows to play

audio files. It was first launched in 2008. As the first

on the market, it offered both music pieces and

podcasts on multiple mobile platforms. Currently,

its library contains over 60 million songs. Its free

version enables to: access the full library, and

playback (interspersed with advertisements).

Whereas, the premium version enables to: play

content without advertisement, even offline, and with

higher quality (bitrate).

Spotify supports different file formats for content

distribution from creators, including FLAC (Free

Lossless Audio Codec) and WAV. Then, audio files

are encoded using either: Ogg Vorbis (bitrates of

96, 160, 320 kbps), AAC (128, 256 kbps),

or HE-AACv2 (24 kbps). Premium users have the

ability to choose one of the following bitrates:

automatic (depending on the network connection

parameters), low (approx. 24 kbps), normal (approx.

96 kbps), high (approx. 160 kbps), very high (approx.

320 kbps). With a dedicated application, mobile users

can not only search for songs or create their own

playlist, but also listen in a group session mode or

even control playback on another compatible device.

4.2 Tidal

Tidal is a service containing over 55 million songs

and more than 200,000 music videos and movies.

In order to consume content, one needs to purchase

one of the two available subscription versions:

Premium or Hi-Fi.

Premium allows to play audio in standard quality

that is either normal (depending on connection speed)

or high (AAC at 320 kbps), as well as video in HD

quality. Additionally, users can download content and

play it offline. The Hi-Fi version offers playback in

the Hi-Fi format (uncompressed music files at

1411 kbps) and MQA (Master Quality Authenticated)

format (recordings from the studio). It offers similar

capabilities as Spotify, except for remote control and

group sessions.

4.3 Netflix

This platform is focused on audio-video content

consumption, such as movies, series and other

materials. However, the library is strictly dependable

on the region in which the user is located. Content

consumption in SD, HD and 4K formats is only

possible with a subscription.

A dedicated application is available on a variety

of consumer devices. Additionally, users can

download content directly to their device and watch it

while being offline. It offers a variety of user profiles

and related suggestions based on similar and/or

previously watched content, as well as a resume

playback option when the viewing process was

interrupted.

4.4 Twitch

This streaming platform was designed in order to

connect the gaming and broadcasting industry.

Currently, Twitch allows creators not only to upload

and share content with others, but also earn money

from ads and subscriptions. The displayed audio-

video quality is dependable on current network

conditions (auto mode). However, one can chose one

of the following resolutions: 160p, 360p, 480p, 720p,

720p 60 FPS, and 1080p 60 FPS. The dedicated

application enables to select from a variety of

broadcast categories, including type of gameplay,

individual creators. It offers the possibility to watch

saved broadcasts or their fragments, adjust playback

settings or even start a fresh live streaming session.

4.5 YouTube

This platform is available in a free version

(with advertisements) as well as a premium one

(no displayed advertisements). However, both

version allow creators to earn money. The premium

version enables to play audio and/or video with the

screen turned off. The resolution ranges from:

144p, 240p, 360p, 480p, 720p 60 FPS, 1080p 60 FPS,

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

238

even up to 4K. The application offers multiple search

options, including movies, playlists, even channels

broadcasting live.

5 ABOUT THE STUDY

The study was carried out using a set of 50 mobile

devices coming from different manufacturers.

Each terminal was running Android 10 and had a

8-core processor and 4 GB of RAM. The display

resolution was equal to Full-HD. All consumer

electronics were compatible with Wi-Fi 802.11

a/b/g/n/ac, as well as 2G, 3G, and 4G cellular

networks.

The serving network infrastructure was realized

with a typical Wi-Fi access point, with 2x2 MIMO

antenna array, operating in the 2.4 GHz frequency

range. After a preliminary benchmark, the Internet

connection was set to the following throughput

values:

 Download speed: maximum 300 Mbps, typical

225 Mbps, minimum 150 Mbps;

 Upload speed: maximum 40 Mbps, typical

30 Mbps, minimum 20 Mbps.

All data sourced from mobile devices,

for monitoring as well as further processing purposes,

were gathered in a wired manner, in order not to

influence the wireless connectivity, using a custom

Linux-based software. The current status was

refreshed every second.

The streaming services were installed in the

currently available distribution, sourced from the

Android dedicated application market. The research

campaign was composed of a set of scenarios,

including both Wi-Fi and cellular connectivity,

together with audio and mixed audio-video content,

as shown in Table 3.

Table 3: Investigated research scenarios.

Name

Wireless

interface

Type of

conten

Scenario 1 Wi-Fi Audio-Video

Scenario 2 Wi-Fi Audio

Scenario 3 Cellula

Audio-Video

Scenario 4 Cellula

Audio

In each of the four scenarios, we have predefined

the initial throughput value, based on type of

streaming services as well as quality (related bitrate).

The list of settings, for each respective scenario,

are shown in Tables 4-7.

Table 4: Initial parameters for scenario 1.

Approach

no.

Streaming

application

Content

quality

Initial

throughput

[kbps]

1 Netflix 240p 1024

2 Netflix 480p 1024

3 Netflix 1080p 1024

4 YouTube 240p 1024

5 YouTube 480p 1024

6 YouTube 1080p 1024

7 Twitch 240p 1024

8 Twitch 480p 1024

9 Twitch 1080p 1024

In case of scenario 1, the content quality ranged

from 240p up to 1080p, regardless of the type of

utilized streaming application.

Table 5: Initial parameters for scenario 2.

Approach

no.

Streaming

application

Content

quality

Initial

throughput

[kbps]

1 Spotify Low 512

2 Spotify Normal 512

3 Spotify High 512

4 Tidal Normal 512

5 Tidal High 512

For scenario 2, the content quality ranged from

low up to high for Spotify, and from normal to high

for Tidal.

Table 6: Initial parameters for scenario 3.

Approach

no.

Streaming

application

Cellular

networ

Content

quality

1 Netflix 3G 480p

2 Netflix 4G 480p

3 YouTube 3G 480p

4 YouTube 4G 480p

5 Twitch 3G 480p

6 Twitch 4G 480p

Scenario 3 was focused on investigating different

audio-visual content distribution streaming

applications, available in 480p resolution, via 3G or

4G terrestrial radio interfaces.

Table 7: Initial parameters for scenario 4.

Approach

no.

Streaming

application

Cellular

networ

Content

quality

1 Spotify 3G High

2 Spotify 4G High

3 Tidal 3G High

4 Tidal 4G High

Whereas scenario 4 was aimed at investigating

audio content distribution applications, available in

high quality, transmitted via 3G and 4G as well.

Study on CPU and RAM Resource Consumption of Mobile Devices using Streaming Services

239

In each of the four scenarios, we have performed

typical user activities, including: moving forward and

backward, skipping and selecting another material,

selecting and switching to and from a playlist, turning

full screen mode on and off.

6 RESULTS

Results, concerning all the aforementioned scenarios,

user activities, as well as devices, have been

averaged, concerning utilized CPU and RAM

resources, are shown in Table 8.

Obtained data indicate, quite surprisingly that the

quality of the consumed content itself does not affect

the CPU usage. In case of RAM, the situation is quite

the opposite. However, this increase is not linear with

the rise of quality of media. This fact indicates that

although RAM is more affected than CPU, the overall

usage depends on a number of factors.

Additionally, larger deviations were observed

during the 1080p content playback. This surely was

related to data buffering, resulting from a seldom

bottleneck in available bandwidth. Moreover,

according to obtained results, the type of Internet

connection did not directly affect the CPU and RAM

usage.

When analyzing particular streaming services,

it can be noticed that they strictly depend on the

particular application. The Netflix mobile application

consumed an average of approx. 40%, whereas

YouTube and Twitch apps used approx. 35% and

50%, respectively. The average RAM usage was

lowest in case of YouTube, resulting in approx. 8%,

whereas Netflix and Twitch apps required a little

more, namely 10% on average.

As expected, streaming audio files required less

processing power than streaming mixed audio-video

files. The Spotify platform used 30% of the CPU

processing power, whereas Tidal required only

approx. 20%. In case of both applications, the RAM

usage oscillated around 8-10%.

7 CONCLUSIONS

The conducted research had shown that the mere

change in quality of consumed content did not

significantly affect the usage of CPU and RAM

resources. In case of a dedicated mobile application,

the type of Internet connection did not contribute to a

significant change in the resource consumption as

well.

Table 8: Overall results concerning CPU and RAM usage with respect to type of streaming service, type of content,

and type of network connectivity.

Scenario Approach no. Avg. CPU usage [%] Std. deviation Avg. RAM usage [%] Std. deviation

S1 1 38.79 18.38 9.44 0.80

2 41.29 18.89 10.25 0.58

3 40.74 16.53 10.20 0.50

4 34.31 21.99 8.06 0.52

5 35.01 20.19 8.05 0.35

6 58.46 22.67 7.81 0.26

7 50.47 11.31 11.95 0.28

8 53.23 13.69 9.54 0.22

9 75.56 22.16 8.83 0.24

S2 1 33.16 6.38 9.49 0.14

2 29.71 8.00 10.27 0.15

3 30.39 5.92 7.80 0.21

4 16.64 10.57 10.45 0.43

5 21.44 14.16 7.20 0.21

S3 1 34.25 18.40 7.89 0.70

2 34.17 19.62 11.17 0.51

3 32.82 24.92 8.90 0.33

4 41.30 27.47 9.52 0.36

5 43.10 13.81 12.36 0.59

6 44.19 12.11 9.47 0.75

S4 1 27.55 10.44 7.18 0.22

2 61.85 14.76 10.53 0.44

3 19.39 19.73 7.51 0.67

4 20.50 16.74 7.82 0.45

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

240

On the other hand, this research experiment also

shown that high network bandwidth and stable

connection enables high-quality media streaming

without the need for buffering.

Streaming services encode the transmitted

multimedia implicitly, which may result in the direct

usage of CPU and RAM resources. This could be one

of the reasons why did the resource usage of a mobile

device differ. These differences may also result from

the developers’ approach to optimizing mobile

applications, etc. Currently, a broad range of Android

mobile devices is freely available on the market.

That is why most developers try to make their

products widely acceptable (e.g. due to the number of

distributions of an operating system available on the

market). The set of smartphones, utilized during this

study, is new and up to date. This may be one of the

reasons why the differences in resource usage did not

significantly differ.

It seems that the question regarding code

optimization, resource usage, network connectivity,

etc., remains open. Future investigation may be

related with a broader range of consumer devices,

including a wide variety of manufacturers, different

distributions of the Android operating system, as well

as diverse Wi-Fi access point manufacturers and

cellular network providers.

REFERENCES

Bouraqia, K., Sabir, E., Sadik, M., Ladid, L. (2020).

Quality of experience for streaming services:

measurements, challenges and insights. IEEE Access, 8,

13341-13361.

Falkowski-Gilski, P. (2020). On the consumption of

multimedia content using mobile devices: a year to year

user case study. Archives of Acoustics, 45(2), 321-238.

Falkowski-Gilski, P., Uhl, T. (2020). Current trends in

consumption of multimedia content using online

streaming platforms: a user-centric survey.

Computer Science Review, 37, 100268.

Gast, M. S. (2005). 802.11 Wireless Networks:

The Definitive Guide, O’Reilly Media. Sebastopol,

edition.

Gazdar, A., Alkwai, L. (2018). Toward a full peer to peer

MPEG-DASH compliant streaming system. Multimed

Tools and Applications, 77, 15829-15849.

Hoßfeld, T., Seufert, M., Sieber, C., Zinner, T.,

Tran-Gia, P. (2015). Identifying QoE optimal

adaptation of HTTP adaptive streaming based on

subjective studies. Computer Networks, 81, 320-332.

Iwacz, G., Jajszczyk, A., Zajaczkowski, M. (2008)

Multimedia Broadcasting and Multicasting in Mobile

Networks, John Wiley & Sons. Chichester, United

Kingdom, 1

edition.

Kohli, C. (2020). The replacement of conventional

television by streaming services. International Journal

of Research in Engineering, Science and Management,

3(10), 59-67.

Kryvinska, N., Greguš, M. (Eds.) (2019). Data-Centric

Business and Applications. Lecture Notes on Data

Engineering and Communications Technologies,

Springer. Cham, 1

edition.

Meraj, M., Kumar, S. (2015). Evolution of mobile wireless

technology from 0 G to 5 G. International Journal of

Computer Science and Information Technologies, 6(3),

2545-2551.

Muscat, S. (2019). Music in the digital age: streaming and

downloading. Bachelor’s thesis, University of Malta.

Polak, L., Milos, J. (2020). Performance analysis of LoRa

in the 2.4 GHz ISM band: coexistence issues with

Wi-Fi. Telecommunication Systems, 74, 299-309.

Raca, D., Leahy, D., Sreenan, C. J., Quinlan, J. J. (2020).

Beyond throughput, the next generation: a 5G dataset

with channel and context metrics. In MMSys’20,

11th ACM Multimedia Systems Conference, 303-308.

Shen, X., Lin, X., Zhang, K. (Eds.) (2020). Encyclopedia of

Wireless Networks, Springer. Cham, 1

edition.

Vetro, A. (2011). The MPEG-DASH Standard for

Multimedia Streaming Over the Internet. IEEE

Multimedia, 18(4), 62-67.

Study on CPU and RAM Resource Consumption of Mobile Devices using Streaming Services

241