Conference PaperPDF Available

A Software Defined Network Based Research on Fairness in Multimedia


Abstract and Figures

The demand for online distribution of high quality and high throughput content has led to non-cooperative competition of network resources between a growing number of media applications. This causes a significant impact on network efficiency, the quality of user experience (QoE) as well as a discrepancy of QoE across user devices. Within a multiuser multi-device environment, measuring and maintaining perceivable fairness becomes as critical as achieving the QoE on individual user applications. This paper discusses application-and human-level fairness over networked multimedia applications and how such fairness can be managed through novel network designs using programmable networks such as software-defined networks (SDN).
Content may be subject to copyright.
A Software Defined Network Based Research
on Fairness in Multimedia
Ahmed Osama Basil
University of Northampton
Northampton, United Kingdom
Mu Mu
University of Northampton
Northampton, United Kingdom
Ali Al-Sherbaz
University of Northampton
Northampton, United Kingdom
The demand for online distribution of high quality and high
throughput content has led to non-cooperative competition of
network resources between a growing number of media appli-
cations. This causes a significant impact on network efficiency,
the quality of user experience (QoE) as well as a discrepancy
of QoE across user devices. Within a multi-user multi-device
environment, measuring and maintaining perceivable fairness
becomes as critical as achieving the QoE on individual user
applications. This paper discusses application- and human-
level fairness over networked multimedia applications and
how such fairness can be managed through novel network de-
signs using programmable networks such as software-defined
networks (SDN).
Networks Network resources allocation
;Network algo-
rithms;Network performance evaluation;Network dynamics;
Human-centered computing
Human computer interac-
tion (HCI);
fairness; quality of experience; human factor; software defined
networking; multimedia
ACM Reference format:
Ahmed Osama Basil, Mu Mu, and Ali Al-Sherbaz. 2019. A Software
Defined Network Based Research on Fairness in Multimedia. In
Proceedings of 1st International Workshop on Fairness, Account-
ability, and Transparency in MultiMedia, Nice, France, October
25, 2019 (FAT/MM ’19), 8 pages.
Recent years have seen a dramatic increase in the general
quality of experience expected by the users [
] over media
delivery platforms such as online video streaming. As the
Permission to make digital or hard copies of all or part of this work
for personal or classroom use is granted without fee provided that
copies are not made or distributed for profit or commercial advantage
and that copies bear this notice and the full citation on the first
page. Copyrights for components of this work owned by others than
ACM must be honored. Abstracting with credit is permitted. To copy
otherwise, or republish, to post on servers or to redistribute to lists,
requires prior specific permission and/or a fee. Request permissions
FAT/MM ’19, October 25, 2019, Nice, France
©2019 Association for Computing Machinery.
ACM ISBN 978-1-4503-6915-2/19/10. . . $15.00
number of online user applications and use devices grows, the
capability of end devices varies significantly and resources
such as network bandwidth are often shared between multiple
user devices. Hence the quality that can be achieved on
a single device is often limited by its share of resources.
Conventional best-effort networking infrastructures allocate
resources based on client requests and high-level service level
agreement (SLA) without the awareness of application or user-
level requirements. This often leads to unfairness perceivable
by the end user. There is a considerable amount of literature
on the concept of fairness. However, most studies have their
own scopes of vision, such as identifying the fairness issue
that needs adjusting, and conducting the research needed to
rectify a single issue. However a major challenge with this
kind of topic is the perspective, as mentioned before, fairness
can be looked at from different levels, and aspects. Fairness
can be considered from a hardware level or a software level, it
can be showed in QoE and quality of service (QoS). The issue
of fairness has been a controversial and much disputed subject
within the field of multimedia. There is a lack of a holistic
research of fairness on online multimedia. The purpose of
this paper is to review recent research into the definitions of
fairness and how researchers attempt to tackle the related
challenges. We also investigate how programmable networks
such as SDN can be used to improve the fairness of online
multimedia applications.
Our work provides an important opportunity to advance
the understanding of fairness and how it can be organized and
achieved. The use of micro-segmentation to achieve fairness
is the aspect that this paper will focus on. Investigation and
research will cover fairness in aspects that exceeds resource
allocation and QoE. Fairness of security and the role of
network management to make a network fair and safe will
also be considered in this paper. Most current literature on
fairness pay particular attention on partial issues or single
level, and by solving them, partial fairness is gained. (example:
[4, 18, 34, 46, 55]).
The aim of the paper is to provide a conceptual framework
supported by SDN-based monitoring and management for
online multimedia. Main contributions of this paper include:
Explore existing definitions of fairness and their impact
on online multimedia applications.
A fairness flow model and the relationship between its
levels and multimedia.
A human-level fairness consideration that exceeds the
application level’s fairness definitions and point of view.
The impact of Human-to-Computer fairness on media
A proposed fairness framework structure through pro-
grammable network management.
The overall structure of the study takes the form of six
sections. In section 2 the definition and understanding of
fairness at different levels will be explained; ranging from
network-level fairness to human-level fairness and the pro-
cess to measure them. Section 3 presents the findings of the
research on tools used to measure fairness. Section 4 gives
the proposition of how to increase the efficiency of achieving
fairness for media applications for better quality of experience
and service through a framework. Finally, section 5 tie up
the various theoretical and empirical strands in order to sum
up the findings, along with the inclusion of a discussion of
the implication of the findings to future research into this
research area.
2.1 Definitions of Fairness
Fairness may have different definitions and interpretations
developed in various contexts. It is necessary here to explore
the linguistic definitions “fairness”. In Cambridge Dictionary
fairness is defined as “considering everything that has an
effect on a situation, so that a fair judgment can be made”, it
is also defined as “the quality of treating people equally or in
a way that is right or reasonable” [
]. Oxford Dictionaries
defined fairness as “Impartial and just treatment or behaviour
without favouritism or discrimination” [
]. In “Vocabulary”
online dictionary, fairness is defined as “The quality of making
judgments that are free from discrimination” [14].
Fairness is a term frequently used in literature, to date
there is a clear consensus about its meaning, as shown above
the definitions share a common factor which is the distribu-
tion of the fairness target object in an equal and organized
manner between any and all contributed parties. There is
a degree of uncertainty around the terminology in the field
of technology. Moreover these definitions are limited to the
verbal sense of the sentence which is dependant on a scenario,
however they all focus on equality, justice, quality of distri-
bution, and the fairness target audience’s satisfaction. This
section aims for the understanding of these definitions in the
field of technology in media streaming.
2.2 Fairness in Media Streaming
Fairness can be addressed and researched from many point
of views. In some cases, researchers dismiss the idea that
fairness can be targeted, because the methodology of equal
opportunity to the individuals in resource sharing may not
always mean equal allocation of resources, however a fair
allocation can always be the outcome of a process where
individuals do not have equal opportunity [
]. Thus the
objective and target of any fairness must be explained. Not
to forget that fairness on different networking levels will be
discussed, from a single user within the network and the
network as a whole.
The past twenty years have seen increasingly rapid ad-
vances in the field of media streaming, it has evolved greatly
from the ability to download videos to streaming them [
Due to the high demand in media streaming, media companies
and internet providers take QoE in high priority and serious-
ness [
]. The high demand naturally resulted in high number
of requests, real time requirement of the video feedback, and
high bit rates of the media content [
]. On a user-level, the
users’ expectations include the ability to stream with ease in
their environment while engaging in other network activities.
Thus media fairness can be improved by ameliorating the ini-
tial delay, and interruptions of the playback which is known
as stalling [
]. These aspects are improved by creating new
algorithms that aim to make a fair distribution of resources
for all users to maximize users’ experience and satisfaction.
2.3 Fairness Flow Model
The Open Systems Interconnection Model (OSI) is one of
the most known models that describe traffic flow from the
Network Level, to the Application Level. This section will look
at the perspective of fairness from each of the relevant levels,
highlighting the Network, transport, and application levels for
the software level. Along with a new Level titled the Human
Level which will state the importance of fairness between
human and computer [
]. Figure 1 shows the Fairness Flow
Model (FFM) which details the flow of fairness on three
different levels. This section will explain those levels and
previous research and techniques discussed on them.
Figure 1: Fairness Flow Model (FFM)
2.3.1 Network and Transport Level Fairness. Generally for
networking and media distribution, the network and transport
levels are two of the most important levels to be discussed
and defined while noticing their affect on the concept of fair-
ness. Network-level is important to us because its fairness
affects all other levels’ fairness. It considers transmission pri-
ority, the level of network congestion, QoS and occasionally,
dynamic routing for the determination of the cost from one
network node to another and identifying the best path for
traffic flow [
]. Often in the network-level routing addressing
schemes, models and guidelines are re-visited and adjusted to
best suit the network’s accessibility for data transmission and
resulting in a better QoE for the end-user. On the other hand,
the transport level provides reliable data transmission. If the
network level was defined as the start of a point-to-point
connection process then we can define the transport level as
the end process of the source to destination procedure. Trans-
port level is responsible for virtual circuit management, error
tracking, correction and recovery, and more importantly flow
control and multiplexing. These responsibilities are highly
important and should always be considered when working
with the term “fairness”.
There are many network and transport level protocols
ranging from Internet Protocol (IP), Internetwork Packet
Exchange (IPX), Routing Information Protocol (RIP), and
Open Shortest Path First (OSPF) in the Network-level.
Where the transport-level includes; Transmission Control
Protocol (TCP), User Datagram Protocol (UDP), and Se-
quenced Packet Exchange (SPX) with many others. The
following subsection will discuss the TCP fairness and the
algorithms proposed to show network-level fairness and limit
steady-packet drop rate.
According to a recent study mentioned in [
] and [
after a clear observation of UDP and TCP based protocols,
it was deducted that UDP provides no fairness at all. Due
to the fact that it is an un-ordered lightweight datagram
based service. Analysis show that UDP and TCP usage-
ratios in media-service companies lie between 5% and 20%
depending on the popularity of the media server and the
characteristics of the end-user’s settings and choices. Even
though UDP is preferred for applications that are based
on media streaming or VoIP and most types of tunneling
applications, it will be difficult to replace the value of TCP
due to UDP’s unreliability and lack of fairness mechanisms.
UDP Fairness; UDP is regarded as a communication pro-
tocol to TCP, where its target is to establish low-latency
and loss-tolerating connections between servers on the web.
UDP is a best-effort protocol that sends datagrams, unlike
TCP which has the ability to break large sets of data into
packets and re-send any lost packets to assemble a correct
sequence for the packets, UDP simply sends packets. Pack-
ets may take random paths to reach destination and often
results in non-retrievable lost packets or received packets out
of order due to route delays. UDP is favoured by some media
application such as gaming or IPTV streaming, where the
loss of few packets will unlikely affect the resulted outcome
in a notable user-end perspective [
]. Applications that re-
quire loss-less data transmission may favour UDP as well.
Using application-level error control, those applications can
re-transmit lost packets, resulting in this protocol being a
better choice for large files’ data transfer rate compared to
Since UDP applications have increased, many attempts
where targeted at achieving fairness while using protocols
to assist UDP-based High-speed transport protocols. In a
performance evaluation [
], RUBDP, Tsunami, UDT, and
PA-UDP high speed UDP transport protocols where used
to gather protocol data, the researchers looked at the inter-
protocol fairness and intra-protocol fairness.
Given that e-science (distributed network collaboration en-
vironment) applications are in more demand, communication
patterns started preferring point-to-multipoint or multipoint-
to multipoint rather than point-to-point [
]. Intra-protocol
fairness shows multiple flows that are using the same protocol
and their ability to share bandwidth fairly and/or use the
same service with no connectivity obstacles. Since network
applications vary in the use of protocols, inter-fairness term
is used to explain an entire flow of data based on different
protocols and distribute the link bandwidth fairly between
each other. If fairness is achieved on both inter and intra-
protocol then the data flow will be fair, giving a better QoE
to the user.
2.3.2 Application Level fairness. The application level is
defined as user-oriented level. This level is responsible for pro-
viding an interface to the embedded network services for the
end-user. It is also responsible for making services such as file
transfer, file management and information processing. The
protocols that are involved and used in the application level
include: Telnet, File Transfer Protocol (FTP), Hypertext
Transfer Protocol (HTTP), and Simple Network Manage-
ment Protocol (SNMP) [
]. This level also highlights the
importance of QoS and QoE and how the application-level
makes the best effort to deliver the required features and
expectations. Some researchers have taken the opportunity
to study the application-aware network management and
try to enhance it in many ways. A research tried to achieve
fairness of adaptive audio application for multi-hop wireless
networks. The authors highlight the idea that a redundant
speech captioning scheme is necessary and when connections
are adaptive in multi-hop networks perception, efficient band-
width usage, and fairness is improved [27, 28].
Some other research explored application-level QoS fairness
in wireless video scheduling. It was argued that the initial
delay for pre-fetching video at the client buffer should be
shorter due to buffer limitations and application level user’s
QoE. A cross-layer optimized mutli-user video adaptation and
scheduling scheme for wireless video communication was pro-
posed in [
]. This research focuses on the video content
delivered to each user and targeting quality of user’s satisfac-
tion, video throughput was attempted to be maximized. Let
the instructions that are sent through the available channel
data rate at time kbe
. Let the average video throughput
be t(k) and
as the total number of video requests from
video streaming users. The scheduler decides on the target
for the channel throughput at any given time with source
encoding rate shown by Equation 1 [42].
𝑖=1𝑎𝑖𝑘·𝑡𝑖,𝑙 𝑘(1)
This scheme was used for video communication over pack-
etised networks. It was tested in multiple environments to
evaluate the QoE for users while playing continuous high def-
inition videos and monitor fairness. Using the scheme it was
noted that the cross-layer shows a decrease in the undesired
pauses in playback which increased the QoE for users.
Figure 2: Human Level Affects, Relations and Research on
Moreover authors also investigated fair load share in multi-
source multimedia overlay in application level multi-cast.
They discussed the efficiency and scalability of IP multicas-
ting and its ability to deliver high quality video and audio
streams for group communication. Since the spread in IP
multicast is limited in near future unless substantial infras-
tructure modifications takes place, the authors proposed an
Application Layer Multicast (ALM) scheme with its ability
to support group communication independent of network
level support. The target was to achieve fair load sharing and
overall communication quality among the participants in a
multi-source multimedia communication [35, 39, 44, 57, 59].
2.3.3 Human Level fairness. Human factors are essential
to understand and maximise fairness. Studies have shown
that fairness can be observed on multiple points of view. In
our scope, it is essential to discuss the fairness perceived
between human and machine, and how it affects fairness for
human to human, leading to a wider scope of vision and
understanding of robot to robot fairness. Many researchers
discussed robot to robot fairness but this type of fairness
can be viewed in many ways, it could be simple machine
data distribution fairness or allocation fairness of resources
between devices equipped with Artificial Intelligence. Figure
2 shows the aspects that will be discussed in this subsection.
Human to Human Fairness;
Many scholars and researchers have discussed the importance
of fairness, and how it affects social life. Fairness is quite
complex due to the fact that it can be referred to in countless
scenarios. The central understanding and mature concept of
fairness in humans is shown in the principles governing the al-
location of resources, so-called distributive justice [
Psychological understanding of human behaviour of how peo-
ple react to unequal distributions such as absence of richer
contextual information like neediness and deservedness is the
key to unlocking human fairness [
]. The concept of hu-
man fairness has dominated the research field of exploration
of studies in disciplines, such as psychology, economy and
anthropology. If we look in technology we can see a clear
resemblance of fairness and its attributes. Simply on net-
working and source allocation; host devices compete to get
more resources, timeliness, and access to the Internet over a
shared connection. We can conclude that the fairness rules
made in technology are based or similar to the social real-life
fairness and justice. To be fair between humans and what
they get in QoE along with QoS provided for them is to get
equal resources from (for example) a governing body, and an
exponential growth in success with effort made by the human,
in what they desire to achieve in life. Naturally we can see a
clear link between fairness in humans and technology.
The ability to predict the provider’s QoS and the user’s
QoE is highly important to us. With the correct additions
and interactions, fairness can lead to best possible QoE on a
human-user level. Studies have shown that QoE management
mechanisms are found usually in Application and/or Network
levels where a process of mapping functions between QoS and
QoE when optimizing the system occur. Some researchers
have used illustrative numerical examples to show that the
choice of different type of metrics for quality estimation cor-
responds to optimizing the system for different types of users,
leading to different QoE management outcomes such as opti-
mal and fair QoE resource allocation [
]. There are different
types of users and thus service providers must address each
of their target audience in specific way to their needs and ex-
pected services to gain the maximum QoE from a user’s point
of view. Many authors have journeyed to achieve fairness in
the field of media technology, an interesting approach of this
(in addition to what was mentioned previously in other levels)
can be the adaptation algorithmic logic for increasing QoE
fairness for HTTP Adaptive Video streaming [
], where a
simulative performance environment evaluation is conducted
to compare the QoS and QoE fairness to achieve fairness
between users or service providers. Previous studies have
proven that QoE directly affects QoS provided by the service
providers [
] and since that is true, then the most desired
approach for Human to Human fairness is to maximize QoE
while ensuring fairness among users.
Human to Computer & Human to Robot Fairness;
To understand and view the fairness between human and
machine, HCI and HRI must be explained. Human to com-
puter interactions (HCI) has emerged as the center of the
computing research, moreover it is highly useful for under-
standing and improving interactions with computer-based
technologies. As time progressed with computer technologies,
breakthroughs led to robotic technology, which naturally
represents huge implications for the Human to Robot in-
teractions (HRI). HCI can have many modes including, (I)
data; is the communication stage of human and computer,
providing signs such as figures, colors and graphs. (II) Im-
age; the ability of computer recognition on images which is
shown in image processing, recognition and perception. (III)
Voice; the combination of audio frequency and stored data
to communicate. (IV) Intelligent reactions; computer’s pre-
diction to human actions based on the data given from their
behaviors and needs (AI). These interactions are examples
of how HCI can be described [
]. Previous studies suggested
that a significant aspect of collaborative environments and
interactions is fairness [
]. In this scenario there must be
some form of an understanding and given properties that
outline fairness to humans, computers and robots. A task
can be given to any of these entities, however with a fairness
point of view where each has its/his own fair sources and
tools, even the highest difficulty tasks can be accepted if all
the collaborators are convinced that they are being treated
fairly. Authors have also pointed out that in HRI, some issues
rise in the human’s point of view. For instance, it is impor-
tant to consider the fact that human perception of robots is
different from other computer technologies, due to the idea
that people humanize robots. Also robots are able to learn
about themselves and their surroundings and make their own
decisions depending on the situation, this is a clear difference
between HRI and HCI [
]. HCI is a static and restrictive
environment where it can only be controlled by a human
entity, HRI are shown to be more dynamic with their interac-
tions [
]. Internet of Things (IoT) devices can be described
here as a form of computing environment. Resource sharing
between human and IoT devices can be a combination of a
static and a dynamic environment. Fairness between them
must be considered as IoT devices can have machine learning
algorithms that sort and distribute resources to humans and
other connected devices. QoS and QoE must be achieved
through techniques such as Max-min fairness for resource
sharing to be achieved between those two entities.
Network fairness can be measured in many different ways, this
section will explain the aspects and techniques used to mea-
sure fairness quantitativelyto insure maximum understanding
of the nature of fairness and its properties. Quantitative re-
search is a numerical based research. It relies on numbers as
the main unit of analysis, this type of research is highly used
in many scientific research fields. Quantitative results usually
give accurate data based on measured algorithms, which is
why it is a key category of techniques that must be discussed
in the process of Fairness measurements [5, 12, 52].
3.1 Quantitative measurements
3.1.1 Jain’s Fairness Index. Jain’s fairness index [
] is
one of the most recognized and used indexes of fairness
𝐽𝑥1, 𝑥2,...,𝑥𝑛=
represents the users and this equation rates the fairness
of a set of values where
is the throughput for the
connection and
is the sample coefficient of variation.
This index gives the result of 1 for a best case allocation
and a
worst case representation [
]. This index aims to
achieve equal share of bottleneck;optimal allocation or equal
fraction of optimal allocation to users. It is scale independent
(applies to any number of users
), bounded between 0 and 1
where variance, standard deviation and relative distance are
not bounded. Jain’s fairness also shows a direct relationship
of more generated fairness in a higher index and vice-versa
3.1.2 Max-min Fairness. Max-min fairness is the process
of applying flexible sort of allocation. Max-min fairness can
be useful in developing an intelligent fairness mechanism,
this is due to the fact that the Max-min allocation technique
allows an increase in data flow and balances by creating a de-
crease in data flow in another allocation to create an equally
balanced flow of data. To expand on this, multimedia appli-
cations can achieve max-min fairness in machine learning on
a network-level. To have the ability to recognize what the
user is using the network for is essential to achieving multi-
media fairness. Max-min fairness modal aims to distribute
network bandwidth equally with infinitesimal increments to
all flows until one is satisfied. With the right understanding of
Max-min fairness and a machine learning modal, multimedia
applications can achieve fairness, which increases the QoE
on an application level [8, 33, 50, 53].
3.2 Qualitative measurements
Qualitative measurement techniques are as important to this
research as quantitative techniques. These type of techniques
pays more attention to non-numerical aspects of measure-
ments. A major qualitative technique is QoE fairness mea-
surements. This technique focuses on considering the QoE as
perceived by the end user [
]. This is important
in co-operations where operators want to keep their users suf-
ficiently satisfied in network management. Many techniques
were proposed especially for the field of adaptive video stream-
ing [
]. The most obvious technique is a questionnaire
survey, distributed among end-users with their suggestions
for increasing QoE fairness. QoE cannot be measured on
ratio scales, thus previous quantitative measurements will
not be applicable. One example to measure QoE can be an
interval scale such as a simple 5-point mean opinion score
scale (MOS). Users with can rate the experience from 1 being
the lowest to 5 being the highest quality on the MOS scale.
Software-Defined networking, with its own architectural ap-
proach, it is thought of as the separation of the management
of the control plane of devices from the underlying data plane
that forwards network traffic [
]. SDNs mainly introduces an
abstraction layer separating network configuration from the
physical communication resources. This is highly useful due
to the fact that a network operating system running inside
a control layer which is sandwiched between the application
and infrastructure layers allow more room for applications to
re-configure dynamically to adapt to their security, scalability
and manageability needs [37].
The purpose of the current study was to determine the
definition of fairness on different levels and points of view
in multimedia and how it can be affected by and achieved
through networking mechanisms. The most important finding
Figure 3: Proposed Fairness Theoretical Framework Structure
to emerge from this study is the importance of the human-
level fairness as an addition on the resulted Fairness Flow
Model in Figure 1. The Human aspect and point-of-view
on multimedia fairness is quite important because it has a
direct affect on all the other levels in the FFM. The gen-
eralisability of these levels is subject to certain limitations.
For instance, future research must expand the network-level
fairness within the FFM with an addition of hardware-based
level fairness. Future research will include an investigation
on Physical, and Session-level fairness and their perspective
on SDN algorithms in multimedia. An additional proposi-
tion is to create a plug-and-play test bed and using SDN
to program a network controller to understand user’s needs
and requests with applying Max-min fairness strategy after a
thorough research on the best language to base the test bed
on, such as; OpenFlow or P4. This will allow a network con-
troller to distinguish between users’ needs, where if one user
requires more media network resources, then the controller
will grant the user its needed bandwidth, while decreasing
the bandwidth for another user who does not need its entire
bandwidth. This will create best QoE possible for both and
all users.
Figure 3 shows a proposed framework structure on achiev-
ing fairness in the use of multimedia applications with SDN.
Framework Inputs; in this framework, the network pro-
tocol inspection using the flow manager (an SDN appli-
cation) along with the Media Presentation Description
(MPD) Parser help the controller identify the number
of devices in the network that are requesting stream-
ing bitrates and the MPD files requested by the users.
This information is essential to the SDN controller as
it will use the Application Rule Manager to filter the
provider’s content for security measures. This process
will occur as the requested content provider reaches
the SDN controller. The controller at this stage gains
the user’s media duration and the requested encoding
Framework Functions; in addition to the filtering that
occurs above, the flow manager will ensure the accurate
flow to the switch. This affects the decisions made by
the framework’s intelligence aspect because it makes
sure that all users receive the requested video stream
from the provider, through the controller.
Framework Intelligence; using the max-min fairness
allocation of resources technique, MPD parser, and
application rule manager will not only identify the
network traffic but will also distribute the traffic and
bitrates fairly and securely to the user.
However more research is required to determine the efficacy
of the proposed framework with a functional test-bed. More
aspects to research can include a Fairness time period (dura-
tion), the importance of long-term and short-term fairness
and which is more applicable for better demonstration of
defining and identifying fairness. Despite its exploratory na-
ture, this study offers some insight into fairness measurement
tools such as quantitative techniques, further research must
include qualitative techniques at measuring fairness to further
the understanding of Fairness in multimedia technologies.
Understanding cross-device and cross-user fairness has be-
come as crucial as the QoE on individual user devices. This
paper discusses different levels of fairness considerations in
multimedia applications. We also discussed how perceivable
fairness is linked to resource allocation and traffic engineer-
ing at the network level and how emerging programmable
networks such as SDN can be used as a tool to improve
fairness. A cross-layer fairness framework is also proposed to
harness the capabilities of new network designs and growing
availability of computing resources in future networks for
fairness-aware content distribution. Our future work will look
into implementations in a smart home environment.
This work is supported by UK Research and Innovation
(UKRI) under EPSRC Grant EP/P033202/1(SDCN).
J. S Adams. 1965. Inequity in social exchange. Adv. Exp. Soc.
Psychol. 2, 267âĂŞ299 (1965).
Aristotle. 1999. The Nicomachean Ethics. Harvard Univ. Press
A. V. Babu and L. Jacob. 2007. Fairness Analysis of IEEE 802.11
Multirate Wireless LANs. IEEE Transactions on Vehicular
Technology 56, 5 (Sep. 2007), 3073–3088.
Ahmed O Basil, Mu Mu, and Michael Opoku Agyeman. 2019. A
Multi-Modal Framework for Future Emergency Systems. In IEEE
Smart World Congress 2019 (SWC 2019). Leicester, United
Kingdom (Great Britain).
S. F. Bessane, M. S. Camara, I. Fall, and A. Bah. 2018. Causal
model of performance measurement systems by combining qualita-
tive and quantitative models for robust results. In 2018 Interna-
tional Conference on Intelligent Systems and Computer Vision
(ISCV). 1–7.
B Butler. 2019. What SDN is and where itâĂŹs going. [online]
Network World.
G. Chao. 2009. Human-Computer Interaction: Process and Princi-
ples of Human-Computer Interface Design. In 2009 International
Conference on Computer and Automation Engineering. 230–233.
J. Choi. 2016. Power Allocation for Max-Sum Rate and Max-Min
Rate Proportional Fairness in NOMA. IEEE Communications
Letters 20, 10 (Oct 2016), 2055–2058.
S. CicalÚ, N. Changuel, V. Tralli, B. Sayadi, F. Faucheux, and S.
Kerboeuf. 2016. Improving QoE and Fairness in HTTP Adaptive
Streaming Over LTE Network. IEEE Transactions on Circuits
and Systems for Video Technology 26, 12 (Dec 2016), 2284–2298.
J. Pan D. Ghadiyaram and A. C. Bovik. 2015. A Time-varying
Subjective Quality Model for Mobile Streaming Videos with
Stalling Events (proceedings of spie applications of digital image
processing xxxviii, ed.). San Diego, CA, USA.
M Deutsch. 1975. Equity, equality, and need: what determines
which value will be used as the basis of distributive justice? J.
Soc. Issues 31, 137âĂŞ149 (1975).
M. Di Penta and D. A. Tamburri. 2017. Combining Quanti-
tative and Qualitative Studies in Empirical Software Engineer-
ing Research. In 2017 IEEE/ACM 39th International Confer-
ence on Software Engineering Companion (ICSE-C). 499–500.
Oxford Dictionaries. 2019. fairness | Definition of fairness in Eng-
lish by Oxford Dictionaries. (2019). https://en.oxforddictionaries.
Dictionary. 2019. Definition of fairness. (2019). https://www.
Cambridge Dictionary. 2019. Definition of Fairness. (2019). https:
Myriam El Mesbahi, Nabil Elmarzouqi, and Jean-Christophe La-
payre. 2014. Fairness Properties for Collaborative Work Using
Human-Computer Interactions and Human-Robot Interactions
Based Environment: “Let Us Be Fair”. In Distributed, Ambient,
and Pervasive Interactions, Norbert Streitz and Panos Markopou-
los (Eds.). Springer International Publishing, Cham, 319–328.
Nabil Elmarzouqi, Eric Garcia, and Jean-Christophe Lapayre.
2008. CSCW from Coordination to Collaboration. In Computer
Supported Cooperative Work in Design IV, Weiming Shen, Jian-
ming Yong, Yun Yang, Jean-Paul A. Barthès, and Junzhou Luo
(Eds.). Springer Berlin Heidelberg, Berlin, Heidelberg, 87–98.
U. Engelke and H.-J. Zepernick. 2007. Perceptual-based quality
metrics for image and video services: A survey. Proc. 3rd
EuroNGI Conf. Netw., Trondheim, Norway. pp. 190âĂŞ197.
WikiMedia Foundation. 2019. Fairness measure. (May 2019).
Elkhatib Yehia Broadbent Matthew Mu Mu Race Nicholas
Georgopoulos, Panagiotis (Ed.). 2013. Towards network-wide
QoE fairness using openflow-assisted adaptive video streaming.
Vol. Proceedings of the 2013 ACM SIGCOMM Workshop on Fu-
ture Human-centric Multimedia Networking. ACM SIGCOMM.
[21] Hagen. 2011. TCP Fairness.
M. Han and J. P. Hong. 2012. Balancing QoE and fairness of lay-
ered video multicast in LTE networks. In 2012 14th International
Conference on Advanced Communication Technology (ICACT).
T. Hoçfeld, P. E. Heegaard, L. Skorin-Kapov, and M. Varela.
2017. No silver bullet: QoE metrics, QoE fairness, and user
diversity in the context of QoE management. In 2017 Ninth
International Conference on Quality of Multimedia Experience
(QoMEX). 1–6.
T. Hoçfeld, L. Skorin-Kapov, P. E. Heegaard, and M. Varela.
2017. Definition of QoE Fairness in Shared Systems. IEEE
Communications Letters 21, 1 (Jan 2017), 184–187. https://doi.
Raj Jain. 2019. Fairness: How to measure quantitatively. Re-
searchGate (05 2019).
[26] Raj Jain, Dah Ming Chiu, and Hawe WR. 1998. A Quantitative
Measure Of Fairness And Discrimination For Resource Allocation
In Shared Computer Systems. CoRR cs.NI/9809099 (01 1998).
M. Kazantzidis and M. Gerla and. 2001. Permissible throughput
network feedback for adaptive multimedia in AODV MANETs. In
ICC 2001. IEEE International Conference on Communications.
Conference Record (Cat. No.01CH37240), Vol. 5. 1352–1356
M. I. Kazantzidis, L. Wang, and M. Gerla. 1999. On fairness and
efficiency of adaptive audio application layers for multihop wireless
networks. In 1999 IEEE International Workshop on Mobile
Multimedia Communications (MoMuC’99) (Cat. No.99EX384).
Sara Kiesler and Pamela Hinds. 2004. Introduction to This Special
Issue on Human-robot Interaction. Hum.-Comput. Interact. 19,
1 (June 2004), 1–8.
N. Kongwatmai and K. Rojviboonchai. 2012. Qualitative study
of TCP NewRAPID. In 2012 9th International Conference on
Electrical Engineering/Electronics, Computer, Telecommunica-
tions and Information Technology. 1–4.
D. Lee, B. E. Carpenter, and N. Brownlee. 2010. Observations
of UDP to TCP Ratio and Port Numbers. In 2010 Fifth Inter-
national Conference on Internet Monitoring and Protection.
Y. Li, D. Li, W. Cui, and R. Zhang. 2011. Research based on
OSI model. In 2011 IEEE 3rd International Conference on
Communication Software and Networks. 554–557. https://doi.
Y. Li, M. Sheng, C. W. Tan, Y. Zhang, Y. Sun, X. Wang, Y. Shi,
and J. Li. 2015. Energy-Efficient Subcarrier Assignment and Power
Allocation in OFDMA Systems With Max-Min Fairness Guaran-
tees. IEEE Transactions on Communications 63, 9 (Sep. 2015),
W. Lin and C. C. J. Kuo. 2011. Perceptual visual quality metrics:
A survey (vol. 22, no. 4, pp. 297âĂŞ312 ed.). J. Vis. Commun.
Image Represent.
J. Makishi, D. Chakraborty, T. Osada, G. Kitagata, A. Takeda, K.
Hashimoto, and N. Shiratori. 2010. A Fair Load Sharing Scheme
for Multi-source Multimedia Overlay Application Layer Multicast.
In 2010 International Conference on Complex, Intelligent and
Software Intensive Systems. 337–344.
Chuk Moozakis Margaret Rouse, George Lawton. 2018.
UDP (User Datagram Protocol). Online. (June 2018).
Dan C. Marinescu. 2018. Chapter 2 - Cloud Service Providers
and the Cloud Ecosystem. In Cloud Computing (Second Edition)
(second edition ed.), Dan C. Marinescu (Ed.). Morgan Kaufmann,
13 – 49.
Blake P. R. Steinbeis N. & Warneken F. McAuliffe, K. 2017. The
developmental foundations of human fairness. Nat. Hum. Behav.
1, 0042 (2017).
A. Mourad and M. Ahmed. 2008. A Scalable Approach for Appli-
cation Layer Multicast in P2P Networks. In 2008 Sixth Annual
IEEE International Conference on Pervasive Computing and
Communications (PerCom). 498–503.
S. Orzen. 2014. Interaction understanding in the OSI model func-
tionality of networks with case studies. In 2014 IEEE 9th IEEE
International Symposium on Applied Computational Intelligence
and Informatics (SACI). 327–330.
J. Otwani, A. Agarwal, and A. K. Jagannatham. 2015. Opti-
mal Scalable Video Scheduling Policies for Real-Time Single-
and Multiuser Wireless Video Networks. IEEE Transactions
on Vehicular Technology 64, 6 (June 2015), 2424–2435. https:
T. Ozcelebi, M. O. Sunay, M. R. Civanlar, and A. M. Tekalp. 2006.
Application-Layer QoS Fairness in Wireless Video Scheduling. In
2006 International Conference on Image Processing. 1673–1676.
Latre Steven Famaey Jeroen De Turck Filip Petrangeli Stefano,
Claeys Maxim. 2014. A multi-agent Q-Learning-based framework
for achieving fairness in HTTP Adaptive Streaming. IEEE Net-
work Operations and Management Symposium (NOMS) (2014).
X. Rao, J. Xu, Q. Xu, and Y. Zhou. 2009. Study of the efficiency
optimization of Application Layer Multicast. In 2009 IEEE Sym-
posium on Industrial Electronics Applications, Vol. 1. 170–173.
J Rawls. 1971. A Theory of Justice. Harvard Univ. Press (1971).
M. Reisslein S. Chikkerur, V. Sundaram and L. Karam. 2011.
Objective video quality assessment methods: A classification,
review, performance comparison (ieee trans. broadcast., vol. 57,
no. 2, pp. 165âĂŞ182 ed.). IEEE Trans. Broadcast.
M. Seufert, S. Egger, M. Slanina, T. Zinner, T. Hoçfeld, and
P. Tran-Gia. 2015. A Survey on Quality of Experience of HTTP
Adaptive Streaming. IEEE Communications Surveys Tutori-
als 17, 1 (Firstquarter 2015), 469–492.
M. Seufert, N. Wehner, and P. Casas. 2019. A Fair Share for
All: TCP-Inspired Adaptation Logic for QoE Fairness Among
Heterogeneous HTTP Adaptive Video Streaming Clients. IEEE
Transactions on Network and Service Management 16, 2 (June
2019), 475–488.
M. Seufert, N. Wehner, P. Casas, and F. Wamser. 2018. A Fair
Share for All: Novel Adaptation Logic for QoE Fairness of HTTP
Adaptive Video Streaming. In 2018 14th International Confer-
ence on Network and Service Management (CNSM). 19–27.
H. SHI, R. V. Prasad, E. Onur, and I. G. M. M. Niemegeers. 2014.
Fairness in Wireless Networks:Issues, Measures and Challenges.
IEEE Communications Surveys Tutorials 16, 1 (First 2014),
A. Sideris, E. Markakis, A. Trigonis, G. Alexiou, E. Pallis, and C.
Skianis. 2014. MPEG-DASH over IDVB-T: The QoE unfairness
issue. In 2014 IEEE 19th International Workshop on Computer
Aided Modeling and Design of Communication Links and Net-
works (CAMAD). 70–74.
L. R. Student, R. Goubran, and F. Kwamena. 2018. Com-
puter vision-assisted human-in-the-loop measurements: augment-
ing qualitative by increasing quantitative analytics for CI situ-
ational awareness. In 2018 IEEE International Conference on
Computational Intelligence and Virtual Environments for Mea-
surement Systems and Applications (CIVEMSA). 1–6. https:
R. Sun, M. Hong, and Z. Luo. 2015. Joint Downlink Base
Station Association and Power Control for Max-Min Fairness:
Computation and Complexity. IEEE Journal on Selected Ar-
eas in Communications 33, 6 (June 2015), 1040–1054. https:
Sebastian Thrun. 2004. Toward a Framework for Human-robot
Interaction. Hum.-Comput. Interact. 19, 1 (June 2004), 9–24.
Y. Wang. [n. d.]. Survey of objective video quality measurements.
X. Wu and A. A. Chien. 2004. Evaluation of rate-based transport
protocols for lambda-grids. In Proceedings. 13th IEEE Interna-
tional Symposium on High performance Distributed Computing,
2004. 87–96.
G. Ye, C. Wu, J. Yue, S. Cheng, and W. Cheng. 2008. Research on
Design of Structured Application Layer Multicast Prototype about
Self-Similar Structured Multicast (S3M). In 2008 International
Seminar on Business and Information Management, Vol. 1.
K. Zeng, H. Yeganeh, and Z. Wang. 2016. Quality-of-experience
of streaming video: Interactions between presentation quality and
playback stalling. In 2016 IEEE International Conference on
Image Processing (ICIP). 2405–2409.
X. Zhang, X. Li, W. Luo, and B. Yan. 2009. On Improving
the Multicast Performance of Application Layer Multicast. In
2009 First International Workshop on Education Technology
and Computer Science, Vol. 1. 903–907.
Zhaojuan Yue, Yongmao Ren, and Jun Li. 2011. Performance
evaluation of UDP-based high-speed transport protocols. In 2011
IEEE 2nd International Conference on Software Engineering
and Service Science. 69–73.
Full-text available
Cloud service providers (CSP) provide on-demand cloud computing services, reduces the cost of setting-up and scaling-up IT infrastructure and services, and stimulates shorter establishment times for start-ups that offer or use cloud-based services. Task reneging or dropping sometimes occur when a task waits in the queue longer than its timeout or execution deadline, or it is compromised and must be dropped from the queue or as an active queue management strategy to avoid tail dropping of tasks when the queues are full. Reneged or dropped tasks could be resubmitted provided they were not dropped due to security reasons. In this paper, we present a simple M/M/c/N queueing model of a cloud computing physical machine, where the interarrival times and the services times are exponentially distributed, with N buffer size and c virtual machines running in parallel. We present numerical examples to illustrate the effect of task reneging and task resubmission on the queueing delay, probability of task rejection, and the probability of immediate service.
The paper proposes an algorithm of bandwidth distribution, ensuring fairness to end-users in computer networks. The proposed algorithm divides users into satisfied and unsatisfied users. It provides fairness in terms of quality of experience (QoE) for satisfied users and quality of service (QoS) for unsatisfied users. In this paper, we present detailed comparisons relevant to service providers to show the advantages of the proposed algorithm over the popular max-min algorithm. Our algorithm is designed to provide service providers with a mechanism to minimize the number of end-user terminations of service, which is one of the most desired factors for service providers.
Conference Paper
Full-text available
Internet of Things (IoT) envisages a world of connected devices paired with automation systems that aid everyone in their daily activities and communities at large. One of the most prominent use cases of IoT is emergency services. Based on the lessons learned from the tragic Grenfell Tower incident in the UK, we designed an IoT framework that aims at supporting rescue services to quickly locate and identify victims in an event of an emergency by using a range of heterogeneous smart systems. Based on the framework design, a prototype is developed using an RFID access control and a drone system. We also incorporate design specifications that minimize the risk of data and privacy breaches.
Full-text available
User-centric service and application management focuses on the Quality of Experience (QoE) as perceived by the end user. Thereby, the goal is to maximize QoE while ensuring fairness among users, e.g., for resource allocation and scheduling in shared systems. Although the literature suggests to consider consequently QoE fairness, there is currently no accepted definition of QoE fairness. The contribution of this paper is the definition of a generic QoE fairness index F which has desirable key properties as well as the rationale behind it. By using examples and a measurement study involving multiple users downloading web content over a bottleneck link, we differentiate the proposed index from QoS fairness and the widely used Jain’s fairness index. Based on results, we argue that neither QoS fairness nor Jain’s fairness index meet all of the desirable QoE-relevant properties which are met by F. Consequently, the proposed index F may be used to compare QoE fairness across systems and applications, thus serving as a benchmark for QoE management mechanisms and system optimization.
This paper presents a novel adaptation logic for HTTP adaptive streaming (HAS), which achieves not only a high quality of experience (QoE) but also high QoE fairness among independent and heterogeneous clients. The algorithm forces video clients to adapt the requested quality level based on the current network conditions and their individual bit rate requirements, such that the overall quality levels selected by all currently active streaming clients are fairly distributed, i.e., they do not diverge too much. The design of the algorithm is inspired by the well-known transmission control protocol (TCP) congestion control, and drives heterogeneous clients to independently converge on similar quality levels without the need for communicating with each other and/or with a centralized controller in the network. By defining quality levels with equal visual quality, and preparing video representations accordingly, the quality level fairness is extended to QoE fairness. In this paper, the design of the TCP-inspired adaptation logic (TCPAL) is described and a simulative performance evaluation is conducted to compare the QoE and QoE fairness of the proposed algorithm with other HAS adaptation logics. TCPAL is evaluated both in scenarios with stable and fluctuating streaming capacity, and the impact of its parameters is explored. The results suggest that TCPAL performs on par with other HAS adaptation logics in terms of QoE and QoE fairness for low link capacities, but significantly improves the QoE fairness for increased link capacity. Moreover, the fairness achieved by TCPAL does not degrade in situations with fluctuating streaming capacity.
New behavioural and neuroscientific evidence on the development of fairness behaviours demonstrates that the signatures of human fairness can be traced into childhood. Children make sacrifices for fairness (1) when they have less than others, (2) when others have been unfair and (3) when they have more than others. The latter two responses mark a critical departure from what is observed in other species because they enable fairness to be upheld even when doing so goes against self-interest. This new work can be fruitfully combined with insights from cognitive neuroscience to understand the mechanisms of developmental change.