63 by linxiaoqin


									               6th International Quality Conference
               June 08th 2012
               Center for Quality, Faculty of Engineering, University of Kragujevac

    Andreja Samčović
                                            ASSURANCE OF QUALITY OF
                                           EXPERIENCE AS A FRAMEWORK
  Faculty of Transport and                  TOWARD NEXT-GENERATION
    Traffic Engineering,                           NETWORKS
   University of Belgrade,
           Serbia                    Abstract: This article aims to invoke assurance of
    andrej@sf.bg.ac.rs               Quality of Experience (QoE) as a framework toward
                                     next-generation networks. QoE can be viewed as the
                                     end-user’s perception of the quality. The goal of this
                                     paper is to present some major problems of the today’s
                                     telecommunications networks and how they can be
                                     solved by the techniques related to QoE.
                                     Keywords: Telecommunications networks, Quality of
                                     Service (QoS), Quality of Experience (QoE), Video

    1. INTRODUCTION                                    equipment, data encoding, protocols,
                                                       terminals, etc. Each of the end-user
     Assurance of quality of experience                requirements     translates   to various
(QoE), that is, service quality as                     technological and business challenges.
subjectively perceived by the user, is an              Accordingly, services must be provided
important challenge facing network                     over any medium and networking
operators and service providers in current             technology, and by any operator.
telecommunication                 networks.                 It is expected that a variety of
Convergence between fixed and wireless                 services, from low demanding to real-time
networks as well as within wireless                    broadband, will be delivered to the end-
systems using different technologies                   user regardless of the type of access
makes it possible to use a large variety of            network, or end user device. Services must
applications that include surfing the web,             be provided in a multi-domain and multi-
sending emails, listening to music,                    operator environment. As a result, various
watching movies, playing games, or GPS                 technological challenges must be faced.
(Global Positioning System) navigation,                Moreover, users expect to be able to use a
on a variety of terminals located in                   given service continuously while on the
different geographical       environments.             move without a noticeable deterioration of
Customers      require     high      quality           service quality.
telecommunication services regardless of                    This paper is organized in the
localization and time constrains. In other             following way: in Section 2 the major
words, they want to receive any service,               problems of today’s networks are given, in
anytime, anywhere, and on any device.                  Section 3 several definitions of QoE are
     In    general,   these four       user            presented, in Section 4 some techniques
requirements motivate the need for                     related to QoE are selected. Section 5
convergence from the user point of view.               presents QoE with respect to convergence
From the network point of view, the                    requirements, while conclusions are
situation is more complex. The user                    presented in Section 6.
perception of the quality is influenced by
several elements associated with end-to-
end service delivery, namely: network,

                                      6th IQC June, 8 2012                                   575
      2. MAJOR PROBLEMS OF                       overprovisioning, i.e. increase network
      TODAY’S NETWORKS                           efficiency. Currently, stable services are
                                                 assured due to greatly overdimensioned
     Network operators are confronted with       networks. These networks operate at their
massive network traffic increases while          capacity limit in peak times only. Most of
seeking to reduce investment and                 the times resources are unused which is
operating costs for their networks. New          cost intensive and leads to unnecessary
service offerings such as video streaming        high environmental load.
and personalized services led to a steep              Fibre-to-the-Home (FTTH) or Fibre-
increase of network traffic [1]. Five main       to-the-Cabinet (FTTCab) roll-outs will
trends can be identified that require more       increase network capacity greatly.
intelligent, adaptive network management         However, these networks require massive
mechanisms [2]:                                  capital expenditures into the infrastructure
1. Rich media consumption. The increasing        while postponing the impending problem
availability of IPTV (Internet Protocol          only. Additionally, fibre networks do not
TeleVision), leads to surges in network          lead to efficiency increases. To the
traffic. Especially in the early evening         contrary, at the beginning the degree of
hours high network traffic peaks can be          overprovisioning          will      increase
observed.                                        significantly instead of being reduced [3].
2. Service personalization. Besides traffic      Thus, other possible solutions for these
neutral service personalization, e.g.            challenges such as Quality of Service
personal settings in web platforms, other        (QoS) and Quality of Experience (QoE)
personalization can induce changes in the        are currently subject to research.
network. The latter is valid for Video-on-
Demand (VoD) platforms such as
Google’s       YouTube.       Instead     of          3. DEFINITIONS OF
broadcasting linear television without                QUALITY OF EXPERIENCE
responses from users, VoD services
require dedicated connections – unicasts –            Especially quality sensible services
to each user.                                    require high-speed broadband Internet
3. Time, place and device sovereignty.           connections with real-time, interactivity,
Smartphones and tablet computers pave            security and reliability capabilities.
the way for independent media offerings of       Whereas the term Quality-of-Service is not
the future. These will allow watching any        used consistently in the literature, it
video content at any time on any different       usually implies the possibility to
devices. To realize such services, content       differentiate individual services and the
needs to be streamed over unicasts in fixed      possibility to allocate different quality
and mobile networks.                             parameters to services [4]. Technically,
4. Quality expectations. After years of          usually four parameters are used to
low-quality video offerings in the web –         determine the quality of a data connection:
mostly due to poor Internet connections –        the available bandwidth, delay time, jitter
end users are becoming increasingly              and packet loss. With these parameters,
sensible to quality issues. Especially IPTV      different service classes or priority levels
offerings need to maintain a perceived           can be created, ranging from level 0 called
quality level similar to that of other           “best-effort”, to level 7 called “layer 2
television transmission technologies to          network control reserved traffic”, with
succeed.                                         latency and jitter less than 10 ms.
5. Efficiency increases. The network             However, QoS does not address the
operators’ wish to decrease the degree of        subjective end user perception of quality

576                                     A. Samčović
that is harder to measure [5].                        4. TECHNIQUES RELATED
     The notion of Quality-of-Experience              TO QUALITY OF
(QoE) is more user-centered than QoS. It              EXPERIENCE
aims at linking together the technical
parameters described above and the users’              Several research projects address
perception of quality. Several definitions        QoE, ranging from systematic QoE
of QoE exist:                                     measurements to the development of a set
• Mostly used is the definition of ITU-T          of technologies that aim to improve
(International Telecommunication Union)           different aspects from the network to the
SG12 that describes QoE as “overall               service layer. The latter found a multitude
acceptability of an application or service,       of possibilities to improve the perceived
as perceived subjectively by the end user”        quality. The following six selected
that ”may be influenced by user                   techniques show the range of possibilities
expectations and context” [6].                    that exist exemplarily:
• Lopez et al. describe QoE as “extension         1. Monitoring and traffic estimation
of the traditional QoS in the sense that          mechanisms.        Allow     forecasts    of
QoE provides information regarding the            congestion situations and triggering
delivered services from an end-user point         adequate reactions to congestion problems
of view” [7].                                     at occurrence.
• Soldani et al. define QoE as “how a user        2. Scalable video. Can be used in at least
perceives the usability of a service when in      two cases. First, the variety of end user
use – how satisfied he/she is with a service      devices can be served with the correct
in terms of, e.g., usability, accessibility,      resolution, minimizing CPU load on the
retainability and integrity”[8].                  devices. Second, downscaling of video in
• And rather recently Fiedler et al. defined      case of traffic peaks allows continuation of
QoE as a concept that describes “the              streaming instead of complete failures.
degree of delight of the user of a service,       3. Routing, notification and admission
influenced by content, network, device,           control mechanisms. Increase network
application, user expectations and goals,         efficiency by optimizing link usage,
and context of use” [9].                          provide technical solutions to trigger
     All definitions except for the very          reactions in case of service failure, allow
broad one by Lopez et al. have in common          notifying end users about current and
that quality levels are defined by the user’s     estimated problems.
perception in addition to measurable              4. Caching. Caching within the access
network parameters. The user’s perception         network, often referred to as microcaching,
may be influenced by the network, the             allows answering similar requests fast and
context and his/her expectations. The             without causing traffic in higher network
extended set of influencing factors can be        aggregation levels.
addressed on very different levels. In the        5. Video streaming based on Mean
context of QoS, network improvements              Opinion Scores (MOS). Studies found that
were mostly developed on the lower OSI            the Mean Opinion Score fluctuates
(Open Systems Interconnection) levels to          depending on the kind of the movie despite
improve and control the QoS service               of the same bit rate, resolution, etc. By
parameters. When taking into account user         implication this means that the perceived
perceptions, improvements need to be              quality on a certain level can be achieved
realized on higher levels as well, i.e.           with      different    video     parameters,
optimizations up to OSI layer 7 – the             potentially allowing either improving or
service layer – need to be addressed. Table       economizing video streaming services.
1 summarizes the three concepts.

                                     6th IQC June, 8 2012                                577
6. Policy-based EPGs. Electronic Program        also depends on many orthogonal factors,
Guides (EPG) can be improved based on           as the end-user’s device, the environment
manually or automatically generated user        in which the service is received, and the
policies. These can be created based on         type of service, which is shown in Figure
previous user behavior, manually selected       1. The orthogonal factors are especially
preferences etc.                                important for voice and video services. For
     Summarized, QoE improvements are           instance, voice quality assessment may be
technically possible. It also shows that        affected by background noise, type of
most QoE improving technologies need to         equipment (headphones, speakers), and
be implemented or supported by the              type of content (music, news, telephone
network. Thus, the network operators need       conversation). The rating of video quality
to adapt their network accordingly.             is influenced by screen llumination and
Whereas the different approaches promise        size (e.g., mobile phone or PDA (Personal
to increase customer satisfaction and           Digital Assistant) screen outdoor vs. large
increase network efficiency these benefits      LCD (Liquid Cristal Display) TV at
cannot be expected to cover capital             home), viewing distance, and content
expenditures and operating costs for the        (video call, action movie, ‘‘talking
network operator. Thus, new revenue             heads’’). User evaluation of a service is
streams are necessary to incentivize the        also affected by subjective factors
adoption of QoE technologies by network         including psychological and sociological
operators.                                      aspects such as emotions, expectations and
                                                experience with similar services, opinions
                                                of others, etc. In general, a service with the
      5. QoE WITH RESPECT TO                    same       network-level       QoS/GoS/QoR
      CONVERGENCE                               support might be experienced differently
                                                by a user depending on several factors
      REQUIREMENTS                              [10].
                                                     The above mentioned components
     QoS, GoS (Grade of Service) and
                                                influencing QoE, except for the
QoR (Quality of Resilience) describe
                                                psychological and sociological ones are
various intrinsic characteristics of a
                                                more or less related to network
network while customers’ satisfaction with
using services is commonly described as
                                                     To ensure that customers receive a
QoE. It strongly depends on intrinsic
                                                service with a high QoE – that is they are
network features and performance,
                                                satisfied with the service – all factors
although this relation is not always
                                                influencing QoE must be taken into
straightforward. There is no simple
                                                account. QoE expectations related to
mapping        between      QoS/GoS/QoR
                                                different applications and services translate
parameters and QoE. These relations can
                                                into differentiated QoS, GoS, and QoR
be better understood only in the context of
                                                performance offered by the network. There
particular applications. However, efforts
                                                are no mechanisms in place provisioning
are ongoing towards finding mathematical
                                                QoE directly. Instead, in order to achieve a
relationships between QoS parameters and
                                                desired QoE level, QoS/GoS/QoR
QoE, expressed quantitatively by the mean
                                                provisioning mechanisms must be selected
opinion score (MOS) value. It must be
                                                and designed appropriately. The complete
noted that such formulas are derived for a
                                                set of solutions necessary to meet various
particular application under several
                                                QoE requirements in converged networks
assumptions. Provisioning of QoS, GoS
                                                should also address issues related to
and QoR at an appropriate level is crucial
                                                several      orthogonal      factors,     e.g.,
for achieving high QoE, however, QoE

578                                    A. Samčović
capabilities of user terminals.                     including both technical and user aspects.
                                                    Those two domains need to be brought
                                                    together, and concepts as well as methods
    6. CONCLUSION                                   need to be combined in order to fully
                                                    understand and improve a product’s Quality
     In this paper, assurance of Quality of         of Experience. The integrated framework
                                                    presented in this paper provides a detailed
Experience (QoE) as a framework toward
                                                    look at the different techniques related to the
next-generation networks is studied. In
                                                    QoE, from both technical as well as user
today’s ICT-environment, it is no longer
sufficient to measure only ‘technological’
                                                         Growing user expectations related to
performance or Quality of Service (QoS)
since it is not the final goal anymore. The
                                                    perceived quality of services accessible
central goal should be to deliver high Quality      anywhere, anytime, on any user device and
of Experience (QoE) to the user. During the         on any media are setting new challenges
development of new systems and                      for technology developers and service
applications, it will be crucial to gain            providers. These expectations can only be
adequate insight in the user’s expectations         fulfilled if an appropriate set of metrics
regarding Quality of Experience, its different      reflecting quality of experience is defined
components and its relation with technical          and interoperability between different
performance metrics.                                converging networks is assured.
     In order to achieve this, a
multidisciplinary approach is called for,


[1] Casier, K. et al., "Adoption and pricing: the underestimated elements of a realistic IPTV
     business case", IEEE Communications Magazine, Vol.46, No.8 (2008) 112-118.
[2] Latre, S. et al, "Online estimation of the QoE of progressive download services in
     multimedia access networks", Proceedings of the International Conference on Internet
     Computing, 2008.
[3] Degrande, N., "Increasing the user perceived quality for IPTV services", IEEE
     Communications Magazine, Vol.46, No.2 (2008) 94-100.
[4] International Telecommunication           Union,    ITU-T     Recommendation        G.1000,
     Communications quality of service: a framework and definitions, 2001.
[5] Meddeb, A., "Internet QoS: pieces of the puzzle", IEEE Communications Magazine,
     Vol.48, No.2 (2010) 86-94.
[6] International Telecommunication Union, ITU-T Recommendation G.1081, 2007.
[7] Lopez, D. et al, "Adaptive multimedia streaming over IP based on customer oriented
     metrics", International Symposium on Computer Networks, 2006.
[8] Soldani, D., Li, M., Cuny, R., QoS and QoE Management in UMTS Cellular Systems,
     Wiley, 2006.
[9] Fiedler, M., Kilkki K., Reichl, P., From Quality of Service to Quality of Experience,
     Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik, Germany, 2009.
[10] Scholda, P. et al., "Quality of resilience as a network reliability characterization tool",
     IEEE Network, Vol.23, No.2 (2009) 11-19.

                                      6th IQC June, 8 2012                                    579
               Table 1 – Quality concepts, potential improvements and measures

               Concept           Description                   Realized on         Measures
                                                               OSI Layers

               QoE               Extension       of    QoS     Layer 1-7           Network
                                 understanding with user                           Context
                                 perceptions,        quality                       Usability
                                 optimization up to the                            User expectations
                                 service level
               QoS               Classification into quality   Layer 1-4           Bandwidth
                                 classes based on                                  Delay
                                 measurable                                        Jitter
                                 parameters, pricing                               Packet loss
                                 to quality classes, quality
                                 optimizations on the
                                 network level

               Best effort       Basic availability of         Layer 1-7           Bandwidth (no
               Internet          Internet                                          assurances)
                                 connectivity and services

      Enviromental,      User profile        specific             Pricing policy
      Psychological                          features                              Terminals,
      Sociological                                                                 coders

                                                     QoE                           Type of

           QoS                                    Grade of                         Quality of
                                                  Service                          Resilience

                             Figure 1. Different factors influencing QoE

580                                          A. Samčović

To top