Professional Documents
Culture Documents
Hyun-Jong Kim - QoS - QoE Correlation Model - 2008 PDF
Hyun-Jong Kim - QoS - QoE Correlation Model - 2008 PDF
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
research status of these two kinds of evaluation 2.2. The study of relationship between QoS
methods are illustrated as followed paragraphs: and QoE
Currently, the evaluation methods for the speech
service are mature. For subjective evaluation methods, Currently, the investigation of QoS and QoE
opinion rating (MOS) based on customer’s satisfaction correlation is continued. Khirman and Henriksen were
has been studied to assess the perceptual QoS. It is trying to relate the objective network service conditions
specified in ITU-T recommendations E.800 initially [1]. with the human perception of the quality of the service.
On another hand, several objective quality assessed Their subject has been widely investigated for voice
methods has been proposed in ITU-T, such as P.861[2] delivery and it is widely acknowledged that the
PSQM (Perceptual Speech Quality Measure), P.862[3] relationship between voice transmission conditions and
PESQ (Perceptual Evaluation of Speech Quality) and the human perception of quality is far from linear[7].
G.107 E-Model[4]. They discuss in detail how the human satisfaction of
For the video service evaluation, subjective video HTTP service is affected by the two main network QoS
quality evaluation method is the most reliable video parameters, namely network delivery speed and latency.
quality measurement method. A group of viewers is However, it is difficult to represent the feature of the
selected and gathered in a room, the measurement provided and various services from only the bandwidth
environment is specified in the ITU-T Recommend- and latency time in the integrated network environment.
ation P.910[5]. For the research of objective video In [8] the authors thought that pervasive computing
quality method, some estimation software has been environment brings the method of evidence context
developed which can analyze the video signals and related to QoE. They studied the QoE evaluation
produce the quality evaluation results. One traditional method in pervasive computing environment, and
objective video quality measurement, Peak Signal to proposed the enhanced QoE evaluation parameter
Noise Ratio (PSNR), has been widely used in many model. In [8] rough-set based algorithm is proposed to
applications to assess video quality. reduce context attributes and determine the weight of
PSNR does not take the visual masking phenomenon each attribute, the algorithm has been validated on
into consideration. In other words, every single pixel video streaming service, and the architecture of QoE
error contributes to the decrease of the PSNR, even if evaluation system is described. As a mass of evidence
information related to the experience of users can be
this error is not perceived. So, MPQM (Moving
gathered through the context-awareness computing, the
Pictures Quality Metric) was proposed for the objective
calculation results of QoE evaluation method can
the video quality measurement[6][10]. MPQM is an
highly match the real feeling of users. However, the
objective quality metric for moving picture which method needs to be enhanced along with the
incorporates human vision characteristics. MPQM development of pervasive computing.
represents the typical image quality assessment models
based on the error sensitivity. The widely adopted
assumption of these models is that the loss of 3. The QoS parameters related with end-
perceptual quality is directly related to the visibility of
user
the error signal.
From current research status of evaluation method,
This clause addresses the QoS quality parameters
we can see that the subjective method based on user
which can be considered for the QoE evaluation.
survey can reflect the experience of user more directly
Transfer Capacity is a fundamental QoS parameter
and match well to the feeling of user. However, this
kind of method has several problems, such as, it having primary influence on the performance perceived
required special environment and equipments, needs a by end-users. Many user applications have minimum
mass of people to participate the test. In conclusion, capacity requirements; these requirements should be
subjective video quality measurement cannot provide considered when entering into service agreements. And
real-time and in-service quality monitoring for real- lost bits or octets can be subtracted from the total sent
time video applications. So the application of the in order to provisionally determine network capacity.
method is limited. An independent definition of capacity is for further
study.
It is assumed that the user and network provider
have agreed on the maximum access capacity that will
720
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
be available to one or more packet flows in a specific 3.1. Delay
QoS class (except the Unspecified class in Table Ⅰ). A
Delay manifests itself in a number of ways,
packet flow is the traffic associated with a given
including the time taken to establish a particular service
connection or connectionless stream having the same
from the initial user request and the time to receive
source host (SRC), destination host (DST), class of specific information once the service is established.
service, and session identification. Other documents Delay has a very direct impact on user satisfaction
may use the terms microflow or subflow when referring depending on the application, and includes delays in
to traffic streams with this degree of classification. the terminal, network, and any servers. Note that from a
Initially, the agreeing parties may use whatever user point of view, delay also takes into account the
capacity specifications they consider appropriate, so effect of other network parameters such as throughput.
long as they allow both network provider enforcement
and user verification. For example, specifying the peak 3.2. Delay variation
bit rate on an access link (including lower layer
overhead) may be sufficient. The network provider Delay variation is generally included as a
agrees to transfer packets at the specified capacity in performance parameter since it is very important at the
accordance with the agreed QoS class. transport layer in packetized data systems due to the
The network performance objectives may no longer inherent variability in arrival times of individual
be applicable when there is packets submitted n excess packets. However, services that are highly intolerant of
of the capacity agreement or the negotiated traffic delay variation will usually take steps to remove (or at
contract. If excess packets are observed, the network is least significantly reduce) the delay variation by means
allowed to discard a number of packets equal to the of buffering, effectively eliminating delay variation as
number of excess packets. Such discarded packets must perceived at the user level (although at the expense of
not be included in the population of interest, which is adding additional fixed delay).
the set of packets evaluated using the network
performance parameters. In particular, discarded 3.3. Information loss
packets must not be counted as lost packets in assessing
Information loss has a very direct effect on the
the network's IPLR performance. A discarded packet
quality of the information finally presented to the user,
might be retransmitted, but then it must be considered
whether it be voice, image, video or data. In this
as a new packet in assessing network performance. context, information loss is not limited to the effects of
Each network QoS class creates a specific bit errors or packet loss during transmission, but also
combination of bounds on the performance values. This includes the effects of any degradation introduced by
clause includes guidance as to when each network QoS media coding for more efficient transmission (e.g. the
class might be used, but it does not mandate the use of use of low bit-rate speech codecs for voice).
any particular network QoS class in any particular
context. Table 2. Guidance for IP QoS classes [12]
Table 1. IP network QoS class definitions and QoS
network performance objectives [12] Applications (ex) Network techniques
class
Real-time, jitter sensitive,
Network QoS Classes Constrained routing and
0 high interaction (VoIP,
performance Class Class Class Class Class Class distance
VTC)
Parameter 0 1 2 3 4 5 Real-time, jitter sensitive, Less constrained routing
100 400 100 400 1
IPTD 1s U interactive (VoIP, VTC) and distances
ms ms ms ms Transaction data, Highly Constrained routing and
IPDV 50ms 50ms U U U U 2
interactive (Signalling) distance
1x 1x 1x 1x 1x Transaction data, Less constrained routing
IPLR U 3
10-3 10-3 10-3 10-3 10-3 interactive and distances
-4
IPER 1x 10 U Low loss only (short
4 transactions, bulk data, Any route/path
video streaming)
Traditional applications of
5 Any route/path
default IP networks
721
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
calculated by the equation (1). The QoE class measured
4. The proposed QoE measurement scheme by the QoS quality parameter of a network-level is
using the QoS parameters mapped like the existing MOS grade with 5 classes.
The α represents the QoS quality class of the network-
The provided QoS-QoE correlation of the service level which has to be provided as the QoS class
can variously show up in the integrated network constant in order to satisfy the required QoE class. The
environment by QoS quality parameters considered for minimum range of the QoE quality standard of the
a kind of service and QoE. We propose the correlation provided service is determined by this value.
model approach method that can reflect all elements Next, β is determined according to the class of
influencing on QoE in order to solve these problem. service as the grade of service constant. Finally, K is
The relation between the QoE and QoS cannot be the Scale constant for showing the satisfaction about
easily mapped to the service which can be provided in the use service. This generalizes the subjective quality
the converged network. According to the traffic about the using service. Although the same QoS quality
is provided in the wire-network environment, this
characteristic of each service, the QoS level which is
variable considers elements influencing on the
required in order to satisfy the QoE class is different.
user experience quality according to the radio channel
Moreover, the QoE influence element at the terminal
and terminal. According to the class of service, the next
layer which is required in order to evaluate QoE need paragraph shows the QoE measurement scheme that
to reflect. The normalized QoS score of the service uses QoE and QoS correlation model.
which can be provided can be obtained through the
next equation from the integrated network. 4.1. The Guaranteed service
Here, the used variable and constant are determined The most of guaranteed services is the real-time
by elements influencing on QoE including an voice / video service like the VoIP, Video Phone and
environment and the using service kind, terminal Video Conference service. The stable bandwidth has to
position, using codec, and etc. The QoS of this be provided for the seamless usage of this service and
equation is the normalized QoS score which is the minimum connection delay time has to be
guaranteed. The quality parameter needed for the
722
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
quality measure of this service is delay, jitter and Guaranteed service and this can be controlled by the
packet Loss rate, and etc. And the call connection yield, grade of service constant (B). The premium service can
and the use of service duration and audio / image MOS divide the quality of service according to this linearity.
value are needed when the quality measure of this This accommodates the various class of the AF PHB of
service considers QoE. DiffServ. According to select the appropriate QoE
The QoS-QoE correlation model of the Guaranteed class constant, a network and service providers can
service can represent the on / off model. If the organize the various premium service QoS-QoE
QoS score which is required in order to guarantee the correlation models.
QoE of this service is unable to be satisfied, the
Guaranteed service cannot be provided. As shown in 4.3. The Best Effort service
Figure 1, the QoS score has to be provided over the
minimum 50 in order to guarantee the QoE of the The best-effort service can include preexistence
Guaranteed service. As to this model, the β-value is internet services and the quality level is similar to it.
small compared with the other model and this means This says to be the more sensitive QoS-QoE correlation
that the minimum guaranteed class of the QoS which model the QoS parameters than QoE. This means that
has to be supported is very sensitive. although the QoS parameter performance is improved,
there is a limit in the QoE grade improvement which
Here, the shown numerical value is one example and
the end user experiences.
actually evaluates the quality customer satisfaction
measurement about the service user and need to
calculate the average distribution of QoE in order to QoE
5
establish the exact relation.
4
In the case of the premium service, it is more flexible 2 Best Effort Service
723
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
Quality VSWR(Voltage Standing
5. The example of the QoE measurement Indicator Wave Ratio),
using the QoS-QoE correlation model (CQI) CINR (Carrier to
Interference plus Noise
The process of measuring the QoE of the Premium Ratio)
service by using the QoS-QoE correlation model is as
follows. The Table 1 shows the weighted value of the
each QoS parameters needed to the QoE evaluation of The QoS normalized by the equation (1) is
the Premium service. substituted for the equation (2), and then the QoE class
If quality data measured about the video conference can be evaluated. As shown in Figure 4, we can predict
service of the Premium service are as follows, we can e the class of Premium service B for
xpect the QoE of a user through the QoS-QoE correlati the experience quality of an end-user in the
on model. above environment. We can confirm whether it is
satisfied the class of service which the first user joins
from this measurement result or not.
• Delay (D) = 125ms Because of being the result that the QoE is
• Jitter (J) = 45ms evaluated by using the objective network quality
• Packet loss rate (L) = 2 x 10-5 parameters, it can be regarded as the more objective
QoE than the other measurement schemes.
• Packet error rate (E) = 1 x 10-6
Here, the used in order to digitize the QoE class
• Bandwidth (B) = 2.5Mbps network quality index and the standards aren't perfect,
• Call success rate(S) = 99.0% but are one example. The investigation of the weighted
• QoS = F(D, J, L, E, B, S) = 8+10+7+10+10+8=53 value about the network performance index has to be
separately executed. The proper quality parameter
Table 3. QoS parameter and range related with score need to be selected against service users through
the question survey.
QoE
Call success
100 ~ 99.9% 10 Because existing QoE measurement schemes
99.9% ~ 99.0% 9 Undefined measure QoE by using a part QoS measurement
rate (S)
99.0% ~ 98.0% 8
parameters, they are difficult to reflect of the various
HO within the cell N/A
services. The quality measurement scheme that uses the
Handover
HO between the cells N/A Undefined preexistence QoS measurement parameter can provide
(H)
HO between the frequencies N/A
the objective quality information in a network-level.
However, this quality measurement scheme is difficult
Channel RF Power, RSSI, N/A Undefined
724
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.
to expect the quality which the service end-user
experiences. So, in this paper, by using the quality of [3] ITU-T P.862, “Perceptual evaluation of speech quality
(PESQ),” Feb 2001.
service information measured in a network-level, we
explain the proposed QoS-QoE correlation model for [4] ITU-T G.107, “The E-model, a computational model for
the objective QoE, and describe the example of the use in transmission planning,” Mar 2005.
service QoE evaluation.
In the future, the research for setting up the [5] ITU-T P.910, “Subjective video quality assessment
quantified QoS class toward QoS parameter of the methods for multimedia applications,” Sep 1999.
various network-levels in the integrated network
[6] Wang, Y. “Survey of Objective Video Quality
environment is needed. Moreover, the research about
Measurements,” Tech report, Worcester Polytechnic Institute,
which how will set up the value of α, β and K in June 2006.
consideration of the real time multimedia service kind
and service usage environment is required. [7] Stas Khirman, Peter Henriksen, “Relationship between
Quality-of-Service and Quality-of-Experience for Public
Internet Service,” PAM 2002, March 2002.
Acknowledgment
[8] Liu Li-yuan, Zhou Wen-an, Song Jun-de, “The Research
of Quality of Experience Evaluation Method in Pervasive
This work was supported by the IT R&D program Computing Environment,” PCA 2006, Aug 2006.
of MKE/IITA. [2008-S-006-01, Development of Open-
IPTV (IPTV2.0) Technologies for Wired and Wireless [9] Hideaki YAMADA, Norihiro FUKUMOTO, Manabu
Networks] ISOMURA, “A QoE based service control scheme for RACF
in IP –based FMC networks,” CEC/EEE 2007, July 2007.
*Corresponding Author: Seong Gon Choi (sgchoi@cbnu.ac.kr) [10] NetPredic, Inc. White Paper, “Performance Analysis for
Video Streams across Networks,” December 2003.
725
Authorized licensed use limited to: UNIVERSIDAD POLITECNICA DE VALENCIA. Downloaded on July 15,2010 at 23:14:43 UTC from IEEE Xplore. Restrictions apply.