Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

2013 IEEE 20th International Conference on Web Services

Evaluating Quality of Web Services: a Short Survey

Olga Kondratyeva1, Ana Cavalli Natalia Kushik, Nina Yevtushenko


Department of Software Networks Department of Radiophysics
TELECOM SudParis Tomsk State University
Evry, France Tomsk, Russia
{olga.kondratyeva, ana.cavalli}@it-sudparis.eu ngkushik@gmail.com, ninayevtushenko@yahoo.com

Abstract — This paper presents a short survey on the quality authors have collected the information about existing web
evaluation of web services. The most popular metrics for services and their QoS [see, for example, 1]. A special data
estimating such quality and user perception of web services are set that could be used by people from QoS evaluating
Quality of Service (QoS) and Quality of Experience (QoE), community for verifying their methods and techniques
which represent objective and subjective assessments against counterexamples (benchmarks) is presented in the
correspondingly. For different types of web services, the values database [7]. A number of tools [see, for example, 11-14]
of QoS and QoE are measured in different ways. In this paper, were developed for the QoS evalution.
we consider various definitions of QoS based on web service
parameters and describe several methods for evaluating QoS
and QoE. We start with experimental evaluation of QoS based
201
on network traffic analysis and further turn to model based
methods for QoS estimating. Existing relationships between Number of publications
different kinds of service quality evaluation are also discussed. 151

Keywords - web service; quality of service; trace models; QoS


101
attribute/parameter; composite service quality. QoE

I. INTRODUCTION 51

Nowadays web services are widely used in various


applications such as online multimedia services, booking 1
systems, online stores, etc. [1]. Most of web services are 1996 1998 2000 2002 2004 2006 2008 2010
Year
distributed for free via Internet and the quality of such Fig. 1. Number of publications related to QoS and QoE of web services
services especially involved in critical systems [2] has to be
studied thoroughly. Since web services use the so-called In most papers the quality of service is defined on a set of
service-oriented architecture (SOA), which is highly attributes (or parameters) [6, 7, 15, 16] where these attributes
heterogeneous and platform-independent [3], the estimation mostly come from the traffic analysis. As it is mentioned in
of their quality is rather complicated against ordinary web [15], the major attributes to define the QoS are the response
applications. The functionality of web services is delay, the package loss percentage, the service access facility
scrupulously tested by other methods [4] and for this reason, (the percentage when service is available), reliability (the
the functionality of web services is not considered when number of successful operations per certain time period), etc.
estimating their quality. For estimating the quality of web services based on the
In order to track the growing popularity of the topic we above parameters heuristic methods are usually proposed
studied a number of digital libraries including IEEE XPlore, [6, 17]. A usual way when evaluating the QoS based on
ACM Portal, Elsevier Online Library and Springer Online traffic analysis is to use linear combinations of weighted
Library. The obtained results for one of such libraries (the network parameters [16]. Sometimes coefficients of the
IEEE Xplore [5]) are shown in Fig. 1. One may notice that formula are not given as they can be a know-how of a
first, the popularity of the topic in 2010 has been increased company that evaluates the service quality for its purposes.
almost 40 times compared with 2000 and secondly, the For many services that are located in repositories the QoS is
number of research devoted to the Quality of Experience evaluated based on parameters mostly related to traffic
(QoE) is less than that devoted to the Quality of Service analysis. However, the parameters used in traffic analysis are
(QoS). Other metrics for evaluating the quality of web more concerned about the transport level than about service
services are not taken into account, since QoS and QoE still issues; thus, new parameters appear that have to be
remain prevailing metrics by now [6]. considered when evaluating the service quality. Such
Additionally, the importance of the topic is confirmed by parameters may be the amount of money to be spent, service
a number of existing and developing repositories of web reputation, the comfort of a solution proposed by the service,
services where the QoS is somehow calculated [see, for etc. A single utility function that maps the values of all
example, 7-10]. There exist several publications where the parameters into a single resulting value still is used for QoS
___________________________________________________
1
– On leave from Tomsk State University

978-0-7695-5025-1/13 $26.00 © 2013 IEEE 587


DOI 10.1109/ICWS.2013.83
evaluation [18] but usually it is difficult to represent all these interfaces that are defined and described using XML-based
parameters by means of a single value. Correspondingly, the languages. The service description should contain the
service quality is represented as a set or a vector of description of service semantics and a machine-processable
heterogeneous attributes [19, 20]. The parameters which are description of the messages that are processed by the service
related to the traffic analysis are usually evaluated based on [30]. One of the basic standard languages is the web service
the traffic monitoring but such parameters as the comfort of a description language (WSDL). A number of service
proposed solution, the number of money to be spent etc. can depositories publish and provide the automatic analysis of
hardly be estimated applying traffic based methods and thus, WSDL service files; for example, in [10], a ‘WSDL
there is a need for more complex models to be involved. Analyzer’ extracts the list of supported operations and a
Taking into account sequences of various user requests, required transport protocol that usually is HTTP. Complex
cancellations etc., web services are often described by the services may be derived as composition of simpler ones, and
sets of permissible sequences of actions [1, 3] and for this special languages for the composite service description are
reason, researchers turn their attention to trace models such developed. Workflows which support the execution logic of
as flow graphs, Markov chains, weighted and probabilistic composite complex services can be described using the
automata, Petri nets, fuzzy logic [21-27], etc. business process execution language (BPEL) that for each
We also mention that in spite of the fact that QoE and component service, defines which messages it gets from and
QoS are known to be most popular metrics for evaluating the sends to other components. The sequences of message
quality of web services, still somehow it may happen that exchange occurred in the system are usually described using
QoE and QoS are not enough to correctly specify the real the web service choreography description language (WS-
quality [28]. Thus, people from Computer Science industry CDL) [31]. Canonical descriptions of services allow an
propose to use other metrics to evaluate the quality of web automatic search for a service with the required
services. For example, A. Moorsel in his paper [28] functionality, though, the information on the quality of a
discusses the so-called Quality of Business (QoBiz) that is selected service is usually not presented.
evaluated in terms of money being received during executed The QoS (Quality of Service) can be defined as a vector
transactions. Another specific metrics is Quality of Design with components which are values of given attributes
(QoD) [29] that might be interpreted as the quality of (parameters), such as delay time, number of packages being
interaction between an end-user and a client application. lost, reputation, corresponding to a given web service. Such
This paper contains a short survey of existing methods vector can be mapped into a single value using an
and tools for evaluating quality of web services. In appropriate computable function [32] and the result of this
particular, we also discuss the quality of a web service function can be an integer, rational, (fuzzy) logic constant,
composition for which a number of basic patterns is etc. The QoE is supposed to represent human satisfaction
considered. We further discuss the relationship between with the service and thus, is highly subjective. Still, there is a
various metrics and how QoS can be used for the QoE lot of research on automatic QoE computing based on the
evaluation. values of QoS attributes [19].
The rest of the paper is organized as follows. Section II Almost all existing methods for QoS evaluation can be
contains preliminaries. Section III is devoted to methods for implicitly divided into two groups. Methods of the first
QoS evaluation. Methods are divided into those which rely group mostly use passive and/or active network traffic
on network traffic analysis and those that use trace models in analysis to deduce some statistics about “good” and “bad”
order to describe web service and to estimate its quality. packages, packages being lost or being delivered to a wrong
Section III also contains the state of the art for evaluating host, etc., and are well established. Such statistics allows
QoS of web service composition while in Section IV, we drawing a conclusion not only about network quality but
briefly discuss the relationship between some known about the user satisfaction with service as well [see, for
metrics/functions which are used for estimating the quality of example, 6]. Methods of the second group use more complex
web service. Section V concludes the paper. formal models that describe the service properties based on
parameters essential for the quality evaluation, and these
II. PRELIMINARIES parameters can be different from those utilized in traffic
analysis. In this paper, we briefly discuss the above types of
Web service can be defined as a composition of web
quality evaluation and classify them w.r.t. to
applications where a server (client) in one application can be
models/techniques they are based on.
turned into a client (server) in another application. A similar
Most services can be considered as composite services
definition is given, for example, in [30], where web service
containing simpler services as components and thus, two
is defined as “a software system designed to support
questions arise: how to evaluate the quality of such
interoperable machine-to-machine interaction over a
composite service (composition analysis) and which
network”. The communication of simple component services
components (with the same functionality) should be selected
may be organized via a remote procedure call (RPC
in order to reach the best quality of their joint work
technology), though, the simple object access protocol
(decomposition/selection problem). In Section III, we briefly
(SOAP) is more often used. In order to support main features
discuss the problem of the composite service QoS
of the heterogeneous service oriented architecture (SOA), the
evaluation.
interaction with services is processed through public

588
III. METHODS AND TOOLS FOR QOS EVALUATIONS Consider the finite state system in Fig. 2 that describes a
vacation planner service. This automaton with slightly
A. QoS evaluation based on traffic analysis changes is taken from [35, 36]. As usual, we use “?” for
For basic (non-composite) web services heuristic input actions while using “!” for output actions.
methods are very popular when evaluating the quality [see,
for example, 15]. Most of these methods are based on !date
?date ?fstat
network traffic analysis. One of the reasons can be that q1 q2
historically the quality was first estimated for networks; the ?date
results then were used for the multimedia quality evaluation. !not reserved
q0 ?exit q4 q3
Correspondingly the same methods were inherited for the
evaluation of the quality of web services. Based on passive !reserved
and active traffic analysis results some statistics are collected
!date
like the number of packages being lost, the maximal time q6 ?hstat
!not reserved q5
delay between transactions, the probability of losing a
package, bandwidth etc. [6] and how these parameters
influence the traffic QoS evaluation. As an example of such Fig. 2. Automaton for vacation planner web service
monitor, one may turn to the paper [6] where the authors
focus on two network parameters. Those are network At the initial state q0 the vacation planner “asks” a user to
delivery speed (bandwidth) and latency and the authors have define preferable dates of traveling and replies at state q1
experimentally shown that end-user experience would be with the dates available for plane tickets. The user enters
better only if the response latency will be below 50mSec his/her opinion (agree/disagree) about the ticket and the
while effective network bandwidth plays a crucial role in the service moves to state q3 where two options are available. If
end-user satisfaction. the travel conditions do not satisfy the user then the system
Traffic network parameters are weighted very precisely moves to state q4 where the user can change the travel dates
for audio and video transmitting systems, since they are and continue or can quit the service via state q7 with the
known to be closely related to the end user satisfaction. As corresponding output exit. If the user accepts the ticket
an example of a metric for estimating the best route for conditions provided by the service at state q3 then the service
delivering a message, one may turn to Cisco protocol IGRP proposes the dates for booking a hotel room, moves to state
[33]. The best route is selected based on integrated metrics q5 and “asks” the user if he/she accepts these hotel
conditions. If the conditions are accepted (not accepted) the
§ K ⋅B · § K5 · output is reserved (not_reserved). In both cases, the system
m = ¨ K1 ⋅ Be + 2 e + K 3 ⋅ Dc ¸ ⋅ ¨¨ ¸¸ , moves to the final state that coincided with the initial state in
© 256 − L ¹ © r + K4 ¹ our example. Taken apart traffic parameters we can evaluate
where K1, K2, K3, K4, K5 are coefficients which have values the quality of this service using a vector with two
between 0 and 1 and are defined by a network administrator, components. The first component is responsible for the
Be is a channel bandwidth calculated as the ratio of the execution time while the second is concerned about the
integer 107 to minimal bandwidth (Kb/sec), L is a traffic service popularity, i.e., the value of this component that is
load with values 1, …, 255, Dc is one tenth of overall delay between 0 and 1 corresponds to the percentage of users who
measured in microseconds, r is a relative reliability, e.g., the succeeded in planning their vacations via this service. Thus,
percentage of successfully transmitted packets via this route. QoS = ¢t, r² where t is a rational corresponded to the
If K5 = 0 then the reliability is not taken into account. By maximum time (in seconds) needed for the query execution
default, K1 = K2 = 1, K3 = K4 = K5 = 0, and Cisco and r is popularity. In other words, if QoS = ¢35, 0.5² then
recommends not to change these parameters. each query is executed in at most 35 seconds and on average,
Similar formulae for the QoS evaluation are used for 50% of travellers who use this service will successfully plan
multimedia services [see for example, 34]. We note that their vacations. Parameters related to the traffic analysis can
coefficients of the formula may not be given in the literature hardly help to evaluate the above values and for this reason,
and sometimes they are considered as know-how of the the researchers are now switching to trace model based
company that evaluates quality of its services. Still, there are evaluation of the quality of web services.
some statistical data available that may help developers and B. Trace models for evaluating QoS of web services
testers to evaluate QoS of services they are using.
An obvious advantage on this approach is that the quality Different formal models are considered when describing
estimation based on the traffic analysis is well established web service or/and evaluating its quality. Taking traffic
and is widely used, especially in the industry. However, analysis apart, these models can be considered as
these approaches which work rather well for evaluating semidiscrete (semicontinuous) trace models, since usually
audio and video transmitting systems become almost useless when describing a web service [4, 21, 25] a state transition
when the parameters of interest are related to the user graph is used where the edges of the graph are weighted,
satisfaction such as the comfort of proposed solutions, the possibly, with probabilistic values. We note that trace models
amount of money to be spent etc. In order to illustrate this are efficiently applied for verification, testing, or automated
fact we consider a simple example. design of web services [35-37], and therefore, integrating

589
quality parameters into such models for evaluating the CS, the question is what is the quality Q of the composition
quality seems to be perspective. Lately, a number of methods of component services. As mentioned above, in this paper,
for the automatic translation of the web service description in we do not consider the functioning of component services
one of the XML-based languages (WSDL, BPEL, WS-CDL, and messages they exchange. Since we are only interested in
etc.) into corresponding formal models are developed task invocations between components, the QoS of a
[31, 38-40]. In [38], for example, the authors develop a tool composite service can be evaluated using a corresponding
for the two-way automatic translation between extended workflow, and in Fig. 3 we consider basic composition
versions of BPEL specifications and timed coloured Petri patterns in which the workflow can be decomposed. When a
net. In other papers [for example, 31, 40] finite state models component service is invoked, it executes some task
are considered. In [40] a push-down automaton is derived for (according to the composition requirements), and after
XML parsing and its application for web services is completing the task, the service either produces the result if
considered. Automata augmented with timed parameters are it is the final task, or invokes other components to execute
in the focus of [31], where it is proposed how to derive a further composition tasks. The same component service can
timed automata from the WS-CDL description of a service be used for executing different composition tasks. To avoid
composition. In [39], proper grammar rules for web service any ambiguity we further refer to the component service
interface verification are generated from WSDL description. when some task is executed by this service.
Despite the fact that in these papers the service quality issues The simplest workflow compositional pattern is
are not explicitly considered, the automatic translation of sequential (Fig. 3a) where the composite service is organized
service descriptions into formal models allows to use these as follows: when service S0 completes a task service S1 is
formal models for the quality evaluation process. invoked. In a conditional pattern (Fig. 3b), also referred to as
When evaluating such attributes of a web service as XOR-split, S0 invokes one and only one of services S1 … Sk
reliability, reputation, popularity etc. probabilistic models depending on the results of the task execution. When
such as Hidden Markov Models [16], probabilistic automata probabilities are involved for each possible invocation the
[26], fuzzy logic [27] etc. can be effectively used. The use of k
most of these models provide an opportunity to conclude that equality ¦ pi = 1 must be held. When service S0 invokes
a given web service is good with a probability p or rather i =0

good than bad and those parameters are also very important several services S1, …, Sk parallel pattern (Fig. 3c), or AND-
as it is illustrated by the above example. The probability can split pattern, is considered. Services executing tasks in
be treated as the ratio of the number of successful events to parallel can be further merged in order to execute a required
the overall number of events (statistical definition), and thus, later task. If the next service is invoked only when all
may be of a good help when calculating relative parameters preceding services have completed their tasks, the
for a service like popularity, reliability etc. synchronizing, or AND-joint, pattern (Fig. 3d) is considered.
Consider a graph in Fig. 2. In order to estimate the time Otherwise, if the next service is invoked after the first
that is needed to process a query we can accordingly weight service completes its task, the concurrent pattern (Fig. 3e), or
edges corresponded to each task and apply the methods used XOR-joint, is at hand. When some tasks should be repeated,
for analyzing a weighted automaton [24]. On the other hand, the loop pattern (Fig. 3f) is considered, and the number of
if we would like to estimate the popularity of the service then repetitions may be either known a priori or can be calculated
transitions should be labeled with corresponding during the task execution. Without loss of generality, we
probabilities that usually are derived due to known statistics. consider loops with a single starting point (service S1 in Fig.
In this case, the methods for analyzing a probabilistic 3f) and any number of exits (service Si in Fig. 3f). For each
automaton [25] will be involved for the quality evaluation. exit point probabilities of continuing the loop or of going out
We note that almost for all models that take probabilities may be specified.
into account the main problem is how the probabilities are
distributed, and thus, some companies/institutions keep those S1 S1
probabilistic parameters as know-how. p1
. .
||
S0 S1 S0 … . S0 … .
C. QoS for web service composition . .
pk
When a new service is derived by composing already Sk Sk
existing web services, possibly developed by a third-party, a b c
the question arises whether existing components can provide
a required composition quality or some component services
S1 S1 …
should be re-designed. As before, we consider the QoS as a Si
vector of values of selected quality attributes such as cost, . . |
. … . … S1
Sk+1 Sk+1
execution time, reliability, etc., and when evaluating the . . …
quality of service composition, it is usually supposed that the Sk Sk
Sk
values of quality attributes for components are given by
service providers or so-called service brokers [e.g. 41]. d e f
Given a set {S1, S2, … , Sn} of component services, their Fig. 3. Basic compositional patterns: (a) sequential, (b) conditional, (c)
QoS vectors {Q1, Q2,…Qn}, and the composition structure parallel, (d) synchronizing, (e) concurrent, (f) loop

590
TABLE I. AGGREGATION FUNCTIONS FOR QOS ATTRIBUTES

Compositional pattern
QoS Attribute Reference
Sequential XOR-split Parallel AND-joint Parallel XOR-joint Loop
k k k k k
Cost (c) ¦ ci ¦ ci ¦ ci ¦ ci n ¦ ci [20, 41]
i =1 i =1 i =1 i =1 i =1
n
k k k k
1 − ∏ (1 − ai ) § k ·
Availability (a) ∏ ai ∏ ai ∏ ai ¨ ∏ ai ¸ [18, 41]
i =1 i =1 i =1 i =1 © i =1 ¹
k k
Execution time, response time (t) ¦ ti max(ti ) max(ti ) min(ti ) n ¦ ti [20, 41]
i =1 i =1
n
k k k k
§ k ·
Reliability, success rate (r) ∏ ri ∏ ri ∏ ri ∏ ri ¨ ∏ ri ¸
[18, 21,
41]
i =1 i =1 i =1 i =1 © i =1 ¹
1 k 1 k 1 k 1 k 1 k
Reputation (q) ¦ qi ¦ qi ¦ qi ¦ qi ¦ qi [18, 21]
k i =1 k i =1 k i =1 k i =1 k i =1

For all basic patterns and their combinations the question only sum and multiplication operators do not seem to be
of overall QoS evaluation has been studied properly and for a sufficient and thus, novel corresponding aggregation
number of QoS attributes so-called aggregation functions functions have to be elaborated. Moreover, some functions in
have been elaborated. The result of each aggregation the table are rather questionable. For example, it is a bit
function is the value of a corresponding attribute calculated questionable that the reputation of the sequential
for a proper pattern. Some of these functions without composition is just the average of component reputations. In
considering probabilities for XOR-splits are summarized in our opinion, this function should depend on service
Table I. In Table II, probabilities for XOR-splits and loops objectives and for some services be rather the minimum of
are taken into account when constructing aggregation reputations of component services.
functions. In these tables, the integer k denotes the number of Table II contains no aggregation functions for sequential
involved services while the integer n is used for the number and parallel patterns since they coincide with those in
of loop iterations, and pi is the probability of invocating the Table I; for a conditional pattern the average quality
service Si. The last column contains the references where evaluation is calculated, though the quality often is computed
such aggregation functions are taken from. [41] for each path separately, which allows to assess the
In fact, the set of functions in Table I is incomplete, since worst case, the best case, or the quality along the most
more attributes can be considered such as popularity, probable execution path. When probabilities are given, the
especially for social networks [42]. When involving some quality of a loop pattern is calculated for an arbitrary number
logic models, such as fuzzy logic, k-value logic etc., or of iterations.
modular arithmetic models, aggregation functions which use

TABLE II. AGGREGATION FUNCTIONS FOR QOS ATTRIBUTES WHEN PROBABILITIES ARE INVOLVED

Compositional pattern
QoS Attribute Reference
XOR-split Loop

k k (∏ j −1
)(1 − p )(¦ c + ∏
i =0 p i j
j
i =1 i
k
i =1 pi ¦i = j +1 ci )
k

¦ p i ci ¦
(1 − ∏ p )
Cost (c) 2
[20, 41]
i =1 j =1 k
i =1 i

k
¦ pi a i ¦
k(∏ p )(1 − p )∏
j −1
i =0 i j
j
i =1 a i
Availability (a)
i =1 j =1 (1 − ∏ p a ) k
i =1 i i
[20]

k k (∏ j −1
p )(1 − p )(¦ t + ∏
i =0 i j
j
i =1 i
k
i =1 pi ¦i = j +1 ti )
k

¦ pi t i ¦
(1 − ∏ p )
Execution time, response time (t) 2
[20, 41]
i =1 j =1 k
i =1 i

591
A vacation planner in Fig. 2 can be represented as the requests is available. Experimental results are provided in
composition of two services: Flight Booking (FB) and Hotel order to illustrate the effectiveness of a proposed service
Booking (HB) services. In the workflow in Fig. 4, nodes S filtering and ranking algorithm. In the ‘Selector’ tool [14]
and F are initial and final nodes correspondingly. functional and non-functional requirements for a composed
After asking for a flight ticket a user has three options: to service are explicitly separated, and the tool takes into
ask for a hotel room when flight dates are set; to change account both quality and context parameters.
flight dates or to quit the service when the flight dates cannot
be changed for some reason. In the case of booking a hotel IV. THE RELATIONSHIP BETWEEN DIFFERENT METRICS
room a request is processed only once. In both cases when a FOR QUALITY EVALUATING
hotel room is booked or not booked the service is quitted. The main purpose of any web service is to satisfy an
Correspondingly, using the above tables we can calculate the end-user and there still is no confidence that a user is
QoS of the vacation planner when knowing the QoS of FB satisfied even when the QoS is rather high. For this reason,
and HB services. another metric QoE for web services [28] is very popular
and there are many research papers on automatic estimation
of the QoE value and these methods can be implicitly
S FB F divided into five groups.
In many cases, the QoE is calculated manually based on
the expert evaluation. This approach is somehow inherited
HB
from multimedia services, for example, when transmitting
Fig. 4. Workflow for the vacation planner service
video information after a number of experiments the experts
can evaluate the QoE (as an integer between 1 and 10, for
We also note that sometimes workflows, which are often instance) while varying the number of lost packages or
used for calculating the composition QoS, are insufficient for response delays, and after many experiments the
the precise evaluation of the quality of the composite service. dependency of QoE on QoS parameters is extrapolating.
Thus, more complex trace models such as finite transition The same experiments may be done for web services [see,
systems are needed to accurately evaluate the quality of the for example, 19]. More complex functions are considered
composition when service component quality is given. for automatic calculating QoE based on the known QoS
Another interesting issue devoted to the problem of a value.
quality-aware component selection is not considered in this The experimental evaluation of user satisfaction also can
paper. The latter is often reduced to specifying a relevant be done based on passive and active traffic analysis as
utility function and optimizing its value using classical local proposed in [6]. End-user activity is being monitored, and
and global optimization algorithms [43-48]. In the next influence of service quality parameters (like latency,
subsection, we present some existing tools for deriving a bandwidth, etc.) on frequency of cancellation and reload
quality-aware composition. requests is assessed.
D. Tools for deriving a quality-aware composition Some approaches are rather concerned on determining a
A number of tools and automated methods for deriving a ‘good’ continuous approximation function based on QoS
quality-aware composition based on service selection have attribute values referring to psychophysics and cognitive
been developed [11-14]. Most approaches are developed studies. If such function can be approximated using
under the assumption that the compositional workflow (or a exponential/logarithmic dependencies between QoS and
work plan) is given, and problems of deriving service QoE [17] then the QoS evaluation is considered as
composition and quality-aware service selection are satisfactory. This dependency is similar to the dependency
separated. However, in the ‘Qsynth’ tool [12] the authors of human perception on physiological stimulus magnitude
combine these two sub-problems and prove that it allows known as logarithmic Weber-Fechner law when interpreting
increasing the scalability and enhancing the properties of the the values of QoS parameters as stimulus and the QoE value
QoS-aware composition. In order to handle QoS information, as perception.
special monitors are implemented, which dynamically collect Methods of another group are based on artificial
the QoS information, track history statistics and predict QoS intelligence techniques when automatic calculating the QoE
parameters in the real time using some simulation value, such as fuzzy logic, neural networks etc. Fuzzy logic
techniques. The composition is derived to assure both is used to calculate the correlation between measured
functional and quality constraints and the authors claim that network parameters and to predict the value of QoE [see, for
a proposed algorithm can provide the global optimal example 27, 34]. Since this correlation differs from service
solution. In [13], another approach that combines the to service and is usually significantly non-linear, it is
derivation of a quality-aware service composition and the essential to introduce adaptive algorithms to derive function
quality-aware component selection is proposed based on
coefficients automatically for given services. In [27], for
classification techniques. Since the main objective of a
service is to satisfy an end-user, the authors mention that a example, neural network learning algorithms are used to
better selection can be performed when the history of user develop the fuzzy logic based tool to determine values and

592
grades (excellent, good, average, or poor) for services. As this problem still remains unsolved. As a perspective, we are
mentioned above, the quality of novel tools may be verified planning to try our hand in developing a method for
with statistical data published in datasets [7]. automatic QoE evaluation based on trace models.
The last group of methods deals with new metrics rather
ACKNOWLEDGMENT
different from QoS and QoE. The Quality of Business
(QoBiz) [20] and Quality of Design (QoD) [29] can serve as The work is partly supported by Futur et Ruptures
example of such metrics. In the paper [20], the QoBiz is scholarship from Fondation Telecom (France) for Olga
evaluated in terms of money being received during executed Kondratyeva, and by Russian Ministry of High Education
transactions. Another specific metric QoD can be and Science (contract  14.37.21.0622, Russia).
interpreted as the quality of interaction between end-user
and client application. Below we show a picture (Fig. 5) REFERENCES
how the sets of attributes for some metrics are related.
[1] E. Al-Masri, and Q.H. Mahmoud, “Investigating web services on the
world wide web,” Proc. 17th international conference on World Wide
Web (WWW 08), ACM New York, NY, USA, 2008, pp. 795-804,
The set of QoS The set of doi:10.1145/1367497.1367605.
parameters QoBiz Automatic evaluation
parameters The QoE value [2] C.C. Chang, and L. Hsueh-Ming, “Integration of heterogeneous
The set of QoD
medical decision support systems based on web services,” Proc.
parameters Ninth IEEE International Conference on Bioinformatics and
BioEngineering (BIBE 09), IEEE Press, June 2009, pp.415-422,
doi:10.1109/BIBE.2009.59.
Fig. 5. The relationship between sets of attributes used for different metrics
when evaluating the quality of web services [3] F. Curbera, et al., “Unraveling the web services web: an introduction
to SOAP, WSDL, and UDDI,” IEEE Internet Computing, vol. 6, no.
2, March-April 2002, pp. 86-93, doi: 10.1109/4236.991449.
As it is shown in Fig. 5, the main purpose still remains
[4] L. Bentakouk, P. Poizat, and F. Zaïdi, “Checking the behaviroal
to automatically calculate the user satisfaction (QoE) based conformance of web services with symbolic testing and an SMT
on the known values of attributes. Based on the results of solver,” Proc. Test and Proof conference, (TAP 11), Zurich, 2011,
this paper we expect that trace models could be very helpful pp. 33-50.
for such evaluation [5] IEEE Xplore digital library, url: http://ieeexplore.ieee.org.
[6] S. Khirman, and P. Henriksen, “Relationship between Quality-of-
V. CONCLUSION Service and Quality-of-Experience for public Internet service,” Proc.
Passive and Active Measurement Workshop 2002 (PAM 2002), Fort
In this paper, we have discussed an approach for Collins, Colorado, USA, March 2002.
automatic QoS calculating that remains one of most general [7] E. Al-Masri, and Q.H. Mahmoud, “Quality of web services dataset,”
metrics when evaluating the web service quality. The url: http://www.uoguelph.ca/~qmahmoud/qws/index.html.
methods and techniques for QoS evaluation implicitly fall [8] Directory of Public SOAP Web Services, url: http://www.service-
into two groups. The first group of methods is based on repository.com.
experimental evaluation of QoS and QoE. Such experimental [9] Online community for the Universal Description, Discovery, and
evaluation might be performed based on passive or active Integration OASIS Standard, http://uddi.xml.org.
network traffic analysis and for these techniques linear [10] Web service depository XMethods, url: http://www.xmethods.net:
combinations are used as a formal model. The second group [11] J. Rao and X. Su, “A survey of automated Web service composition
of methods for evaluating QoS of web services contains methods,” in Proceedings of the first International workshop on
Semantic Web Services and Web Process Composition, San Diego,
methods which rely on more complex formal models such as California, USA, 2004, pp. 43–54.
trace models. We have also discussed how the quality of a [12] W: Jiang, C: Zhang, Z: Huang, M: Chen, S: Hu1, “QSynth : A Tool
composite web service can be evaluated by the use of such for QoS-Aware Automatic Service Composition”, in Proceedings of
models. IEEE International Conference on Web Services (ICWS), 2010, pp.
At the same time, the main purpose of any web service is 42-49.
to satisfy an end-user and despite of the fact that QoS [13] W. Nagy, H. M. O. Mokhtar, Ali El-Bastawissy, “A Flexible Tool for
contains a number of new attributes which are far away from Web Service Selection in Service Oriented Architecture”, in
International Journal of Advanced Computer Science and
traffic analysis and can be automatically calculated using Application; Dec 2011, vol. 2 issue 12, pp. 191-201.
formal models, there is no confidence that a user is satisfied [14] A. Alti, A. Boukerram, P. Roose, “Selector: A tool for dynamic
even when the QoS is rather high. For this reason, we briefly service selectionand management”, in Journal of Computing, vol. 4,
mentioned other metrics such as QoD, QoBiz, etc. for web issue 4, 2012, pp. 23-32.
services and have described the correspondence between [15] Hyun Jong Kim, Dong Hyeon Lee, Jong Min Lee, Kyoung Hee Lee,
them. More research is needed in order to establish a good Won Lyu, Seong Gon Choi, “The QoE Evaluation Method through
correlation between different metrics/functions used for the QoS-QoE Correlation Model,” Proc. 4th Int. Conf. on Networked
Computing and Advanced Information Management (NCM 08),
evaluating the quality of a web service, in particular, Sept. 2008, vol. 2, pp. 719-725, doi: 10.1109/NCM.2008.202
between automatically estimated quality attributes and the
[16] A. Morais, A. Cavalli, “Deliverable D2.1 – State of the art of
QoE, which is related to the quality of service that satisfies SQM/CEM technology, tools, and standartization”, IPNQSYS project,
an end-user. There is a number of research work on how to 2012, url: http://projects.celtic-initiative.org/ipnqsis/IPNQSIS-D21.pdf
automatically evaluate QoE based on the QoS results, but

593
[17] P. Reichl, S. Egger, R. Schatz, and A. D'Alconzo, “The Logarithmic [33] Charles L. Hedrick Rutgers. “An Introduction to IGRP”, url:
Nature of QoE and the Role of the Weber-Fechner Law in QoE http://www.cisco.com/en/US/tech/tk365/technologies white paper091
Assessment,” Proc. IEEE International Conference on 86a008000c8ael.shtml
Communications (ICC 2010), IEEE Press, May 2010, pp.1-5, [34] J. Pokherel, B. Wehbi, A, Morais, A. Cavalli, and E. Allitaire,
doi:10.1109/ICC.2010.5501894 “Estimation of QoE of video traffic using a fuzzy expert systems,”
[18] S.X. Sun, and J. Zhao, “A decomposition-based approach for service Proc. IEEE 10th Consumer communication & networking conf.
composition with global QoS guarantees,” Information Sciences, vol. (CCNC 2013), January 2013, pp. 224-229, doi:
199, Sept. 2012, pp.138–153, doi:10.1016/j.ins.2012.02.061. 10.1109/CCNC.2013.6488450
[19] F. Lalanne, A. Cavalli, and S. Maag, “Quality of experience as a [35] J.P. Escobedo Del Cid, Symbolic test case generation for testing
selection criteria for web services,” Proc. 8th Int. Conf. on Signal orchestrators in context, PhD thesis, Telecom SudParis, 2011.
Image Technology and Internet Based Systems (SITIS 2012), IEEE [36] C. Gaston and P. Le Gall, “About incremental model-based testing of
Press, Nov. 2012, pp.519-526, doi: 10.1109/SITIS.2012.81 web service orchestrations,” Slides from TAROT’2012,
[20] J. Cardoso, A. Sheth, J. Miller, J. Arnold, and K. Kochut, “Quality of url:http://tarot2012.univ-fcomte.fr/?talks.
Service for Workflows and Web Service Processes,” Journal of Web [37] I: Rodriguez, G: Diaz, P: Rabanal, J:A. Mateo, “A centralized and a
Semantics, 2004, pp.281-308 decentralized method to automatically derive choreography-
[21] J. El Hadad, M. Manouvrier, and M. Rukoz, “TQoS: Transactional conforming web service systems”, in The Journal of Logic and
and QoS-aware selection algorithm for automatic web service Algebraic Programming, vol. 81, issue 2, Feb. 2012, pp. 127–159
composition,” IEEE Transactions on Services Computing, vol. 3, [38] M. Diaz, V. Valero, H. Macia, J:A: Mateo, G: Diaz, “BPEL-RF Tool:
issue: 1, 2010, pp. 73-85, doi:10.1109/TSC.2010.5. An Automatic Translation fromWS-BPEL/WSRF Specifications to
[22] S. Dustdar, and W. Schreiner, “A survey on web services Petri Nets”; in Proceedings of The Seventh International Conference
composition,” International Journal of Web and Grid Services, vol. 1 on Software Engineering Advances (ICSEA), 2012, pp. 325-330.
iss. 1, August 2005, pp. 1-30. [39] S. Halle, G. Hughes, T. Bultan, and M. Alkhalaf, “Generating
[23] M.C. Jaeger, G. Rojec-Goldmann, and G. Muehl, “QoS aggregation Interface Grammars from WSDL for Automated Verification of Web
for web service composition using workflow patterns,” Proc. Services”, in Proceedings of 7th International Joint Conference,
Enterprise Distributed Object Computing Conf. (EDOC 2004), 2004, ICSOC-ServiceWave, Nov. 2009, pp. 516-530.
pp.149 – 159. [40] W. Zhang, “An Adaptive XML Parser for Developing High-
[24] Proceedings of the annual IFIP conference TestCom (International Performance Web Services”, in Proceedings of IEEE Fourth
conference on Testing Communication Systems) (formely IWPT International Conference on in eScience, Dec. 2008, pp. 672-679.
(International Workshop on Protocol Testing) and ICTSS [41] H. Zheng, W. Zhao, J. Yang, and A. Bouguettaya, “QoS analysis for
(International conference on testing communicating systems and web service compositions with complex structures,” IEEE
software), 1988-2012. Transactions on service computing, in press,
[25] P.-C. Héam, O. Kouchnarenko, and J. Voinot, “Component doi:10.1109/TSC.2012.7.
simulation-based substitutivity managing QoS aspects.” Electr. Notes [42] E. Billionniere, D. Greiman, and K. Gosha, “A comparison of social
Theor. Comput. Sci. 260, 2010, pp. 109-123. service selection techniques,” Proc. 8th IEEE Int. Conf. Dependable,
[26] B.A. Trakhtenbrot, and Y.M. Barzdin', “Finite automata: behaviour Autonomic and Secure Computing (DASC 09), Dec. 2009, pp.260-
and synthesis (fundamental studies in computer science),” Elsevier 265, doi: 10.1109/DASC.2009.24.
Science Publishing Co Inc., U.S, 1973. 328 p. [43] Y. Yang, M. Dumas, L. García-Banuelos, A. Polyvyanyy, and
[27] B. Pernici, “Evaluating web service QoS: A neural fuzzy approach,” L. Zhang, “Generalized aggregate Quality of Service computation for
Proc. IEEE Int. Conf. Service-Oriented Computing and Applications composite services,” The Journal of Systems and Software, vol. 85,
(SOCA), 2011, pp. 1-6. 2012, pp. 1818– 1830.
[28] A. van Moorsel “Metrics for the Internet age: Quality of Experience [44] Tao Yu, and Kwer-Jay Lin, “Service selection algorithms for web
and Quality of Business,” Technical Report HPL-2001-179, 2001. services with end-to-end QoS constraints,” Proc. IEEE Int. Conf. on
[29] P. Reichl, “From ‘Quality-of-Service’ and ‘Quality-of-Design’ to e-Commerce Technology (CEC 04), 2004, pp. 129-136.
‘Quality-of-Experience’: A holistic view on future interactive [45] X. Deng, and C. Xing, “A QoS-oriented optimization model for web
telecommunication services,” Proc. 15th Int. Conf. on Software, service group,” Proc. 8th IEEE/ACIS Int. Conf. on Computer and
Telecommunications and Computer Networks (SoftCOM 2007), Sept. Information Science, 2009, pp.903-909.
2007, pp.1-6. [46] Yuan Yu Qian, and Hu Xiao Hui, “A web service selection approach
[30] D. Booth, H. Haas, F. McCabe, E. Newcomer, M. Champion, based on the authenticity of QoS data and the confidence of users,”,
C. Ferris, and D. Orchard, “Web services architecture,” W3C Proc. Int. Symp. on Computer Network and Multimedia Technology
Working Group Note, W3C Technical Reports and Publications, Feb. (CNMT 2009), Jan. 2009, pp. 1-5.
2004, url:http://www.w3.org/TR/ws-arch/. [47] L. Zeng, B. Benatallah, M. Dumas, J. Kalagnam, and Q.Z. Sheng,
[31] G. Diaz, J.-J. Pardo, M.-E. Cambronero, V. Valero, and F. Cuartero, “Quality driven web services composition,” Proc. 12th Int. Conf. on
“Automatic Translation of WS-CDL Choreographies to Timed World Wide Web (WWW 03), 2003, pp. 411-421,
Automata”, in Formal Techniques for Computer Systems and doi:10.1145/775152.775211.
Business Processes, Lecture Notes in Computer Science, vol. 3670, [48] Wu Du, and Hong Fan, “An automatic service composition algorithm
2005, pp 230-242. for constructing the global optimal service tree based on QoS,” Proc.
[32] J.E. Hopcroft, R. Motwani, and J.D. Ullman, “Introduction to IEEE International Geoscience and Remote Sensing Symposium
Automata Theory, Languages, and Computation,” 2nd ed., Addison- (IGARSS), 2010, pp. 3976-3979.
Wesley Publishing Company, 2001.

594

You might also like