Interactive assessment of simulated service qualities by business stakeholders: principles and research issues
We present the principles of an approach supporting the stakeholder involvement in a software process for service-oriented systems in a form of assessing the perceived quality of the software under development in its usage context. This method relies on interactive simulation of service performance...
Saved in:
| Date: | 2010 |
|---|---|
| Main Author: | |
| Format: | Article |
| Language: | English |
| Published: |
Інститут програмних систем НАН України
2010
|
| Subjects: | |
| Online Access: | https://nasplib.isofts.kiev.ua/handle/123456789/14657 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Cite this: | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues/ Shekhovtsov, A.// Пробл. програмув. — 2010. — № 2-3. — С. 288-298. — Бібліогр.: 89 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859482864874160128 |
|---|---|
| author | Shekhovtsov, A. |
| author_facet | Shekhovtsov, A. |
| citation_txt | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues/ Shekhovtsov, A.// Пробл. програмув. — 2010. — № 2-3. — С. 288-298. — Бібліогр.: 89 назв. — англ. |
| collection | DSpace DC |
| description | We present the principles of an approach supporting the stakeholder involvement in a software process for service-oriented systems in a form of assessing the perceived quality of the software under development in its usage context. This method relies on interactive simulation of service performance and reliability; simulation models are parameterized by the factors influencing service execution; business stakeholders experience simulated service qualities in simulated usage contexts and assess this experience; the obtained assessments can be later used throughout the system lifecycle as a means of control for the quality of the software under development.
Наведено принципи підходу, що підтримує участь зацікавлених осіб у процесі розробки сервіс-орієнтованих програмних систем у вигляді оцінювання сприйманої якості розроблюваної системи в контексті її використання. Цей підхід спирається на інтерактивне імітаційне моделювання продуктивності та надійності сервісів; параметрами імітаційних моделей є фактори, що впливають на виконання сервісів; зацікавлені особи висловлюють своє відношення до значень продуктивності та надійності, отриманих при взаємодії з імітаційними моделями якості сервісів у рамках виконання імітаційних моделей їх контекстів використання, надані оцінки можуть бути використані на різних етапах життєвого циклу програмного забезпечення як засоби контролю його якості.
|
| first_indexed | 2025-11-24T15:04:58Z |
| format | Article |
| fulltext |
Методи та засоби програмної інженерії
© Vladimir A. Shekhovtsov, 2010
288 ISSN 1727-4907. Проблеми програмування. 2010. № 2–3. Спеціальний випуск
УДК 681.03
INTERACTIVE ASSESSMENT OF SIMULATED SERVICE QUALITIES BY
BUSINESS STAKEHOLDERS: PRINCIPLES AND RESEARCH ISSUES
Vladimir A. Shekhovtsov
National Technical University «Kharkiv Polytechnical Institute»
21 Frunze Str., Kharkiv, Ukraine 61002
Phone: +38057 7076474. Fax: +38057-7076520. Email: shekvl@yahoo.com
We present the principles of an approach supporting the stakeholder involvement in a software process for service-oriented systems in a form
of assessing the perceived quality of the software under development in its usage context. This method relies on interactive simulation of
service performance and reliability; simulation models are parameterized by the factors influencing service execution; business stakeholders
experience simulated service qualities in simulated usage contexts and assess this experience; the obtained assessments can be later used
throughout the system lifecycle as a means of control for the quality of the software under development.
Наведено принципи підходу, що підтримує участь зацікавлених осіб у процесі розробки сервіс-орієнтованих програмних систем у
вигляді оцінювання сприйманої якості розроблюваної системи в контексті її використання. Цей підхід спирається на інтерактивне
імітаційне моделювання продуктивності та надійності сервісів; параметрами імітаційних моделей є фактори, що впливають на
виконання сервісів; зацікавлені особи висловлюють своє відношення до значень продуктивності та надійності, отриманих при
взаємодії з імітаційними моделями якості сервісів у рамках виконання імітаційних моделей їх контекстів використання, надані
оцінки можуть бути використані на різних етапах життєвого циклу програмного забезпечення як засоби контролю його якості.
1. Introduction
Collecting the opinions of business stakeholders is recognized in the current software engineering research and
practice as an important part of the software process. For example, software engineering lifecycle standards such as
ISO/IEC 12207 [35] emphasize the need of collecting such opinions in a form of stakeholder requirements by
establishing special Stakeholder Requirements Definition Process to deal with them. This is particularly true for
service-oriented software systems as developing software services require knowledge of their possible uses which is
difficult to obtain without the involvement of their prospective users [25].
Currently in most cases business stakeholders are involved in the development process to express their opinions
of the prospective system’s functionality [53]. It is not the only possible form of involvement, however: another kind is
the involvement through assessments of the quality of the prospective system.
A motivating example is as follows. Suppose the software under development (SUD) is a service-oriented
system intended for business stakeholders. If they will not have a chance to present their opinions regarding the desired
quality of the prospective system early on its development lifecycle these opinions could be easily overlooked and lost.
As a result, the understanding of the desired quality of the system becomes biased towards the view of the IT people:
“the inmates are running the asylum” [10]. This could lead to the stakeholder dissatisfaction with SUD quality late in
the development lifecycle, the loss of confidence in the developers, and the failure of the project [61].
The current problem is that such assessment and its use in the software process in many domains still remains a
poorly investigated, difficult and error-prone task. Several arguments are in favor of this claim:
1. It is not realistic to expect that the business stakeholders can be fully involved in a software development
process if they cannot experience the SUD. Without such experience, they are forced to be speculative in their opinions
e.g. by formulating narrative statements such as “the system should be reliable” and “the system should have good
performance under any load”. Such speculations are not well suited to be a means of control for the quality of the SUD.
2. The process of assessing the SUD quality depends on the anticipated or implemented interaction of the SUD
with its environment. This assessment obviously may be a complex task especially for service-oriented systems. It is
difficult for stakeholders to invent the interaction scenarios “on-the-fly” without appropriate support.
We propose a method motivated by the above problem. It is based on the investigation of the ways to support the
stakeholder involvement in a form of assessing the perceived performance and reliability of the service-oriented SUD in
its usage context. To implement this support, we make service quality assessment mechanisms accessible to the
stakeholders without IT background. On top of these mechanisms, we establish higher-level policies solving particular
requirements engineering and architectural design problems (requirements elicitation, verification, architecture
assessment etc.). In this paper, we describe the concepts of the method and the proposed assessment mechanisms.
2. State of the art
The importance of the problem of involving business stakeholders in the development process as a means of
control for the quality of the produced artifacts was supported by its extensive discussion in the scientific literature. In
this section, we review several categories of methods addressing this problem mostly following the classification of the
methods to represent the quality of the prospective system proposed by Bosch [7], which includes scenario-based
techniques, prototyping, and simulation.
Методи та засоби програмної інженерії
289
Human interaction with stakeholders. Significant number of methods addresses the problem directly by
performing human interaction with stakeholders and collecting their opinions.
Request-centered techniques. These techniques emphasize the way of questioning the stakeholders and
processing their opinions. Social research interaction techniques [15] such as surveys, interviews, brainstorming,
questionnaires or checklists are used to solve this problem [45, 53] together with software engineering-specific
techniques such as CRC cards [4] or user-centered tabular glossaries [22]. The problem with these approaches is that
many of the stakeholders neither are used to nor trained in reasoning about the system in general and its qualities in
particular without having working experience regarding the targeted SUD.
Scenario-centered techniques. Such techniques [9] organize scenarios of stakeholder interaction with a
prospective system; they often rely on request-centered techniques to process the opinions of stakeholders. In the
manual scenario-centered approach, the stakeholders are requested to go through the scenarios (without actual
interacting with the SUD or its executable model) and express their opinions. The techniques of this kind are
extensively used to evaluate software architectures (examples are such methods as ATAM [41] and PASE [86]) or elicit
or analyze quality requirements (examples are Quality Attribute Workshops [2] and SRA tool [27]). Manual scenarios
are best suited for arranging the assessment of qualities which are revealed through complicated interactions (e.g.
security [71]) or development-related quality attributes such as modifiability [44].
There are shortcomings of these techniques as a means of addressing our problem: (1) they do not allow
stakeholders to experience quality with their senses in natural way (scenarios are usually narrative and do not offer
reality-like experience); (2) manual scenarios do not allow studying the dependency between performance-influencing
factors and the observed performance levels [7]; (3) for reliability, the scenarios in most cases cannot replace interaction
with the real system or its executable model as they are able to express it only by example [1].
Traditional prototyping. These techniques use system prototypes as an aid for stakeholders. The notion of
traditional prototype refers to a scaled-down version of the final system running in its intended context [19].
Horizontal prototypes. Early attempts to involve stakeholders in a development process using exemplification
under the title of rapid prototyping were made in the 80ies [51]. In particular, horizontal prototypes (lacking
implemented functionality) were introduced to simulate user interfaces in order to allow stakeholders to experience their
future environment. Such prototypes are still widely used, their usage is often embedded in the system usage scenarios
[72, 75]. The shortcomings of horizontal prototyping as an approach to address our problem are as follows: (1) it cannot
make stakeholders assess measurable software qualities as there is nothing to perceive as measurable in the prototyping
interactions; obtained information is limited to narrative notes; (2) such prototypes are mostly limited to making
stakeholders look at the system via its prospective user interface; this is of less value for service-oriented systems as the
client user interface can be not readily known at the time of development; (3) they are not well suited to taking into
account factors influencing system execution and experimenting with different sets of values for such factors.
Vertical prototypes. Such prototypes implement some part of the system functionality in deeper (up to complete)
detail. They can be used to make stakeholders experience operational software qualities (such as performance) and
assess such qualities. They are also not very suitable for our problem due to the following limitations (discussed in e.g.
[7]): (1) creating and maintaining such prototypes requires programming skills; (2) running prototype realistically
requires access to the target system including its hardware and human users, otherwise the stakeholder experience will
differ from the reality; (3) it is difficult to experiment with the vertical prototype trying out the different sets of factors
influencing software qualities (doing “what-if” analysis) as such experiments are usually too costly and cannot be
performed by non-programmers; (4) the solutions elaborated for such prototypes cannot be easily reused; (5) it is not
possible to “compress” or “expand” the time during the execution of prototype; it complicates realistic exposing of the
system reliability as it usually requires running in compressed time to reveal the relevant quality values.
«Live» prototyping. Several research-based [29, 39] and industrial (iRise studio, Axure RP etc.; the survey is in
[54]) tools extend the rapid prototyping technique by allowing non-programmers to build and execute interactive
software imitations (called “simulations” in their documentation) to get stakeholder’s feedback.
By simulation, their authors mean an imitation of system behaviour visible through its user interface. To
implement such imitation, they provide interactive user interface animations (similar to those used by rapid prototyping)
driven by the proposed scenarios of SUD behaviour. The only difference from the traditional rapid prototyping is that
these animations are “live”: the imitated program displays an anticipated user interface allowing stakeholders to play
with its visible elements making it show other elements according to the predefined navigational scenario, formulate
requirements while looking at this interface, and attach them to its visible elements. As with rapid prototyping, the
elicited information is limited to narrative notes. Underlying the animation is an interpreter of the language for
specifying scenarios (either proprietary [54] or based on standard notation such as statecharts [29, 39]).
These tools are not directly suitable for our research problem because they implement different type of
stakeholder involvement. Their purpose is to make stakeholders think about the system functionality while looking at
some version of the behaviour of its user interface. To solve our research problem, however, it is necessary to make
stakeholders think about the system quality while perceiving the numerical values of its attributes. This reflects the
fundamental difference between functional and quality requirements: whereas the functional requirements are
specifications of the anticipated systems behaviour (mostly in the narrative form as e.g. in the use case specifications),
the quality requirements are qualities that the product must have accompanied with quantified criteria for measuring
those qualities: “If you cannot quantify and measure a [non-functional] requirement, it is not really a requirement” [61].
Методи та засоби програмної інженерії
290
Another problem with these tools is that they are rather inflexible. As their output is limited to textual (free-
form) requirements their applicability is limited to the problem of requirements elicitation. The problem of stakeholder
involvement in a software process is wider than that: it is desirable to be able to involve business stakeholders into more
extended set of activities, such as requirements verification, requirements negotiation or software architecture
evaluation. These tools also do not target service-oriented systems.
Simulation-based techniques. These techniques use simulation in a sense of “the process of designing and
creating a computerized model of a real or proposed system for the purpose of conducting numerical experiments” [42]
to model service qualities in a way that can be used to support stakeholder involvement.
Service-level simulations. Some solutions simulate qualities of the particular software services or their
underlying components. Performance prototyping technique [31] allows performance simulations for the components to
be integrated into real-life environment as alternative for prototypes, other service performance solutions include
MOSES approach [12] based on UML 2 service representation, [16] implementing context-dependent discrete-event
service performance simulation, [3] using two-level representation of service performance. Service reliability
simulations are proposed in [24, 28], service error distribution is also simulated in [48]. The results of such simulations
are used to guide the evaluation of the software architecture [7, 26], requirements validation etc. The specific problem
with such simulations is that they are supposed to be used in standalone mode without explicit integration into the usage
processes.
Process-level simulations. Such solutions support quality simulations in context of usage scenarios (in a sense of
scenario-centered techniques) or complete business processes. Simulated scenarios [18, 30] are elaborated by analysts
to reveal particular quality characteristics. Business process simulation solutions are numerous (e.g. [36, 64]) but not
much of them produce numerical quality values. Among these solutions, UML-Ψ tool [50] simulates performance of the
processes described by annotated UML diagrams, [20, 21] augments BPM process definitions (expressed in e.g. BPMN
[8]) with information necessary to simulate performance, [73] enhances BPEL [84] programs with Java language blocks
implementing performance and reliability simulation. The specific problem with such simulations is that they are aimed
at process-aware applications [17] executed by BPM engines (where the process is a part of the SUD), so they treat
services as “black boxes” lacking detailed control on their models. As a result, they are less suited to our problem where
processes provide the usage context for the SUD treated as a set of services. Good idea would be to combine service-
level and process-level simulation models into a coherent model being able at the same time to address them separately.
There are also common problems with the current simulation solutions: (1) they, as a rule, are designed for use
by the persons with the IT background – not by business stakeholders; (2) many of them are not interactive [81]: the
results are available after the run is finished; there are, however, approaches to running business processes interactively
mostly based on task modeling [76]; (3) they mostly address particular quality attributes without paying attention to the
inter-attribute tradeoffs.
3. Problem statement
Investigating the state of the art in the area of stakeholder participation in software process activities reveals the
research problems to be addressed by the method. We describe these problems in logical sequence by formulating the
corresponding research questions.
The overall research question that motivates the development of the method is: How to involve business
stakeholders into the software development process as a means of control for the quality of the produced artifacts? As
the concept of software quality is broad, we plan to focus on two specific quality characteristics: performance (treated
as service latency) and reliability. These operational quantitative characteristics are well suited for representation by
simulation and the number of contexts for their assessment is more manageable in comparison with e.g. security (which
requires extensive set of complicated scenarios to be assessed in full). We also focus on service-oriented systems as the
specific category of software under development. As a result, the main research question this method trying to answer is
(RQ) How to involve business stakeholders into the development process for service-oriented software systems as a
means of control for the performance and reliability of the produced artifacts?
As an idea of addressing this research question, we propose to use interactive simulation of service performance
and reliability possessing the following features:
1. Service performance and reliability simulations are parameterized by the factors influencing execution of
service-oriented systems and executed interactively in the context of system usage;
2. Business stakeholders experience simulated service performance and reliability in simulated usage contexts
corresponding to their roles and assess this experience using specific scale;
3. The obtained assessments are used as a basis for software lifecycle activities such as requirements elicitation,
verification, and negotiation and as a means of control for software lifecycle decisions.
The reasons of selecting the simulation as an approach to the stated problem are as follows. It is free from the
problems of horizontal or “live” prototyping as it can really offer the possibility to model the measurable (quantitative)
quality attributes of the system under development in its environment. It is also free from the shortcomings of vertical
prototyping as [55]: (1) simulations can be made configurable by non-programmers; (2) no access to the target system is
required; (3) experimenting with simulation (trying different values for the factors influencing qualities of the target
system) is easier and less expensive (in some situations separate prototype can be necessary for every major
configuration); (4) simulation solutions can be reused as the same model can be instantiated to simulate different usage
contexts; (5) software reliability can be simulated; (6) simulation offers the control over the model time: it is possible to
compress or expand it. On the other hand, known problems with simulation solutions (outlined above) need to be
addressed.
Методи та засоби програмної інженерії
291
After establishing the basic idea of a solution, we formulate more specific research questions. We start from the
basic notion of service quality as we need to establish its conceptual representation (covering both performance and
reliability) for the purpose of the project. The corresponding question is: (RQ-1) How to conceptualize service
performance and reliability in a way most suitable for their assessment by business stakeholders?
Prior to addressing the simulation, we plan to investigate the ways of assessment of quantified service qualities
by business stakeholders. The corresponding question is: (RQ-2) How to establish the interaction procedures allowing
business stakeholder to experience and assess quantified service performance and reliability?
To address this question, we need to obtain the knowledge of the current practice in this field so we formulate
more specific research question: (RQ-2.1) What is the current practice for business stakeholders to perceive and assess
service performance and reliability? After gathering this knowledge, we need to formalize it in a way suitable for reuse
by addressing another more specific research question: (RQ-2.2) How to formalize and reuse common practices of such
assessment?
The reason of conducting this research prior to investigating simulations is that stakeholder perception of
qualities does not depend on their origin so we can use “mock-up” quality values before simulations are available. On
the other hand, while working on simulations, it is desirable to be able to test them via assessment interactions.
Main research question related to the simulation of service performance and reliability is (RQ-3) How to
simulate service performance and reliability in a way compatible with common practices of their assessment by
business stakeholders? To address this question, we need to understand the factors that influence real-world service
performance and reliability and should be taken into account in simulations (such as hardware capabilities, network
bandwidth, peak user load etc); this leads to the specific research questions (RQ-3.1) Which factors influence service
performance and reliability? After that, we need to make use of the gained knowledge in simulation-related context by
addressing the question: (RQ-3.2) How to make simulations of service qualities depend on these factors? After we
understand the influencing factors the next step is to develop the formal simulation models for performance and
reliability in a way suitable for reuse; this leads to the specific research question (RQ-3.3) How to formalize and reuse
simulation solutions for service performance and reliability?
After establishing approaches for simulating and assessing separate service qualities we plan to establish
integrated service-level quality assessment mechanisms by addressing the research question (RQ-4) How to combine
service performance and reliability simulation solutions with the corresponding assessment procedures to establish
service-level mechanisms for interactive assessment of simulated service qualities?
After establishing service-level mechanisms, we plan to investigate the idea of providing them to business
stakeholders in simulated usage contexts; this leads to the research question (RQ-5) How to integrate simulation and
assessment of service qualities into service usage contexts? To make stakeholders experience usage contexts we plan to
simulate business processes providing these contexts; this leads to the research question (RQ-5.1) How to simulate
service usage processes with integrated service-level simulation and assessment mechanisms?
Addressing the process-level research questions leads to elaborating mechanisms defined at the level of the
particular simulation session where a particular stakeholder in a particular role participates in a particular simulated
usage process. To reveal all the stakeholder opinions, the final assessment mechanism needs to gather assessment data
out of all the necessary usage processes, roles and stakeholder sessions; we plan to investigate the idea of such
mechanism by addressing the research question (RQ-6) How to make mechanism of process-driven assessment of
simulated service qualities gather all necessary assessments?
4. Description of the proposed approach
In this section, we describe in detail the proposed approach for answering the stated research questions. The
preliminary research leading to this approach was presented in [40, 69].
General methodology issues. The research is based on model-driven [33, 58] and ontology-based [32] software
engineering methodologies. Ontologies are used to describe the knowledge resulting from the investigation of the
problem space. Mechanisms addressing research questions in a solution space have underlying prescriptive models;
every model is defined by a metamodel. To elaborate this set of basic notions, conceptual modeling and system analysis
methodologies need to be applied; the treatment of system analysis follows the approach of M.Z.Zgurovsky and
N.D.Pankratova [88]. For dissemination, metamodels and ontologies need to be expressed using both practice-oriented
(e.g. UML, OWL) and formal (e.g. set theory, description logic) notations. As a result, a complete formal description of
the approach has to be obtained.
Modeling service quality attributes. To address the research question RQ-1 we provide the definition of quality
for proposed method and the means of conceptualizing the service quality attributes. We use Quality-Aware Predesign
Model for Services QAPM-S [70] for this purpose (extending it to satisfy the needs of the method).
QAPM-S models service operations with the notion of operation-type (defining the operations, their actors and
objects/parameters; these parameters are modeled with the notion of thing-type generalizing attributes, entities or
values) [52]; it also uses tabular model representation (called glossary) which is well understood by stakeholders [22]. It
models service quality as a hierarchy of quality characteristics [38], represents the facts that quality characteristics
influence each other and that stakeholders perceive qualities differently, follows aspect-oriented paradigm [56] in
representing service quality and functionality as separate concerns.
Interactive stakeholder assessment of quantified qualities. This stage of research addresses the question RQ-2.
The relevant research has to be conducted in two steps corresponding to RQ-2.1 and RQ-2.2: gathering knowledge
related to the stakeholder perception of qualities and developing formal interaction models of quality assessments.
Методи та засоби програмної інженерії
292
Stakeholder perception of quantified qualities. To address the question RQ-2.1, it is necessary to investigate the
principles of stakeholder perception of performance and reliability and the relevant representation/assessment scenarios.
For this purpose, it is necessary to establish field studies involving, in particular, interviews and surveys, working with
paper and user interface prototypes. The methodologies used in these studies are information perception and
visualization [14, 83], usability engineering and human-computer interaction applied to software engineering [68, 78],
social scaling [13, 15] etc. The results should describe the relevant knowledge by means of the ontology of stakeholder
quality perception. It organizes gained knowledge and can be used for elaborating interaction procedures. This
knowledge can be useful by itself to better understand people’s needs with respect to the assessments of numerical
software qualities.
Interaction models for quality assessments. Based on this knowledge, we can address the question RQ-2.2 by
formalizing the ways of making stakeholders experience and assess numerical quality values by means of process
models for quality visualization and end-user interactive assessment. For every quality characteristic the corresponding
interactive assessment model IAM needs to be elaborated. It describes the IAE mechanism which:
1. receives the numerical quality value as an input (the mechanism does not depend on its source);
2. makes the stakeholder experience this quality according to this value: (a) visually e.g. by displaying the
control board with a number of visual failure alerts to represent particular level of reliability; (b) by other means e.g. by
making the stakeholder wait for the time related to the particular latency;
3. involves the stakeholder in an interaction related to an assessment of his/her experience; in this case: (a)
appropriate research will be performed to choose scales for these assessments [13, 15]; (b) trade-off dependencies
among qualities can be used as assessment hints, e.g. the request for the latency assessment could include the hint that
improving the latency negatively affects reliability;
4. receives the assessment value from the stakeholder and returns this value as an output.
We aim at facilitating reuse of such models by establishing the IAML library storing IAMs for different quality
characteristics. The modeling techniques to be used for these models include conceptual user interface modeling [58]
and model-based user interface design (MB-UID) [59].
Parameter-dependent simulation of particular service qualities. We address the research question RQ-3 by
establishing the procedures for parameterized simulation of service qualities. These procedures are responsible for
composition and execution of simulation models.
Simulation parameters. To address the question RQ-3.1 it is necessary to investigate the factors influencing
service qualities (we define the notion of simulation parameters for these factors). As the parameters are related to the
real-world problem being simulated, we establish the parameter ontology, where all the knowledge about the factors
influencing service qualities is collected (e.g. their types, interdependencies, character of the influence on the model
etc.). This ontology needs to be consulted while developing parameter-related simulation modules following ontology-
based simulation modeling technique [5]. To address the question RQ-3.2, it is necessary to create a parameter slot in a
simulation model for every necessary parameter; prior to the execution of this model, the parameter value is expected to
be specified for every slot; these values influence the execution. It is also necessary to take into account possible
fuzziness of the parameter values to reflect incomplete knowledge of the factors.
Main simulation structure. The implementation of this structure follows the research on simulation modeling of
service-oriented architectures [3, 74, 77] in that we start from the simulation conceptual model [62] for this structure
with entities representing different parts of the SOA infrastructure (processing and storage hardware, network and
software infrastructure). We consider agent-oriented simulation technique [57] and its applications to SOA simulation
[74] as a means of instantiating this model as the set of communicating agents representing the above entities. Another
research alternative is discrete-event simulation ([42], its application to SOA is in e.g. [16]); these two techniques could
also complement each other. These approaches are better suited for modeling software systems than continuous
techniques such as system dynamics [49].
Composing the simulation models. Simulating particular service quality characteristic requires the particular set
of solutions to be introduced into the simulation model. On the other hand, making simulations depend on the particular
parameter also requires specific set of solutions to be implemented. To address research question RQ-3.3, for every
selected quality characteristics and parameter, we elaborate the set of simulation model solutions best suited for its
support and package it into the simulation module (obtaining the set of quality-related and parameter-related modules).
Separate research directions need to be pursued to establish simulation models for performance and reliability.
For performance, it is necessary to make the module implement the set of modeling procedures implementing actions
specific to service performance [6, 74, 89]. They should rely on a formal representation of performance (we plan to
investigate Colored Petri Nets [85] or their extension of Queuing Petri Nets [43] for this purpose) and utilize software
performance ontology (e.g. [46]). For reliability, the module needs to implement the set of fault injection procedures
[48] (taking into account error propagation [60] and other reliability-influencing issues) depending on selected fault and
failure models [11]; these procedures should rely on service dependability ontology [47].
After the simulation modules are available, we can develop the rules managing their composition to form service
quality simulations. Implementing this composition can be complicated as the particular parameter could influence the
simulation of several qualities whereas the particular quality simulation could depend on several parameters. In addition,
dependencies inside the sets of qualities and parameters as expressed in QAPM-S and parameter ontology need to be
considered as well (for example, simulating the reliability can depend on the simulated latency).
Методи та засоби програмної інженерії
293
Such complicated crosscutting relationships among the modules could be the case for the separation of concerns
inside the simulation conceptual model according to the aspect-oriented paradigm (aspect-oriented simulation modeling
- AOSM). This way, both support for the quality characteristic and parameter dependency can be viewed as separate
simulation concerns. The corresponding quality-related and parameter-related modules can be independently developed
as aspects representing the separated concerns (latency simulation aspect, reliability simulation aspect, network
bandwidth parameterization aspect etc.). For example, fault injection procedures implemented by reliability simulation
aspect could contain “hooks” for possible extensions by other aspects without revealing any information about the
character of such aspects (following the obliviousness principle of aspect-orientation).
We choose to follow the asymmetric aspect-oriented modeling approach [67] where the aspects crosscut the
main simulation structure which is preserved as the structure of the resulting simulation. The weaving of the simulation
aspects into this structure is shown on Fig.1. For example, the composition rule could connect fault injection procedures
implemented by the reliability aspect to particular external variables defined by aspects responsible for user load and
latency (populating the “callback” hooks defined for all these three aspects) and then weave the resulting load-
dependent and latency-dependent fault injection procedures into the simulation models for the hardware and software
infrastructure. The application of the aspect-oriented approach to the service quality modeling is novel; existing AOSM
techniques are specific to other domains [34, 79].
hardware
infra-
structure
class
software
infra-
structure
class
user load
aspect
network
bandwidth aspect
reliability aspect
latency
aspect
Base structure
Simulation model
hardware
infrastructure
class
software
infrastructure
class
Quality-related aspects
Parameter-related aspects
Fig. 1. Weaving the simulation model
This investigation follows the research on assessment interaction models, so we can immediately test simulation
results via the assessment interfaces. During the execution, every service quality simulation accepts the set of values for
all the parameters it depends on, and produces the value (or the set of values) for the simulated quality, so we just
replace mock-up quality sources with simulation modules.
To facilitate reuse of simulation solutions according to the research question RQ-3.3 we aim at creating QAL and
PAL aspect libraries storing simulation aspects corresponding to selected qualities and parameters and the library of
infrastructure simulation solutions (ISL). These simulation support libraries are supposed to be used for implementing
simulation of the complete services (described below). To validate the results of the research at this stage (i.e.
simulation models), we follow the research on simulation models verification and validation (e.g. [66]).
To implement proof-of-concept software support for quality simulations and for service-level execution
mechanisms, we use Java API for the state-of-the art simulation tools (AnyLogic (http://www.xjtek.com) being the main
choice and OMNet++ Java extension (http://www.omnetpp.org) the main alternative).
Standalone assessment of simulated service qualities. We address research question RQ-4 by elaborating IAS
mechanisms (interactive assessment of services) aimed at an assessment of simulated service qualities at the level of the
particular service. Fig.2 depicts the outline of IASC (composition) and IASE (execution) mechanisms.
Composing the service simulation. IASC inputs include QAPM-S representation of the set of qualities of interest
to be simulated and assessed (it can be a subset of the available qualities) and the set of necessary parameters. To get
the integrated quality simulation model QSM, we perform a composition of the simulation aspects corresponding to the
qualities of interest and the necessary parameters on top of the base simulation structure (all obtained from simulation
support libraries). The parameter slots are created for all the necessary parameters. Also, we select the set of interactive
assessment models from IAML for the qualities of interest and integrate them with QSM. The resulting service-level
simulation and assessment model IASM becomes the IASC output. At this stage of the project, it is transferred to IASE
for standalone execution (as shown on Fig. 2).
Service user
Assess-
ment
Simu-
lation
particular
service
Compo
sition
IASC: composition IASE: execution
IAML: assessment
model library
IAM
selection
PAL, QAL, ISL: simulation
support libraries
In
te
g
ra
ti
o
n
Parameter values
Parameter
definitions
QAPM-S: qualities
to assess
Simulated
qualities
Assessments
IASM
selected
IAMs
QSM
particular
run
Fig. 2. Mechanisms for interactive assessment of simulated service qualities
Методи та засоби програмної інженерії
294
Executing and assessing the service simulation. IASE is responsible for execution of both simulation and
assessment interaction submodels of IASM. The input for every IASE run is the set of parameter values corresponding
to the IASM parameter slots. As a result of the run, the set of simulated values for the qualities of interest is obtained
and presented to the service user for assessment via IAE mechanisms described by interaction models integrated into
IASM. The IASE outputs are this set of simulated qualities and the set of assessment results.
Service-level assessment example. Suppose we develop the CheckOrder service and plan to make users as sess
its latency and reliability which depend on network bandwidth and user load. In this case IASC inputs are the qualities
of latency and reliability and the parameters of bandwidth and load. The resulting model IASM(CheckOrder) is
composed from simulation model QSM(CheckOrder) (based on simulation aspects for latency, reliability, bandwidth
and load) and the interaction models for latency and reliability. IASE accepts IASM as the model to execute and, for its
run, obtains the parameter values for bandwidth and load. The outputs for the run are the values for simulated latency
(e.g. 0,5 sec) and reliability (e.g. 99.4%) and their assessments (e.g. the latency score of 8 out of 10).
IASC and IASE can be useful by themselves if the analyst just wants to ask stakeholders to assess the qualities
of the particular service or the list of services without burdening oneself with establishing detailed usage processes.
Process-driven assessment of simulated service qualities. We address research question RQ-5 by elaborating
IAP mechanisms (interactive assessment of processes) aiming at interactive assessment of simulated service qualities in
context of usage processes at the level of the particular process. Fig.3 depicts the outline of IAPC (composition) and
IAPE (execution) mechanisms. They rely on IASC and IASE dealing with individual services.
IASC(s2)
IASC(s3)
IAPC: composition
SIMC(a1)
IAPE: execution
particular usage process
Participant
in a role r1
particular role and run
Participant
in a role r2
IASE(s2)
IASE(s3)
SIME(a1)
Role
model
Assessment and
ineraction sets
QAPM-S: quality
characteristics
Parameter
definitions
Parameter
values
Interaction set
Assessment set
for r1
Assessment set
for r2
Simulated
qualities
Assessments
BPrM: model
of usage process
activity a1
activity a2
service s2
activity a3
service s3
IASM for s3
IASM
for s2
SIM for a1
IAPM
if role
r1 or r2
if role r1
if role r2
SIM for a1
IASM
for s2
IASM
for s3
Fig. 3. Interactive assessment of simulated service qualities in context of usage processes
Composing the process simulation. IAPC forms the simulation model of the usage process making it ready for
interactive assessment of service qualities. It accepts the following inputs:
1. The control flow model (CFM) for the usage process conforming to the network BPM notation such as
BPMN, process chains [80], Colored Petri Nets etc. (in the scope of this project we restrict the accepted notations to
those supported by the chosen BPM simulation tool). It is possible to use existing process model repositories such as
APROMORE [63] for storing and retrieving such models.
2. The role model for the usage process which includes: (a) the set of roles defined for process participants
(clerk, manager etc); (b) different sets of interaction activities for different roles; such activities make participants affect
the state of the process simulation (e.g. by selecting execution paths, specifying values to be used for subsequent
decisions etc); these interactions advance the model time; (c) different sets of assessment activities for different roles;
such activities correspond to the services of interest to be simulated and assessed by stakeholders; assessments do not
change the state of the process simulation (e.g. the model time has to remain the same as the simulation needs to be
“paused” while the stakeholder thinks over the assessment); (d) the sets of qualities of interest and necessary parameters
defined for every service of interest.
3. The flow simulation model (FSM) including all the extensions necessary to implement the process
simulation beyond those presented in a role model (according to e.g. the techniques presented in [65, 82]); to simplify
this model, all except interaction and assessment activities are treated as black boxes.
While composing the integrated model IAPM for the process, IASC mechanism is invoked for every service of
interest. It creates the IASM model for this service which is then integrated into IAPM (this is the case for a2 and a3
activities on Fig.4). For every interaction activity, SIMC mechanism for constructing the interaction model is invoked
and the resulting SIM model (depending on the type of activity) is also integrated into IAPM (see the activity a1 on
Fig. 4). SIMC prototype handling several basic interaction types will be established alongside IAPC.
The resulting model contains (1) the simulation logic defined by CFM and FSM (for the process) and simulation
submodels of different IASM models (for the services of interest); (2) the assessment logic defined by interaction
submodels of these models; (3) the interaction logic defined by the SIM models for all interaction activities.
Методи та засоби програмної інженерії
295
Executing and assessing the process simulation. The IAPM is executed by IAPE. Every run is presented to the
stakeholder belonging to the particular role (on Fig.3 two stakeholders of the roles r1 and r2 are shown interacting with
IAPE; in the scope of the project, we restrict ourselves with a single-user mode when every run involves exactly one
stakeholder; the multi-user mode can be investigated as a project extension). The input for every run of IAPE is the set
of parameter values for all the parameter slots defined for the services of interest.
During the run, the following actions are performed:
1. The basic simulation flow is managed by the model derived from the usage process CFM;
2. When the logic of the run requires invoking an activity representing the service of interest, the simulation of
its qualities and the assessment interaction logic are handled by IASE invoked for its IASM (on Fig.3, this is the case
for a2 and a3 representing s2 and s3). IASE inputs are parameter values for all the slots of this service.
3. When this logic requires interacting with the simulation, the logic of this interaction is handled by the SIME
mechanism invoked for the corresponding SIM (this is the case for a1 on Fig.3). SIME prototype performing several
basic interactions is expected to be elaborated alongside IAPE; we plan it to utilize available interaction modeling
techniques similar to those used for IAM.
The selection of actions could be different for different roles (on Fig.3, role r1 interacts with activities a1 and a2,
whereas role r2 interacts with a1 and a3). The outputs for IAPE run include the set of all simulated quality values for all
the services of interest and the set of corresponding assessment results.
Process-level assessment example. Suppose the task is to make order and bookkeeping clerks assess the services
of the order processing system in the context of order processing. In the CFM for the ProcessOrder usage process
CheckOrder activity (with performance as the quality of interest) in accessible to order clerks whereas CheckPayment
activity (for it, we are interested in reliability) is accessible to bookkeeping clerks; both these activities are backed by
services of interest. These activities are assessment activities for the respective roles. On the other hand,
AskForCanceling activity accessible to order clerks is not backed by a service and serves only to drive the simulation; it
is an interaction activity for this role. IAPC forms the IAPM by combining CFM(ProcessOrder) and
FSM(ProcessOrder) with IASM(CheckOrder), IASM(CheckPayment) and SIM(AskForCanceling) and transfers it to
IAPE for execution. When the simulation logic for the order clerk-driven run passes through CheckOrder, IASE
executes the corresponding IASM and obtains the performance assessment from the clerk as described in an example
for this mechanism; the same happens to reliability of CheckPayment during the run driven by bookkeeping clerk. In
case of passing through AskForCanceling its SIME asks the order clerk to decide what to do next.
Implementing the process simulation. We address research question RQ-5.1 by investigating an approach to
simulation which can be implemented by following existing simulation solutions defined for network-based BPM (such
as BPMN-supporting OXProS [23] or CPN Tools [37]). We propose to evaluate several possible ways to implement
this integration. One possibility is to implement solution in spirit of [3] which uses an implementation of process chains
for the control flow simulation and OMNet++ code to implement service simulations with a resulting model deployed
on OMNet++ engine. This way, IAPE could run process chains simulation code while IASE – service simulation code
(both backed by OMNet++ or AnyLogic). Another approach is to use an idea of [23] which performs template-based
conversion of BPMN diagram into Colored Petri Nets (CPN), this way, BPMN-based CFM can be converted into CPN
representation, augmented with Java-based service simulation code and run via CPN tools or OXProS engine.
Iterative assessment mechanism. We address research question RQ-6 with the iterative IIA mechanism
(iterative interactive assessment) which explores different variants of usage processes, roles and runs. The invariants for
all iterations are the set of services under development (corresponding to the chosen variant of SOA) and the QAPM-S
instance defined for this set. The iterations are as follows: the outer (composition) loop is by process (selecting different
usage processes for the services and composing simulation models with IAPC), the inner (execution) loops are by role
(selecting different roles and, as a result, exploring different subsets of services accessible for the role) and then by
simulation run (making different runs for different users acting in the chosen role). The results are obtained on the every
iteration of the inner loop. As a result, for every service of interest, we obtain simulated quality values and their
assessments in all its key usage contexts.
Cost management. Modular structure of the solutions allows for flexible management of the implementation
and deployment costs as different level of detail can be selected depending on the permissible cost. At the service level,
the cheapest option is to build simulation models out of the predefined aspects and interaction models according to the
selected set of qualities and influencing factors (the configuration is reduced to specifying qualities and factors of
interest). If the cost is not an issue, however, it is possible to prepare custom quality simulations for every service of
interest or combine predefined and custom simulation aspects. At the process level, the configurations vary from
making stakeholders work directly with services, to introducing simple “task list” interfaces for service simulations,
then to relying to available predefined processes (e.g. from APROMORE repository [63]) and to building custom
processes for the problem at hand. This flexibility offers cost management advantages compared to other solutions. For
example even the most complete process-based solutions such as [20, 21] are built in top-down fashion, as a result,
complete process model needs to be built in any case; other techniques are even less flexible. Another kind of cost
advantages are related to the choice of simulation over other quality modeling techniques such as prototyping [7, 55].
Методи та засоби програмної інженерії
296
5. Conclusions and future research
In this paper, we elaborated the concepts for the mechanisms of tool-supported stakeholder assessment of
simulated service performance and reliability in service usage contexts represented by simulated business process
models. We presented service-level and process-level mechanisms, examples of their usage, and possible directions of
their implementation. The development of these mechanisms should rely on research in such fields as human-computer
interaction, social sciences, information visualization, discrete and agent-based simulation of performance and
reliability.
Novelty of the proposed approach. After comparison with state-of-the-art stakeholder involvement techniques
we state that the proposed mechanisms possess the unique combination of features:
1. They rely on simulation models able to express quantified quality attributes; to form service simulations,
such models are integrated using novel aspect-oriented approach;
2. They allow interactive participation in simulations by business stakeholders; such participation is based on
formal interaction models elaborated as a result of studies of real interactions in field settings;
3. They integrate the simulations of service qualities (in particular, performance and reliability) into the
simulations of service usage processes and make these simulations accessible to stakeholders acting in particular roles
so that a set of services available for assessment depends on the role.
Future research directions. In this paper, we limited ourselves to the description of the assessment mechanisms.
These mechanisms form the foundations for the higher-level policies intended to address particular problems of
requirements engineering and architectural design. In particular, requirements elicitation policy to get the threshold
values for quality requirements out of stakeholder assessments and simulated quality values, requirements verification
policy to compare external requirements with requirements elicited via assessment mechanisms, requirements
negotiation policy inspired by systemwise optimization by V.M.Glushkov [87] to mutually adjust the resources of the
organization and the needs of stakeholders. These policies will be the subject of future research.
The ultimate future goal of the proposed method is to establish a lifecycle simulation support by asking the
stakeholders to make assessments of simulated qualities reflecting the current state of the software under development
as the development progresses with a purpose to make these assessments drive the development.
1. Amyot, D., Eberlein, A. An Evaluation of Scenario Notations and Construction Approaches for Telecommunication Systems Development //
Telecommunication Systems. – 2003. – Vol. 24. – N 1. – P. 61-94.
2. Barbacci, M., Ellison, R., Lattanze, A., et al. Quality Attribute Workshops (QAWs), Third Edition. – Carnegie Mellon University. – 2003.
3. Bause, F., Buchholz, P., Kriege, J., Vastag, S. A Framework for Simulation Models of Service-Oriented Architectures // In SIPEW 2008. – LNCS,
Vol. 5119. – 2008. – P. 208–227.
4. Bellin, D., Suchman-Simone, S. The CRC Card Book. – Addison-Wesley. – 1997.
5. Benjamin, P., Patki, M., Mayer, R. Using ontologies for simulation modeling // In WSC'06. – 2006. – P. 1151-1159.
6. Bertolino, A., De Angelis, G., Polini, A. A QoS Test-Bed Generator for Web Services // In ICWE'07. – LNCS, Vol. 4607. – Springer. – 2007.
– P. 17-31.
7. Bosch, J. Design and Use of Software Architectures. – Reading: Addison-Wesley. – 2000.
8. Business Process Model and Notation (BPMN) Version 1.2. –. OMG. – 2009.
9. Carroll, J.M. (ed.). Scenario-Based Design. – Wiley. – 1995.
10. Cooper, A. The Inmates are Running the Asylum. – Sams. – 2004.
11. Cortellessa, V., Grassi, V. Reliability Modeling and Analysis of Service-Oriented Architectures // In Baresi, L., Nitto, E.D. (eds.): Test and Analysis of
Web Services. – Springer. – 2007. – P. 339-362.
12. Cortellessa, V., Pierini, P., Spalazzese, R., Vianale, A. MOSES: MOdeling Software and platform architEcture in UML 2 for Simulation-based
performance analysis // In QoSA 2008. – LNCS, Vol. 5281. – Springer. – 2008. – P. 86-102.
13. DeVellis, R.F. Scale Development: Theory and Applications. – Sage Pubs. – 2003.
14. Diehl, S. Software visualization: visualizing the structure, behaviour, and evolution of software. – Springer. – 2007.
15. Drew, P., Raymond, G., Weinberg, D. Talk and Interaction in Social Research Methods. – Sage Pubs. – 2006.
16. Driss, M., Jamoussi, Y., Jezequel, J.-M., et al. A Discrete-Events Simulation Approach for Evaluation of Service-Based Applications // In Proc.
ECOWS’08. – IEEE. – 2008. – P. 73-78.
17. Dumas, M., Van der Aalst, W., ter Hofstede, A. (eds.). Process-Aware Information Systems. – Wiley-IEEE. – 2005.
18. Egyed, A. Dynamic Deployment of Executing and Simulating Software Components // In Component Deployment. – LNCS, Vol. 3083. – Springer. –
2004. – P. 113-128.
19. Floyd, C. A Systematic Look at Prototyping // In Approaches to Prototyping. – Springer. – 1984. – P. 1-17.
20. Fritzsche, M., Gilani, W., Fritzsche, C., et al. Towards utilizing model-driven engineering of composite applications for business performance analysis
// In ECMDA-FA'08. – 2008. – P. 369-380.
21. Fritzsche, M., Picht, M., Gilani, W., et al. Extending BPM environments of your choice with performance related decision support // In BPM 2009. –
LNCS, Vol. 5701. – 2009. – P. 97-112.
22. Galle, D., Kop, C., Mayr, H.C. A Uniform Web Service Description Representation for Different Readers // In Proc. ICDS'08. – IEEE CS Press. –
2008. – P. 123-128.
23. Garcia-Banuelos, L., Dumas, M. Towards an Open and Extensible Business Process Simulation Engine // In Proc. CPN'09. – 2009.
24. Gokhale, S.S., Lyu, M.R., Trivedi, K.S. Reliability Simulation of Component-Based Software Systems // In Proc. ISSRE’98. – 1998. – P. 192-201.
25. Graham, I. Requirements Modelling and Specification for Service Oriented Architecture. – Wiley. – 2009.
26. Grassi, V., Mirandola, R., Sabetta, A. Filling the gap between design and performance/reliability models of component-based systems: A model-driven
approach // The Journal of Systems and Software. – 2007. – Vol. 80. – P. 528–558.
Методи та засоби програмної інженерії
297
27. Gregoriades, A., Sutcliffe, A. Scenario-Based Assessment of Nonfunctional Requirements // IEEE Trans. Software Eng. – 2005. – Vol. 31. – N 5. – P.
392-409.
28. Grishikashvili Pereira, E., Pereira, R. Simulation of fault monitoring and detection of distributed services // Simulation Modelling Practice and Theory.
– 2007. – Vol. 15. – P. 492–502.
29. Harel, D., Politi, M. Modeling Reactive Systems with Statecharts. – McGraw-Hill. – 1998.
30. Haumer, P., Heymans, P., Jarke, M., Pohl, K. Bridging the Gap Between Past and Future in RE: A Scenario-Based Approach // In Proc. RE'99. – IEEE
CS Press. – 1999. – P. 66-73.
31. Hennig, A., Hentschel, A., Tyack, J. Performance Prototyping - Generating and Simulating a Distributed IT-System from UML Models // In Proc.
ESM’2003. – IEEE. – 2003.
32. Hesse, W. Ontologies in the Software Engineering process // In Proc. EAI'05. – Ceur-WS.org, Vol. 141. – 2005.
33. Hesse, W., Mayr, H.C. Modellierung in der Softwaretechnik: eine Bestandsaufnahme // Informatik Spektrum. – 2008. – Vol. 31. – N 5. – P. 377-393.
34. Ionescu, T.B., Piater, A., Scheuermann, W., et al. An Aspect-Oriented Approach for Disaster Prevention Simulation Workflows on Supercomputers,
Clusters, and Grids // In Proc.DS-RT 2009. – 2009. – P. 21-33.
35. ISO. ISO/IEC 12207:2008, Information technology – Software life cycle processes. –. – 2008.
36. Jansen-Vullers, M., Netjes, M. Business process simulation – a tool survey // In CPN Tools Workshop. – 2006.
37. Jensen, K., Kristensen, L.M. Coloured Petri Nets. – Springer. – 2009.
38. Jureta, I.J., Herssens, C., Faulkner, S. A Comprehensive Quality Model for Service-Oriented Systems // Software Quality Journal. – 2009. – Vol. 17. –
N 1. – P. 65-98.
39. Jwoa, J.-S., Cheng, Y.C. Pseudo software: A mediating instrument for modeling software requirements // Journal of Systems and Software. – 2009,
in press.
40. Kaschek, R., Kop, C., Shekhovtsov, V.A., Mayr, H.C. Towards Simulation-Based Quality Requirements Elicitation: A Position Paper // In REFSQ 2008.
– LNCS, Vol. 5025. – Springer. – 2008. – P. 135-140.
41. Kazman, R., Barbacci, M., Klein, M., Carriere, S.J. Experience with Performing Architecture Tradeoff Analysis // In Proc. ICSE'99. – ACM. – 1999.
42. Kelton, W.D., Sadowski, R.P., Sadowski, D.A. Simulation with Arena. – McGraw-Hill. – 2004.
43. Kounev, S. Performance Modeling and Evaluation of Distributed Component-Based Systems using Queueing Petri Nets // IEEE Trans Soft Eng. –
2006. – Vol. 32. – N 7. – P. 486-502.
44. Lassing, N., Bengtsson, P., Bosch, J., Vliet, H.V. Experience with ALMA: Architecture-Level Modifiability Analysis // Journal of Systems and
Software. – 2002. – Vol. 61. – N 1. – P. 47-57.
45. Lauesen, S. Software requirements: Styles and techniques. – Addison-Wesley. – 2002.
46. Lera, I., Sancho, P.P., Juiz, C., et al. Performance assessment of intelligent distributed systems through software performance ontology engineering
(SPOE) // Software Qual J. – 2007. – Vol. 15. – P. 53-67.
47. Looker, N., Gwynne, B., Xu, J., Munro, M. An Ontology-Based Approach for Determining the Dependability of Service-Oriented Architectures // In
WORDS’05. – 2005. – P. 171- 178.
48. Looker, N., Xu, J., Munro, M. Determining the dependability of Service-Oriented Architectures // Intl J of Simulation and Process Modelling. – 2007. –
Vol. 3. – N 1-2. – P. 88 - 97.
49. Mårtensson, F., Jönsson, P., Bengtsson, P., et al. A Case Against Continuous Simulation for Software Architecture Evaluation // In Proc. ASM'03. –
2003. – P. 97-105.
50. Marzolla, M., Balsamo, S. UML-PSI: the UML Performance SImulator // In Proc. QEST’04. – IEEE. – 2004.
51. Mayr, H.C., Bever, M., Lockemann, P.C. Prototyping Interactive Application Systems // In Budde, R., Kuhlenkamp, K., Mathiassen, L. (eds.):
Approaches to Prototyping. – Berlin: Springer-Verlag. – 1984. – P. 105-121.
52. Mayr, H.C., Kop, C. Conceptual Predesign - Bridging the Gap between Requirements and Conceptual Design // In Proc. ICRE '98. – IEEE CS Press. –
1998. – P. 90-100.
53. McManus, J. Managing Stakeholders in Software Development Projects. – Butterworth-Heinemann. – 2004.
54. Memmel, T., Gundelsweiler, F., Reiterer, H. Prototyping Corporate User Interfaces - Towards a Visual Specification of Interactive Systems // In Proc.
IASTED-HCI 2007. – 2007.
55. Miller, M.J., Pulgar-Vidal, F., Ferrin, D.M. Achieving Higher Levels of CMMI Maturity Using Simulation // In Proc. WSC'02. – IEEE. – 2002.
– P. 1473-1478.
56. Moreira, A., Araújo, J., Brito, I. Crosscutting quality attributes for requirements engineering // In SEKE'02. – 2002. – P. 167-174.
57. North, M.J., Macal, C.M. Managing Business Complexity: Discovering Strategic Solutions with Agent-Based Modeling and Simulation. – Oxford
Univ. Press. – 2007.
58. Pastor, O., Molina, J.C. Model-Driven Architecture in Practice. – Springer. – 2007.
59. Paternò, F. Model-based Design and Evaluation of Interactive Applications. – Springer. – 2000.
60. Popic, P., Desovski, D., Abdelmoez, W., Cukic, B. Error Propagation in the Reliability Analysis of Component based Systems // In Proc. ISSRE’05. –
IEEE. – 2005.
61. Robertson, S., Robertson, J. Mastering the Requirements Process, 2nd ed. – Addison-Wesley. – 2006.
62. Robinson, S. Conceptual modelling for simulation // J Operational Research Society. – 2008. – Vol. 59. – N 3. – P. 278-290.
63. Rosa, M.L., Reijers, H.A., van der Aalst, W., et al. APROMORE: An Advanced Process Model Repository. – QUT Reprints. – 2009.
64. Rozinat, A., Wynn, M., van der Aalst, W., et al. Workflow simulation for operational decision support // Data & Knowl Eng. – 2009. – Vol. 68.
– P. 834–850.
65. Rücker, B. Building an Open Source Business Process Simulation Tool with JBoss jBPM. – Stuttgart UAS. – 2008.
66. Sargent, R.G. Verification and Validation of Simulation Models // In Proc. WSC'08. – 2008. – P. 157-169.
67. Schauerhuber, A., Schwinger, W., Kapsammer, E., et al. Towards a Common Reference Architecture for Aspect-Oriented Modeling // In Proc. 8th Intl.
Workshop on AO Modeling. – 2006.
68. Seffah, A., Gulliksen, J., Desmarais, M.C. (eds.). Human-Centered Software Engineering. – Springer. – 2005.
69. Shekhovtsov, V., Kaschek, R., Zlatkin, S. Constructing POSE: a Tool for Eliciting Quality Requirements // In Proc. ISTA 2007. – LNI, Vol. P-107. – GI.
– 2007. – P. 187–199.
70. Shekhovtsov, V.A., Kaschek, R., Kop, C., Mayr, H.C. Relational service quality modeling // In J.Suzuki (ed.) Developing Effective Service Oriented
Architectures: Concepts and Applications in Service Level Agreements, Quality of Service and Reliability. – IGI Global. – 2010, in press.
71. Sindre, G., Opdahl, A.L. Eliciting security requirements with misuse cases // Req. Eng. – 2005. – Vol. 10. – N 1. – P. 34-44.
Методи та засоби програмної інженерії
298
72. Sutcliffe, A., Ryan, M. Experience with SCRAM: a scenario requirements analysis method // In Proc. ICRE'98. – IEEE CS Press. – 1998. – P. 164-171.
73. Tewoldeberhan, T., Janssen, M. Simulation-based experimentation for designing reliable and efficient Web service orchestrations in supply chains //
El.Commerce Res. and Apps. – 2008. – Vol. 7. – P. 82–92.
74. Thanheiser, S., Liu, L., Schmeck, H. SimSOA: an approach for agent-based simulation and design-time assessment of SOC-based IT systems // In
Proc.ACM Symp. on Applied Computing. – ACM. – 2009. – P. 2162-2169.
75. Tolstedt, J.L. Prototyping as a means of requirements elicitation // In Proc. SAE International Off-Highway Congress. – SAE Technical Paper Series,
Vol. 2002-01-1466. – 2002.
76. Trætteberg, H., Krogstie, J. Enhancing the Usability of BPM-Solutions by Combining Process and User-Interface Modelling // In Proc. POeM'08. –
LNBIP, Vol. 15. – Springer. – 2008. – P. 86-97.
77. Tsai, W.T., Cao, Z., Wie, X., et al. Modeling and Simulation in Service-Oriented Software Development // Simulation. – 2007. – Vol. 83. – N 1. –
P. 7-32.
78. Tullis, T., Albert, W. Measuring the User Experience. – Morgan Kaufmann. – 2008.
79. Um, I.-S., Lee, H.-C., Cheon, H.-J. Determination of Buffer Sizes in Flexible Manufacturing System by using the Aspect-oriented Simulation // In Proc.
ICCAS'07. – IEEE. – 2007. – P. 1729-1733.
80. van der Aalst, W. Formalization and verification of event-driven process chains // Inf Soft Tech. – 1999. – Vol. 41. – N 10. – P. 639-650
81. van der Aalst, W., Nakatumba, J., Rozinat, A., Russell, N. Business process simulation: how to get it right? – TU Eindhoven. – 2008.
82. Waller, A., Clark, M., Enstone, L. L-SIM : Simulating BPMN Diagrams With A Purpose Built Engine // In Proc. WSC'06. – 2006. – P. 591-597.
83. Ware, C. Information Visualization: Perception for Design. – Morgan Kaufmann. – 2004.
84. Web Services Business Process Execution Language (WS-BPEL) Version 2.0. –. OASIS. – 2007.
85. Wells, L. Performance analysis using coloured Petri nets // In MASCOTS’02. – 2002. – P. 217–221.
86. Williams, L.G., Smith, C.U. PASA: A Method for the Performance Assessment of Software Architecture // In Proc. 3rd Workshop on Software
Performance. – 2002.
87. Глушков, В.М. О системной оптимизации // Кибернетика. – 1980. – № 5. – С. 89–91.
88. Згуровский, М.З., Панкратова, Н.Д. Системный анализ: Проблемы, методология, приложения. – К: Наук. думка. – 2005. – 743 с.
89. Менаске, Д., Алмейда, В. Производительность Web-служб. Анализ, оценка и планирование. – К: Диасофт. – 2003.
|
| id | nasplib_isofts_kiev_ua-123456789-14657 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 1727-4907 |
| language | English |
| last_indexed | 2025-11-24T15:04:58Z |
| publishDate | 2010 |
| publisher | Інститут програмних систем НАН України |
| record_format | dspace |
| spelling | Shekhovtsov, A. 2010-12-27T14:04:16Z 2010-12-27T14:04:16Z 2010 Interactive assessment of simulated service qualities by business stakeholders: principles and research issues/ Shekhovtsov, A.// Пробл. програмув. — 2010. — № 2-3. — С. 288-298. — Бібліогр.: 89 назв. — англ. 1727-4907 https://nasplib.isofts.kiev.ua/handle/123456789/14657 681.03 We present the principles of an approach supporting the stakeholder involvement in a software process for service-oriented systems in a form of assessing the perceived quality of the software under development in its usage context. This method relies on interactive simulation of service performance and reliability; simulation models are parameterized by the factors influencing service execution; business stakeholders experience simulated service qualities in simulated usage contexts and assess this experience; the obtained assessments can be later used throughout the system lifecycle as a means of control for the quality of the software under development. Наведено принципи підходу, що підтримує участь зацікавлених осіб у процесі розробки сервіс-орієнтованих програмних систем у вигляді оцінювання сприйманої якості розроблюваної системи в контексті її використання. Цей підхід спирається на інтерактивне імітаційне моделювання продуктивності та надійності сервісів; параметрами імітаційних моделей є фактори, що впливають на виконання сервісів; зацікавлені особи висловлюють своє відношення до значень продуктивності та надійності, отриманих при взаємодії з імітаційними моделями якості сервісів у рамках виконання імітаційних моделей їх контекстів використання, надані оцінки можуть бути використані на різних етапах життєвого циклу програмного забезпечення як засоби контролю його якості. en Інститут програмних систем НАН України Методи та засоби програмної інженерії Interactive assessment of simulated service qualities by business stakeholders: principles and research issues Article published earlier |
| spellingShingle | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues Shekhovtsov, A. Методи та засоби програмної інженерії |
| title | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| title_full | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| title_fullStr | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| title_full_unstemmed | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| title_short | Interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| title_sort | interactive assessment of simulated service qualities by business stakeholders: principles and research issues |
| topic | Методи та засоби програмної інженерії |
| topic_facet | Методи та засоби програмної інженерії |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/14657 |
| work_keys_str_mv | AT shekhovtsova interactiveassessmentofsimulatedservicequalitiesbybusinessstakeholdersprinciplesandresearchissues |