A New Application of Domain Specific Modeling
Towards Implementing an Early Warning Service
Syed Imran, Ioannis Dokas, John Feehan and Franclin Foping
University College Cork, Cork Constraint Computation Centre
Western Gateway Building, Cork, Ireland
Abstract. Domain Specific Modeling (DSM) has been used for a while in many
scenarios, case studies, and applications successfully. This paper intends to present
a new application of DSM for facilitating an Early Warning Service (EWS). The
knowledge about risks and its representation is an important factor for any EWS.
The challenge is to help the users be closer to the understanding of the associ-
ated risks and to facilitate them with a new software application so as to provide
the ability of risk analysis, we propose in this paper a novel application of using
Domain Specific Modeling (DSM) in deploying EWS. The paper presents a case
study in which there is a need to develop an EWS for the personnel of institu-
tions involved in the drinking water delivery governance model in the Republic
of Ireland.
1 Introduction
Access to safe drinking water is a basic human right, but this right is not being met
universally. At least one third of the population in developing countries and almost one
fifth of the global population have no access to safe drinking water. Diseases related to
drinking water continue to be one of the major health problems globally as there are
about four billion cases of diarrhoea each year, out of which 2.5 million cases end in
death [1]. In particular waterborne disease outbreaks have attracted the attention of the
media and of the public, giving a raise on the discussions about the issues of drinking
water safety and risk management of water utilities.
The ultimate goal of our research, which is at its early phase, is to design and de-
velop a domain specific software application that will facilitate Early Warning Service
(EWS) development and is to be used by the personnel of institutions involved in the
drinking water delivery governance model of Ireland, such as the EPA, local authorities
(LA) and the Health Service Executive (HSE). It can be viewed as useful contribution
in supporting a proactive risk management strategy.
The goal of this paper is to propose a component and its high level architecture of
the EWS; that is a Domain Specific Modeling (DSM)application for modeling and rep-
resenting the knowledge about risks in WTP operations. Fig. 1, provides a model of the
interactions between the DSM application and the entities such as personnel of EPA and
WTP caretaker. One of the use cases, for example would be met when an EPA person-
nel is required to perform an audit on WTP using software application for EWS and in
order to do so it is necessary for him to receive input from the caretaker regarding with
Imran S., Dokas I., Feehan J. and Foping F. (2010).
A New Application of Domain Specific Modeling Towards Implementing an EarlyWarning Service.
In Proceedings of the 2nd International Workshop on Model-Driven Architecture and Modeling Theory-Driven Development, pages 99-104
DOI: 10.5220/0003049600990104
Copyright
c
SciTePress
Fig. 1. Functional requirements.
profiling the particular plant and updated working status on water treatment utility. The
possibility of automatically generating code from domain specific models can enhance
the development of the EWS.
The rest of the paper is organized as follows: Section 2 provides the brief description
of the relevant concepts of DSM . In Section 3 a proposed architecture of our novel ap-
plication of DSM will be presented. In Section 4, related work is high-lighted followed
by concluding remarks in Section 5.
2 Concepts
Each of the concepts mentioned below and their research areas has its own substantial
body of literature. In this article, we just scratch the surface of each.
2.1 Knowledge Modeling
Modeling in the context of information technology aims at capturing relevant knowl-
edge in a conceptual, often high-level, and abstract form. An ideal situation from a pure
knowledge perspective is to be able to reuse the knowledge available implicitly and ex-
plicitly, which can be accomplished by the effort of knowledge modeling. The activity
of knowledge modeling is to generate knowledge models, which are interpretable mod-
els that represent the concepts or entities of a real world, their behaviour, properties and
related constraints. Knowledge models are structured representations of knowledge and
they maintain knowledge in the form of symbols, entities, events with states, attributes
and values, in order to represent the pieces of knowledge and their relationship.
The elicitation and modeling of knowledge for engineering tasks, especially re-
lated to a specific domain, has always been a matter of concern. General purpose pro-
gramming languages represents knowledge in an object oriented style, supports think-
ing in terms of classes, objects and methods instead of depiction of acquired specific
knowledge in terms of their domains. Therefore, for knowledge acquisition activities
the knowledge modeling is used as a means of acquiring, validating and storing knowl-
edge for future use.
100
In model-driven paradigm of DSM metamodels are used to define various modeling
languages that represent the key concepts and abstractions for a specific domain. The
domain models, built by the set of processes and programming tools in DSM environ-
ment are formal models which correspond to entities in the real world and conceptually
represents its behaviour, constraints.
2.2 Domain Modeling
Utilizing the DSM approach can be achieved by raising the abstraction level and by un-
derstanding the concepts of specific domain. The definition of the modeling language
will be based on domain concepts identified during domain analysis [2]. The identified
domain concepts of the risk analysis methodologies for drinking water utilities are re-
quired to be mapped to the domain semantics, which is a set of rules that specify the
well-formedness of domain models and are represented by means of graphical nota-
tions; these notations represent the graphical syntax to express the semantics.
For capturing the syntax of the modeling language, a metamodel must describe all
the entities and relationships that may exist in DSM [3]. Metamodels provide object-
oriented specification of the abstract syntax which is the structure of the language
constructs used in Model Driven Engineering where models are primary artifacts. In
other words, metamodeling can be defined as a modeling activity for specifying mod-
els, where models are abstract representations of the real world entities, processes thus
describing the information that concerns with a particular domain. The high level con-
ceptual architecture for defining models for DSM is given in Fig. 2.
In order to complete the DSM definition specifying domain concepts and domain
models is not sufficient enough but domain rules and constraints are also required to
define the legal relationships between domain concepts and their usage. Different kinds
of constraints will be applied to the proposed DSM such as operational constraints
expression for modes of operations relevant to failure preventive measures, compos-
ability constraints expression for compatibility between different alternatives, resource
constraints will be used to indicate the need for specific resources and performance con-
straint [3]. In this consideration of DSM, the constraints and behaviour of DSM will be
defined in accordance to EWS by keeping the integrity of the domain concept.
2.3 Code Generation
In DSM the code is required to be generated from the domain models in order to drive
working software application, a database schema or even some documentation [4]. Fully
functional executable code can be generated from the domain models itself without any
manual intervention. During transformation no manual mapping is required from do-
main models to code; instead automatic code generation is facilitated by implementing
any technology that provides model to text transformations. In Eclipse, Model to Text
transformations (M2T)[5] project focuses on code generation and hosts frameworks
mostly based templates with an ability of generating text from models.
101
Fig. 2. Conceptual architecture of DSM.
3 Proposed High Level Architecture
In this approach, we stress on providing a new standalone DSM application for a EWS
in WTP. Eclipse Rich Client Platform (RCP) will be used to develop it. Fig. 3 illustrates
the higher level architecture and the core components of our system can be described
as follows:
Fig. 3. High level Architecture.
Prospective users include authorized stakeholders, personnel and managers of WTP,
EPA and HSE;
The presentation layer comprise of RCP platform in order to provide a standalone
application. Domain specific editor will serve as user interface for the prospect user
to interact with EWS;
102
An application layer will provide the core logic for different processes. Risk analy-
sis to calculate the likelihood of vulnerabilities. Monitoring mechanism will assist
the users such as plant manager to check, oversee the current status of water treat-
ment utilities. Profiling mechanism will allow the update of physical inventory and
to list what exist in specific plant;
The persisting layer will handles all the logic to save and retrieve the data from
data base. In this layer object-relation mapping is performed to transparently maps
objects to relational data.
4 Related Work
A useful aspect of DSM is in providing the convenience of identifying the concepts
representing the real entities of WTP by utilizing custom graphics within the models.
Hence, offering the convenience to think closer in terms of problem domain and even-
tually providing an effective EWS. The specified domain models can be used to define
other models or to configure other systems perhaps in combination with other enabling
technologies by providing the domain specific application process interface for EWS.
We argue that it is feasible to develop a novel type of software application to provide
EWS.
DSM approach has been successfully applied in a vast array of different domains
to create applications for a broad collection of programming languages and platforms
including many industrial applications [6] such as in insurance sector, mobile phones,
and radios. Their comprehensive study comparison is out of the scope of this paper.
The most recent relevant work we have found is on developing a prototyping ap-
proach using DSM for describing the earthquake detection algorithms for an earthquake
EWS [7],[8],[9].
DSM approaches have been used also for risk and vulnerability modeling in do-
mains such as critical systems engineering [10], grid computing systems security [11].
These approaches are defining the concepts of ”threat” ”risk” and ”vulnerability” in an
appropriate manner but do not integrate the concept of early warning signal, which is
of important in every EWS.
5 Conclusions and Ongoing Research
This paper proposes the use of a DSM in EWS for drinking water utilities. Our in-
novative approach consists of providing DSM to model the risk associated for water
treatment utilities and to describe the architecture of EWS application. However, there
are many technical research challenges that need to be resolved to define the early warn-
ing methodologies within the drinking water treatment utilities and their deployment in
our domain models along with their constraints. These include the modeling of con-
straints as well as the behaviour of the entities of water treatment utilities with respect
to the rules and techniques involved in risk management. Further different aspects of
DSM implementation need to be investigated such as code generation, improved per-
formance, and validity of the models with respect to the domain.
103
Acknowledgements
This work is supported by the research project SCEWA (Grant No 2007-DRP-2-S5),
funded by the Irish Environmental Protection Agency under the DERP grand scheme.
References
1. World Health Organization (WHO).: The world health report 2002 - Reducing Risks, Pro-
moting Healthy Life (2002)
2. Klein, M. R., and Methlie, L. B.: Knowledge-Based Decision Support Systems with Appli-
cations in Business. 2nd edition, John Wiley & Sons, New York (1995)
3. Lin, Y., Gray, J., and Jouault, F.: DSMDiff: a differentiation tool for domain-specific models.
European Journal of Information Systems, Palgrave Macmillan 16 (2007) 349–361
4. Gronback, R. C.: Eclipse Modeling Project: A Domain-Specific Language (DSL) Toolkit.
Addison-Wesley Professional (2009)
5. Model to Text transformations (M2T) project, http://www.eclipse.org/modeling/m2t
6. DSM Case Studies and Examples. DSM Forum, http://www.dsmforum.org/cases.html
7. SAFER - Seismic eArly warning For EuRope, http://www.saferproject.net/index.htm
8. Sadilek, D. A., Scheidgen, M., Wachsmuth, G., Weißleder, S. and den Linden, U.: Towards
Agile Language Engineering, Humboldt-Universitt zu Berlin (2008)
9. Wachsmuth, G.: Modelling the operational semantics of domain-specific modelling lan-
guages, Generative and Transformational Techniques in Software Engineering II. (2007)
06-520
10. Normand, V. and F
´
elix, E.: Toward model-based security engineering: developing a security
analysis DSML, Security in Model-Driven Architecture. (2009) 22
11. Flechais, I. and Mascolo, C. and Sasse, M. A.: Integrating security and usability into the
requirements and design process, International Journal of Electronic Security and Digital
Forensics 1 (2007) 12–26
104