AN INTEGRATED DECISION SUPPORT TOOL FOR EU
POLICIES ON HEALTH, TRANSPORT AND ARTISTIC
HERITAGE RECOVERY
Kanana Ezekiel and Farhi Marir
London Metropolitan University
Department of Computing, Communication Technology and Mathematics
2-16 Eden Grove
London N7 8DB
United Kingdom
Keywords: Enterprise Application Integratio
n, Information Integration and sharing, Integrated software
Abstract: In this paper, we describe an ongoing EU funded project (ISHTAR) that dev
elops an advance integrated
decision support tool (ISHTAR suite) for the analysis of the effects of long-term and short-term policies to
improve the quality of the environment, citizen’s health and preservation of heritage monuments. From the
background of the project, this paper goes on to explain the integration of information of a large number of
tools aimed at semantic approach integration to allow European cities to make balanced decisions on a wide
range of issues such as health, noise, pollution, transport, and historical monument. We also present and
suggest solutions to the information integration problem resulted when attempting to represent and share
information.
1 INTRODUCTION
European cities are faced with common challenges
concerning their quality of life ranging from
degradation of the urban environment, significant
risks for citizens health, traffic congestion and
economic inefficiency to progressive damage of the
artistic and monumental heritage. The process of
integrating such issues across organisation
boundaries is complicated and multidimensional
one. One difficult is derived from the lack of
integrated decision support tools that allow cities to
make balanced decisions on a wide range of issues
such as evaluation of the impacts of various types of
urban policies, actions on the quality of life of
citizens and in particular on traffic congestion, air
quality, citizen health and preservation of cultural
heritage.
The integration of information across
o
rganisation boundaries is a topic of interest to many
enterprises. According to some researchers (Fensel,
2003), (Butler, 2002) between 10% and 30% of
future IT spending will go towards Enterprise
Application Integration (EAI). Application
integration defines a core mission of IT
organizations today (Linkin, 2004). There are
several reasons for the increase in the need of
integration. New systems need to be integrated with
the legacy systems, and organisation or company
mergers require integration of software tools to
overcome problems in heterogeneity between data
sources. Also information sharing is a primary
emerging function (Figallo, 2002). The continue rise
in business opportunities and globalisation requires
capabilities that enable information integration to
ensure diffusion of information across enterprises.
By integrating tools through a middleware,
information can be accessed and shared across
organisational boundaries.
The Ishtar project aims to overcome the problem
of het
erogeneity between data sources by integration
of data sources or software tools using a central
integration manager called Ishtar Suite Manager
(ISM). The ISM manages and monitors processes of
software tools inside the suite. The Ishtar Data
Model (IDM) is proposed as a class/object model.
Each data corresponds to an object belonging to a
class defined in the metadata model. The metadata
model is managed in the ISM by means of a
database. Classes and objects are stored inside the
database. Objects inside the database represent the
actual data exchanged. The Ishtar Suite Interface
(ISI) for mapping users and the ISM is also
considered.
261
Ezekiel K. and Marir F. (2005).
AN INTEGRATED DECISION SUPPORT TOOL FOR EU POLICIES ON HEALTH, TRANSPORT AND ARTISTIC HERITAGE RECOVERY.
In Proceedings of the Seventh International Conference on Enterprise Information Systems, pages 261-266
DOI: 10.5220/0002522702610266
Copyright
c
SciTePress
In view of the above, this paper presents the process
of building an integrated tool for managing and
sharing information in the Ishtar project. We present
the specification of the architecture of the Ishtar
software tool. We also review some applications in
recently developed field of EAI. The paper is
organised as follows: In section 2, we present the
review of the state of the art and then in section 3 we
introduce the Ishtar Suite. In section 4, we present
the information integration problem and suggest
some solutions to the problem. Finally we provide
some conclusions in section 5.
2 STATE OF THE ART
With globalisation, organisations have become more
interconnected. The concept of EAI appeared only a
few years ago and it is now widely used and
explored because of the necessity for organisations
to interconnect their applications and provide new
functionalities to their users. The focus of EAI is on
process integration, structured data, and getting
applications to talk to each other. Although
according to InfoWorld (InfoWorld 2002) EAI
concept is still a tough and expensive nut to crack
despite advances in technology, it is being
implemented in a number of organisations (
Lejeune,
2003). In this section we observe some of the
achievements in the commercial sector and provide a
comparative evaluation with our approach.
WebLogic (BEA, 2004) is an integration server
with a full-featured platform for delivering
enterprise-class applications. WebLogic Integration
server not only provides data and application
integration but also business-to-business integration
and business process management. WebLogic
application Integration server is based on the J2EE
Connector Architecture (J2EE-CA) framework and
XML-based transformation engine. According to
Vijay Sarathy, product manager for Connectors and
JDBC at sun, the J2EE-CA is trying to extend what
JDBC has done for Java and J2EE in terms of
providing uniform connectivity with databases to
heterogeneous database systems. This application
framework links existing systems with new
applications. WebLogic supports leading enterprise
applications and technologies such as Oracle
Applications, MQ Series, SAP, Siebel, etc. It
provides a set of J2EE-CA to integrate with back-
end systems.
BusinessWare (Vitria, 2003) aims its product’s
integration capabilities at technical users.
BusinessWare is designed for application as well as
business-to-business integration. It offers dozens of
adapters for application integration, a connector
development kit, publish/subscribe messaging layer,
a transformation and data mapping tool, business
process management tool, etc. BusinessWare also
embeds rational Rose modelling tool for building
business processes.
Other applications that are close to WebLogic
and BusinessWare include WebSphere, SeeBeyond,
and ActiveEnterprise. WebSphere (Canon, 2004)
provides an EAI design and business-to-business
integration. WebSphere offers a strong
administration and deployment of tools. It not only
provides application integration but also a common
platform for development. This is particularly
appealing to companies that have a uniform platform
to develop their applications. SeeBeyond provides a
flexible way of linking applications and different
parts of a customer's supply chain independent of the
applications being connected. ActiveEnterprise is
designed for enterprise application integration and
business-to-business integration. It offers an end-to-
end integration product that aim at business users
with some technical knowledge. ActiveEnterprise
also offers packaged adapters for integration with
major databases and business systems such as Oracle
applications, SAP, Baan, PeopleSoft, J.D. Edwards,
etc.
Although the EAI technology can effectively
move data between applications in real time, it does
not define an aggregated view of the data objects or
business entities. EAI is less adequately addressed
the need to merge data and information across the
enterprise. Along with the movement to real time,
the need to integrate different kinds of information
has become more important. The Enterprise
Information Integration (EII) does address this issue.
Although there has been a flurry of research
activities (Bergamaschi, 2001), (Mitra, 2001), (Noy,
2000), (Stefan, 1998), etc, in the EII area, still a lot
has to be done especially on the semantic level
integration. The work that most related to ours based
on semantic integration is the one presented in
(Bergamaschi, 2001). The approach of the MOMIS
project (Bergamaschi, 2001) is a semi-automatic
approach to schema integration. MOMIS has a
single mediator, which provides a global data
schema and query interface to users. Data sources
with an ODL wrapper are connected to the
architecture. Differently from our approach, the
Description Logics is used as a kernel language to
interactively set-up the Thesaurus. In our approach,
the structure and context of classes in the schema are
defined using the XML language transformation.
Also the MOMIS approach has not been used in any
major industrial projects and is mainly an academic
research activity.
Another interesting work that is built using
object integration technologies is in (Stefan, 1998).
Stefan focused the attention on suitable architectural
abstraction concepts of components and connectors
ICEIS 2005 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
262
for systems. They propose connectors as pattern-like
transferable abstractions of system level component
interconnection and interoperation.
Unfortunately the use of the above commercial
applications is not contemplated. Most of these
application integration approaches are complex,
costly, and inflexible and require additional design
and development efforts. Moreover, they are aimed
at large application integration in the enterprise. The
Ishtar suite is developed using ideas from both EAI
and EII technologies.
3 INTRODUCING THE ISHTAR
SUITE & THE ARCHITECTURE
The Ishtar project involves seven European cities
(Rome, Bologna, Athens, Paris, Brussels, Graz and
Grenoble) and a number of technical and scientific
organizations (ENEA, WHO, INRETS, TRAC,
ARIA, PHAOS, ASTRAN, etc) having different
roles in the integration process of the necessary tools
and other data sources for the Ishtar Suite.
The integration process includes both old and
new software tools. The suite supports a large
number of software tools including TRANSCAD,
TISIS, COPERT, dBEL, CALINE, PCRAMMET,
WOCCS, SIMT-SSP, ADMS-Urban, SATURN,
TRANUS, VISUM, METROPOLIS, TEE-
TURBAN, KEMIS, etc. Some of the tools need
connection to Microsoft office tools (Excel, Access,
etc), GIS and other tools such as Logical Decision
for Windows, Hielow, etc. Although most tools run
on PC under windows operating system, some run
on workstations under UNIX environment causing
additional difficult.
The architecture of Ishtar Suite (Figure 1)
consists of the following:
Work packages with connector tools for
sending and receiving information to the
ISM.
The Ishtar Suite Manager (ISM) for
managing the processes inside the suite.
The user interface also known as Ishtar Suite
Interface (ISI) for connecting users to the
Ishtar Suite.
3.1 Work packages with connectors
The integration process is based on the seven
distinct steps of evaluation of transport policies
impacts on health and monuments, defined by the
following Work Packages:
WP1: Urban Policies and Actions selection
and analysis
WP2: Integrated transport Analysis Module
WP3: Integrated transport Direct Impacts
WP4: Pollutant and Noise Dispersion
Module
WP6: Advance health Impact Module
WP7: Damage reduction on Artistic
Heritage
WP9: Overall Analysis of case studies
Each work package tool appears as a separate
software tool running its own data input, providing
its own output and possibly having its own user
interface. Using Connectors, work packages are
linked to the ISM (Section 3.2) for exchanging
information with other work package. This allows
information to be shared among applications.
Connectors are pieces of software or interface
components that are used in the integration of
applications and serve as a "wrapper" that mediates
access to an application that was not developed with
integration in mind. Connectors are used to move
data in and out of applications. This is similar to the
ODBC in relational databases, which allow unique
interface to each vendor of RDBMS. In the Ishtar
suite, a Connector performs two kinds of operations:
The management of exchanges between ISM
and tool in work packages. All data used by
work package as input, output or internal data
are sent to ISM by the Connector on the form of
object corresponding to predefined classes of
metadata model. A connector communicates
with the ISM by means of messages in XML
format. Messages are sent to ISM for each
request of data.
The management of execution of work package
tool under the control of ISM. A message may
be sent by ISM to start a work package tool, it is
the connector’s job to prepare the definition of
the object used by the work package tool and
then send to ISM. After approval by ISM, the
work package tool is launched. Two cases are
considered when launching the work package
tool:
o One case is when the work package
tool provides its own user interface.
The data provided by user will be
entered through work package tool’s
own interface. The connector carries
out the management of input data
building corresponding parameters
then send to ISM the object
corresponding to the parameters.
o The other case is when there is no user
interface. The data provided will be
entered through the ISI. Users
parameters are built by ISI and ISM
and then provided to the connector.
AN INTEGRATED DECISION SUPPORT TOOL FOR EU POLICIES ON HEALTH, TRANSPORT AND ARTISTIC
HERITAGE RECOVERY
263
3.2 The Ishtar Suite Manager (ISM)
The ISM performs the following functionalities:
Management of data and information
exchanged between work packages
Management of the current scenario
execution, keeping trace and archiving of
scenarios
Control of the running process in the suite
Monitoring processes of each work packages
Control Ishtar Suite Interface (ISI)
Store data that is run in each scenario in the
suite to a repository called Ishtar Suite
Database (ISDB). Also a GIS tool called
Ishtar Suite GIS (ISGIS) is used for the
management of geographical data and map
displays.
Most applications in the EAI domain use message
exchange communication to send and receive
information (David 1999), (Longo, 2001). The
principle of message communication is that a
message provides more than simple data. A message
contains data and information on action requested by
a tool. A Connector undertakes the connection and
message exchange of the application to the ISM.
Whenever an application needs to be integrated a
connector is added to a work package. Applications
are no longer tightly coupled. Also there is no longer
a bridge via one specific single user interface. The
ISM not only controls the processes but also
manages the information. Clearly information
management in this context is an extremely difficult
activity. In particular there is a need for managing
information transformation. Further details are
articulated in the following subsections.
3.2.1 Management of messages, objects and
data in ISM
All message or data exchanged are described as
objects belong to classes defined as metadata, part of
Ishtar Data Model. The kind of data that ISM
manages includes traffic data, population data,
network data, Meteo data, etc. Data management is
according to a transformation or mapping rule. To
achieve transformation objects must refer to one of
the classes defined in the IDM. The ISM contains a
database (ISDB) with a global schema. The global
schema is composed of hierarchy of global classes
defined for exchange. In each class, attributes are
identified and defined with data types in XML
format. Figure 2 shows the management of
messages, objects and data in the ISM:
Message management gets the objects sent
by tools with associated request.
Object management checks that the object
belongs to a class stored in the class database
and stores the object in the objects database.
3.2.2 Management of processes
The ISM launches a software tool on the request
from user via the interface. It then sends a launching
message to the connector tool. The ISM waits for the
response of the tool. After receiving message ISM
lock all other processes. The ISM is then control the
execution of the tool by means of the exchanged
messages. At the end of the execution of the tools
process, the connector sends a message of execution
end to ISM. ISM sends the information to the user
and unlocks the other processes. Figure 3 shows
management of processes in the ISM phase.
Ishtar Suite Mana
g
er (ISM)
Figure 1: The Architecture and Work Packages (WPs) with Connectors (C1 – C9).
ISDB
Ishtar Suite Interface
(
ISI
)
ISGIS
WP1
C1
WP2
WP2
C2
WP3
WP3
C3
WP4
WP4
C4
WP6
WP6
C6
WP7
WP9
WP7
C7
WP9
C9
ICEIS 2005 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
264
Message management
3.2.3 Management of transformation of data
The ISM undertakes the transformation of data
coming from work package tools and requested by
the current tool. There are two kinds of
transformations. The first one concerns the content
of data. In some cases data requested by tool is not
directly provided by one tool rather a combination of
several data provided by several work package tools.
The ISM has a base rules for transformation to
assure that all data requested are completed from
relevant sources. The other transformation concerns
the format of data so that it fits the expected format.
ISM uses XML format as a core format for
exchanges of data inside the suite.
3.3 Ishtar Suite Interface (ISI)
The ISI is part of Ishtar suite through which user
controls the execution of processes in the suite. The
ISI provides an interactive window where user will
have a set of menus to control the process of Ishtar
Suite. Each menu corresponds to one of the seven
modules defined as work packages. The ISI opens
specific interface window when a module of the
suite is activated. It also provides an interactive
window for each activated module.
The combination of ISM, ISI and connectors
create a powerful architecture for information
integration in the Ishtar Suite. The architecture now
can flow from the ISI to WP9 tool to the ISM then
from ISM to WP3 tool. This makes the Ishtar Suite
adaptable to any existing software tools and does not
force them to conform to any particular form that is
existing tools. It is also designed to embrace new
applications or technologies as they emerge.
4 INFORMATION INTEGRATION
A collection of information or data is needed to
support most decision-making processes.
Information is a central part of any decision support
system. Heterogeneous information drawn from
internal operational tools as well as external data
sources is mapped to the information specified in a
global schema for exchange. Our approach follows
the semantic paradigm, in that a global conceptual
schema is involved and schema objects (sharable
information) are described using a common Ishtar
Data Model (IDM) and XML language. IDM is
defined as subset of Object Data Model. Information
integration is described in detail in (Lejeune, 2003).
4.1 The information Integration
Problem
When sharing knowledge between applications,
integration problems are compounded. Each
Ob
j
ect
Data management
In
p
ut data to tools
Transformation
rules
Classes Objects
Common data
Out
ut data from tools
Fi
g
ure 2: Mana
g
ement of messa
g
es, ob
j
ects and data in ISM.
User
launch WP
Send a launch message to
WPs
WP tool
Wait message from WPs
Manage messages from WP tool
Receive launch start message
Control execution of WP tool
Lock all
other processes
Messages
Info to the use
r
Unlock
p
rocesses
Treat error
Fi
g
ure 3: Mana
g
ement of
p
rocesses in ISM.
User
AN INTEGRATED DECISION SUPPORT TOOL FOR EU POLICIES ON HEALTH, TRANSPORT AND ARTISTIC
HERITAGE RECOVERY
265
organisation has its own set of applications,
technologies and standards. Information integration
is hindered by differences in software and hardware
applications and also by syntactic and semantic
differences in various data and data sources. This is
a recognized problem in the area of EAI where many
applications have been created for the purpose of
information integration. When broadening the scope
of the information integration problem to cross-
organisational boundary, the problem becomes even
more severe because within an organisation there is
certain amount of control over the applications used
by one organisation, while there typically is no
control over the applications used by other
organisations.
What further complicated the matter is the fact
that people who are involved in the integration
process from existing applications may sometimes
have little or no knowledge of the data sources or
software tools that are being used which make
mapping a hard problem.
4.2 Solving Information Integration
Problem
Most current tools, however, only addresses the
problems of platforms and syntactic heterogeneity.
The problem of semantic heterogeneity that is the
differences in the meaning of concepts is, in current
solutions reduced to syntactic rewriting without
explaining the meaning of data. In the Ishtar project,
we try to provide a solution for the information
integration problem across organization boundaries.
We use the Ishtar suite to create integration manager
ISM with a well-defined common Ishtar Data Model
(IDM). The ISM maps data schemas taken from
different sources to the IDM in order to make the
meaning of the concepts explicit. Here the ISM
enables information integration on a semantic level,
but there are still many issues to resolve, such as
standardisation in languages for specification of
mapping rules and the development of reusable,
industrial-strength tools for creation and operational
of these mapping. Standards are arising to mitigate
these problems, but current approaches are specific
for certain organisations and are typically very rigid
in what require from anyone using the standard.
5 CONCLUSION
We have presented the architecture of integrated tool
for the Ishtar project. The architecture is flexible,
allowing the integration of interdependent work
packages. We base our architecture on the main
principles of EAI and EII domains, namely the use
of an integration manager, constituting the core of
the suite, managing the control of processes and data
exchange between software in work packages. The
manager also communicates with all tools by means
of messages and manages exchange of data based on
XML file format. Specific tool called ‘Connector’ is
developed by each work package for
communication.
REFERENCES
BEA, 2004. “Introducing BEA WebLogic Integration”,
Version 8.1 Service Pack2
Linkin, P. 2004. “The Convergence of Application
Development and Integration”
Vitria, 2003 “The leading platform for Agile Business
process integration”, BusinessWare 4
Fensel, D 2003. Ontologies: Silver Bullet for Knowledge
Management and Electronic Commerce, Springer-
Verlag
Lejeune, B. and Flavigny, P. O. 2003, Fifth framework
Programme, Thematic Programme: Energy
Environment and Sustainable Development – Ishtar
Project
Canon, J. 2004 IBM: WebSphere Business Integration
Brokers
Fagillo, C. and Rhine, N. 2002, “Building the Knowledge
Management Network”
Butler, M. 2002 “What’s Better: Application Integration
or Integrated Applications?”
InfoWorld 2002 “Application Integration Survey of IT
leaders”, http://www.infoworld.com/
Mitra, P. and Wiederhold, G. 2001 Algebra for semantic
interoperability of information sources. In IEEE
International Conference on Bio informatics and
Biomedical Engineering, pp174-182
Longo, J. 2001 “The ABCs of Enterprise Application
Integration”
Noy, N. and Musen 2000 Prompt: Algorithm and tool for
automated ontology merging and alignment. In
Proc.17
th
.Conference on Artificial Intelligence
Bergamaschi, S. Castano, S., Beneventano, D. and
Vincini, M. 2001, Semantic Integration of
heterogeneous information sources, Data and
knowledge Engineering, pp 215-249
David, S. 1999 Enterprise application integration
Stefan, T. 1998 “A Connector Model for Object-Oriented
Component Integration”
ICEIS 2005 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
266