SEMANTIC INTEROPERABILITY
Information Integration by using Ontology Mapping in Industrial Environment
Irina Peltomaa, Heli Helaakoski
VTT Technical Research Centre of Finland, P.O. Box 3, 92101 Raahe, Finland
Juha Tuikkanen
VTT Technical Research Centre of Finland, P.O. Box 1100, 90571 Oulu, Finland
Keywords: Interoperability, semantic technology, ontology model, ontology mapping.
Abstract: Interoperability requires two components technical and information integration. Most of the enterprises have
solved the problem of technical integration but at the moment they are struggling with information
integration. The challenge in information integration is to preserve the meaning of information in different
context. Semantic technologies can provide means for information integration by representing the meaning
of information. This paper introduces how to use semantics by developing ontology models based on
enterprise information. Different ontology models from diverse sources and applications can be mapped
together in order to provide integrated view for different information sources. The domain area of this case
study is heavy industrial environment with multiple applications and data sources.
1 INTRODUCTION
Today business processes must be managed and
modified more effective than ever before. Processes
need to adapt to consolidation, mergers and
acquisitions, joint ventures, divestitures, regulatory
compliance issues, shifts in business models,
changing customer expectations, industry
standardization and business process outsourcing
(Smith and Fingar, 2002). Successful management
of a company requires overall management of all
processes and information connected to them. All
changes in business processes have influence on
several information systems in enterprises and
therefore seamless communication and integration of
data and information as well as synchronization of
inter-organizational business processes are complex
problems (Li et al., 2006). Enterprises are using
huge amounts of time and money to solve
interoperability problems of software systems. At
the moment enterprises are using 30 to 40 % of their
IT budget for integration problem (Gartner) and the
average time spent for integration of single interface
is 2 PM. The most worrying thing is that some
enterprises spend up to 80 % of their IT budget for
updating their legacy systems to integration and
extensions (Polikoff and Allemang, 2003). As the
enterprise collaboration is increasing al the time and
requiring more and more integration, the costs of
inadequate interoperability will be unsustainable.
Interoperability is comprised of both technical
integration and information integration. Most of the
current solutions are focused only on technical
integration, to link disparate software systems to
become part of a larger system while information
integration is focused on preserving the semantics
while transforming the context (Pollock, 2001). Yet
any moderately complex integration work requires
both types of integration. There is two strategies for
migrate system integration problem with the
enterprise: the development of an enterprise
message model as a reference point for flexible and
economic integration and the use of a semantic
broker so that each application would not have to
understand the semantics of every other application
(McComb, 2004). Today, enterprises are using the
message model as a technical integration solution
although the architecture still has many unsolved
problems like laborious configuration and
differences in the information interpretation between
designers. Examples of such technical solutions that
use common transportation layer for integration are
465
Peltomaa I., Helaakoski H. and Tuikkanen J. (2008).
SEMANTIC INTEROPERABILITY - Information Integration by using Ontology Mapping in Industrial Environment.
In Proceedings of the Tenth International Conference on Enterprise Information Systems - DISI, pages 465-468
DOI: 10.5220/0001702504650468
Copyright
c
SciTePress
IBM WebSphere, Microsoft Messaging Queuing and
BEA ESB.
Information integration is based on semantic
interoperability that emphasizes the importance of
the information and focuses on enabling content,
data, and information to interoperate with software
systems outside their origin (Pollock and Hodgson,
2004). The aim of semantic interoperability is to
provide the ability to bridge semantic conflicts
arising from differences in implicit meanings,
perspectives, and assumptions for co-operating
enterprises, thus creating a semantically compatible
information environment based on the agreed
concepts between different business entities (Park
and Ram, 2004). The basis for ontology-driven
knowledge management was formed in EU projects
On-To-Knowledge (On-To-Knowledge, 2008) and
continued in Semantic Knowledge Technologies
(SEKT) (SEKT, 2008) that researched the
development and exploitation of semantic
knowledge technologies.
This paper proposes the use of semantics for
solving the interoperability problems above
technical integration. The presented work is
conducted in Sebi-project (Sebi, 2008) that
concentrates on solving interoperability problems in
industrial environment by using semantic
technologies and especially ontology-based
knowledge management. This paper presents the
architectural approach of using semantic
technologies in Sebi project. The presented approach
is tested in a case study of integrating information
from various sources in industrial production
environment. This paper gives a brief introduction to
the objectives of the case study and overviews the
basic elements of the development process.
2 CASE STUDY OF
INFORMATION INTEGRATION
Interoperability between systems can be supported
with semantic technologies by developing a shared
information model for the use of different solutions.
This model provides an integrated view for
heterogeneous data sources within a company.
Integrated view can be composed several ways
according to the requirements. It can support
business process management (BPM) or different
task-specific purposes within BPM like controlling
the delivery reliability or lead time. Semantic
models can also be used in enterprise collaboration.
Using semantically enriched data different parties
having divergence in ontologies are able to
communicate with each other; communication can
involve applications, information systems,
processes, companies in value-chain and humans.
The semantic interoperability research has
categorized three broad research areas: mapping-
based, intermediary-based, and query-oriented
approaches (Park and Ram, 2004). Mapping-based
approach attempts to construct mappings between
semantically related information sources while the
intermediary-based approach may also rely on
mapping knowledge established between a common
ontology and local schemas. Query-oriented
approach is focused on interoperable languages
which can be used for formulating queries over
several databases. Semantic architecture
methodologies are divided into three groups: one-to-
one mapping, single shared ontology and ontology
clustering (Alexiev et al., 2005). The methodologies
have different approaches of using global and local
ontologies, one-to-one paradigm uses local
ontologies alone, single-shared ontology use global
ontology without local ontology and mix of single-
shared and one-to-one mapping uses global ontology
with local ontologies (Bruijn and Feier, 2005).
According to Sebi view the interoperability can
be best achieved by combining all three different
approaches of interoperability: mapping-based,
intermediary-based and query-oriented. In this case
study we have used mixed paradigm with one global
and several local ontologies. The interoperability
architecture is presented in Figure 1.
Integration ontology
Common conceptual model
Various
concept models
Source data
DB 2DB 1
Source application
DB C
Mappings
Source data
DB DDB C
XML
Access A Access CAccess B
Business layer
Integrated access to
heterogeneous source data/
applications for users
(production managers, etc.)
Semantic Layer
Data Layer
Figure 1: Sebi interoperability architecture.
In data layer is the source data from
heterogeneous data sources and applications. The
source data is derived from various databases or
other electronic sources (eg. XML-files, ontology
ICEIS 2008 - International Conference on Enterprise Information Systems
466
files). Semantic layer represents the concept models
of domain information derived from various data
sources. Integration ontology is a common
conceptual model that defines key concepts of the
domain with related relations and attributes.
Integration ontology is developed with ontology
building tool in collaboration with domain experts.
The mappings between concept models and
integration ontology are defined with the same tool
in order to connect the source concept models to
integration ontology. Business layer offers different
views for integrated information according to user’s
need. The information is requested by executing
queries into integration ontology by using
middleware tool. Run-time implementation of
queries enables real-time access to distributed
information and supports for example decision
making in company management.
The case study describes the process of
developing interoperability for industrial
environment and the case enterprise operates in the
domain area of heavy manufacturing. The aim of the
case was to integrate the information used in
manufacturing process control. The enterprise has
several information systems (ISs) for controlling
manufacturing process; each IS having a totally
different concept hierarchy, database structures and
information models. This posed constant problems
for information management and furthermore caused
extra costs and loss of resources. The case involved
three ISs, one containing failure and breakdown
information, another IS controlling machines and
devices and the third IS needed the information from
the previous two. The operating environment of the
case is presented in Figure 2.
Production
system
Product
information
application
ooo
Production, maintenance, planning, etc.
Integration
interface
Event management
application
(measurement
analysation and
processing)
Automation
system
Production
system
Production
system
Measurement data
(production time)
Failure and
breakdown
information
Maintenance
information
application
Automation
system
Automation
system
Figure 2: Operating environment of the case.
The objective was to integrate ISs together in
order to offer a single view into machine, device,
breakdown, and failure information to enable
proactive maintenance in manufacturing process,
planning of maintenance work and controlling the
lead time of maintenance. The main requirements
for the solution were to enable frictionless
information sharing, flexibility of the integration and
effective maintenance of the integration.
The technical integration of the case was solved
by using a message broker that transmits information
between ISs, databases and different data sources. In
the semantic solution source ontologies from
different databases are modeled using ontology
engineering tool. In this case OntoStudio (Ontoprise,
2008) tool was used for developing automatically
source ontologies, designing integration ontology,
defining mappings between source ontologies and
integration ontology and creating intelligent
reasoning. The development of integration ontology
was also developed by OntoStudio. It was a very
demanding phase and it was done in collaboration
with experts from the domain area. The integration
defines shared semantics of the data and contains all
the important concepts with related attributes and
restrictions from the domain area.
The connections between source ontologies and
integration ontology were realized using mappings
and rules. When using mappings, the concept from
source ontology is connected to the suitable concept
in integration ontology. In complex cases, where the
concepts are not correspondent, the mappings can be
made by using reasoning. The reasoning may be
based on similarity of concepts and the meaning of
concepts. In this case the rules were created using F-
logic (Kifer et al., 1995) as an inferencing language.
Reasoning enables upper level mappings, where
rules deliver information to a corresponding place at
the lower level. This makes the maintaining of
mappings easy, because changes in upper level
cause changes in the whole concept tree. By using
rules the amount of mappings can be reduced and it
is easier to piece mappings together. Figure 3
describes a rule that get inputs from source and types
all failures which “FailureType_id” is one to
mechanical failure to the business ontology.
Integration ontology operates as a link between
different ISs by offering access to the information.
The information is queried through integration
ontology which provides access to information in
databases through source ontologies. The queries are
executed to the integration ontology using a
middleware tool.
SEMANTIC INTEROPERABILITY - Information Integration by using Ontology Mapping in Industrial Environment
467
Figure 3: Rule for reasoning the type of the failure.
3 CONCLUSIONS
System interoperability is complex problem which
has become unbearable for enterprises in dynamic
environment. Constant integration work consumes
time and money that enterprises could otherwise use
for development of new systems. System integration
and the maintenance of existing point-to-point
integration solutions are consuming lion’s share of
the companies’ IT budget. By using semantics the
prerequisites to solve the interoperability problem is
significantly increasing.
This paper presents our approach of semantic
solution. During the research different architectures,
methodologies and tools for semantic
interoperability were examined and the most suitable
alternatives where chosen. Chosen ontology
architecture allows flexible adaptation to the
changes which take place in system combination.
The interoperability between systems is
implemented through integration ontology. The
information from existing systems is modelled isto
concept models that are mapped to integration
ontology. Integration ontology offers real-time
access to information in integrated systems for users
and systems. It also offers effective tool for human-
to-human communication. The functionality of the
semantic solution was examined in case for
manufacturing system interoperability.
The development of semantic solution is still on
its early stage. The development requires new
working methods, because new technologies require
different approaches and the old working methods
are not necessarily suitable. The research will
continue by defining the time and costs of ontology
development. Also the technical side of the solution
and rule creation process still needs development.
During the research it was realized that the wide
adoption of semantics is still in future. Considerable
amount of research has been done in the area of
semantic interoperability, but real implementations,
especially in industry, are few. The adoption of
semantic technologies requires hard evidence of the
functionality in real-life cases and quicker
implementation pace. Whole process need to be
handled, mere technical solution is not enough.
REFERENCES
Alexiev, V., Breu, M., Bruijn, J.D., Fensel, D., Lara, R.,
Lausen, H., 2005. Information Integration with
Ontologies: Experiences from an Industrial Showcase.
John Wiley & Sons.
Bruijn, J.D., Feier, C., 2005. D4.6.1.1 Report on ontology
mediation for case studies V1. SEKT-project Report
2005-08-05.
Kifer, M., Lausen, G., Wu, J., 1995. Logical foundations
of object-oriented and frame-based languages. Journal
of the ACM, vol. 42, no. 4, pp. 741-843.
Li, M.-S., Cabral, R., Doumeingts, G., Popplewell, K.,
2006. Enterprise Interoperability research roadmap
v4.0 31 July 2006.
McComb, D., 2004. Semantics in Business Systems.
Morgan Kaufmann Publishers.
On-To-Knowledge, 2008. On-To_Knowledge homepage.
[Online], Available: http://www.ontoknowledge.org/,
14.3.2008
Ontoprise, 2008. Know how to use Know-how Ontoprise
GmbH. [Online], Available: http://www.ontoprise.de/
content/index_eng.html, 14.3.2008
Park, J., Ram, S., 2004. Information Systems
Interoperability: What Lies Beneath?. ACM
Transactions on Information Systems, vol. 22, no. 4,
pp. 595-632.
Polikoff, I, Allemang, D., 2003. Semantic Integration,
Strategies and Tools. TopQuadrant Technology
Briefing.
Pollock, J.T., 2001. The Big Issue: Interoperability vs.
Integration, eAI Journal, October, pp. 48-52.
Pollock, J.T., Hodgson, R., 2004. Adaptive information.
Improving business through semantic interoperability,
grid computing, and enterprise integration. Hoboke:
John Wiley & Sons.
Sebi, 2008. Sebi-project Web site. [Online], Available:
http://www.vtt.fi/proj/sebi/ [14.3.2008]
SEKT, 2008. Semantic Knowledge Technologies
homepage. [Online], Available: http://www.sekt-
project.com/, 14.3.2008
Smith, H., Fingar, P., 2002. Business Process
Management: The Third Wave. Meghan-Kiffer Press.
ICEIS 2008 - International Conference on Enterprise Information Systems
468