SEMANTIC FRAMEWORK FOR INFORMATION

INTEGRATION

Using Service-oriented Analysis and Design

Prima Gustiené

Department of Information Systems, Karlstad University, 651 88 Karlstad, Sweden

Irina Peltomaa, Heli Helaakoski

VTT Technical Research Centre of Finland, P.O. Box 3, 92101 Raahe, Finland

Keywords: Service-oriented analysis and design, Enterprise modelling, Information integration, Semantic technologies.

Abstract: Today’s dynamic markets demand from companies’ new ways of thinking, adaptation of new technologies

and more flexible production. These business drivers can be met effectively and efficiently only if people

and enterprise resources, such as information systems collaborate together. The gap between organizational

business aspects and information technology causes problems for companies to reach their goals.

Information systems have increasingly important role in realization of business processes demands which

leads to demand of close interaction and understanding between organizational and technical components. It

is critical for enterprise interoperability, where semantic integration of information and technology is the

prerequisite for successful collaboration. The paper presents a new semantic framework for better quality of

semantic interoperability.

1 INTRODUCTION

Today’s rapid changes in markets force companies

to produce their products with better quality and

more flexibly, which results in necessity for

introduction of new technological solutions. The role

of information systems, as support for realization of

business process demands, becomes of great

importance. Traditionally there is a gap between two

communities; business administration professionals

and information technology experts. Business people

tend to consider technological issues as a

subordinate aspect in business process and

technology experts consider that business goals

issues do not deserve much attention (Weske, 2007).

As goals stated by business experts at the

organizational level should fit with the outputs from

implementation, it is necessary that all partners

involved have a common understanding of both

organizational and technical aspects. Growing

business enables growing of data. The business will

suffer service disruptions if there is no strategy how

to manage relevant information.

The fundamental problem with conventional

methods for information system development is that

they do not take into account some important

semantic interdependency types between static and

dynamic models, which are crucial for gluing

strategic, organizational and technical descriptions

into one computation independent and integrated

representation (Gustas and Gustiené, 2007). There is

a lack of integrated models and systematic methods

to support business process modelling across

organizational and technical system boundaries.

Semantic problems of communication between

business analysis and design experts lead to

ambiguous and incomplete system requirement

specifications as well as causes enterprise

interoperability problems (Sarjanoja et al., 2008).

In business modelling is an important to

determine how an information system contributes to

the objectives of the organization (Bennett, 2002).

Traditionally graphical representations of enterprise

architectures are constructed fragmentally and not

aligned with information system design. It causes

difficulties to maintain semantic integrity of multiple

enterprise architectural specifications (Gustas and

Gustiené P., Peltomaa I. and Helaakoski H. (2009).

SEMANTIC FRAMEWORK FOR INFORMATION INTEGRATION - Using Service-oriented Analysis and Design.

In Proceedings of the 11th International Conference on Enterprise Information Systems - Databases and Information Systems Integration, pages 64-69

DOI: 10.5220/0001950800640069

 SciTePress

Gustiené, 2008). The success of enterprise

interoperability much depends on how static (data)

and dynamic aspects of enterprises are integrated.

This integration enables preservation of the meaning

of information about the context (Sarjanoja et al.,

2008). The description of integrated service

architectures should be established before

implementation specific solutions are discussed.

The paper proposes an extended framework for

improving semantic quality of business processes

using service-oriented approach. It enables

integration of static and dynamic aspects of business

processes, facilitates information integration across

organizational boundaries and provides possibilities

to check consistency and completeness as well as to

track undesirable system qualities (Gustas and

Gustiené, 2004; Gustiené, 2003).

2 SEMANTIC ISSUES

Semantics i.e. study of meaning is the central part of

communication. We have to understand the meaning

of the message unambiguously in order to reach a

successful communication. Ambiguity is one of the

deficiencies of the natural and system modelling

languages, it causes misunderstanding. Ambiguity of

concepts in system modelling may occur because a

construct, formal expression or natural language

sentence has more than one meaning (Dori, 2002),

or because of incompleteness or inconsistency of

conceptual models.

Ontology captures consensual knowledge in a

generic way to be reused and shared across software

applications and by groups of people (Gomez-Perez

et al., 2005; Gruber, 1995). It defines a common

vocabulary for information sharing in a domain

(Noy and McGuinness, 2001; Uschold and

Gruninger, 1996). Creation of a business ontology

which is describing the semantics of the essential

concepts of company will offer better possibilities

for unified process management and system

interoperability. Through business ontologies it is

possible to view an integrated view of company’s

data (Pollock and Hodgson, 2004).

The most important issue in information system

development is how to manage its complexity.

According to Dietz (Dietz, 2006), complexity can

only be mastered under two conditions: to have a

comprehensive theory about the things whose

complexity one wants to master and the other

condition is that there are appropriate analysis

methods and models based on that theory. To

manage complexity it is necessary to have an

integrated method and a coherent, comprehensive,

consistent and concise conceptual model of the

enterprise. Semantic interoperability can be ensured

by providing contextual knowledge of domain

applications (Ram and Park, 2004). Interoperability

is comprised of both technical integration and

information integration (Peltomaa et al., 2008). The

main technical challenge is the lack of

interoperability of different systems and data sources

thus most of the current solutions are focused only

on technical integration, to link disparate software

systems to become part of a larger system.

Information integration is enabled by semantic

interoperability that emphasizes the importance of

information inside enterprises and focuses on

enabling content, data, and information to

interoperate with software systems outside their

origin (Pollock and Hodgson, 2004).Yet any

moderately complex integration work requires both

technical and information integration.

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 architectures for information

integration are divided within the methodologies

into three groups which are one-to-one mapping,

single shared ontology and ontology clustering

(Alexiev et al., 2005). These methodologies use

differently global ontology together with local

ontology. Either local ontologies are used alone

(one-to-one paradigm), or a global ontology exists

either without (single-shared ontology) or with local

ontologies (mix of single-shared and one-to-one

mapping) (Alexiev et al., 2005; Bruijn and Feier,

2005).

3 SEMANTIC FRAMEWORK

Enterprises need more effective way to manage

information related to their business. The

management of information includes communication

between personnel and the integration of

information in separate information systems.

SEMANTIC FRAMEWORK FOR INFORMATION INTEGRATION - Using Service-oriented Analysis and Design

The ambition of service orientation is to provide

system designers with a constructive way of

integrating business as a set of linked services. It is a

way of designing an integrated business process as a

set of loosely coupled services. Service architectures

can be used for specifications of business processes

in terms of organizational and technical services

(Gustas and Gustiené 2007; Gustas and Gustiené,

2008).

Mappings

Pragmatic level

Semantic level SOAD

Service-oriented approach presented for analysis

and design process (SOAD) (Gustas and Gustiené

2007; Gustas and Gustiené, 2008; Gustiené and

Gustas, 2008) has semantic power to conceptualize

organizational and technical system components, by

distinguishing intersubjective and objective views,

that facilitates better semantic integrity control

between static and dynamic aspects. The advantage

of such modelling is that it integrates semantics of

different aspects in one type of diagram. Conceptual

representation of service architectures define

computation independent aspects that are not

influenced by any implementation solutions and are

more comprehensible for business people as well as

system designers.

Sebi-framework (Peltomaa et al., 2008) defines

framework for information integration using

semantic technologies. Interoperability between

separate information systems is achieved by

developing a shared information model for the

information. Different views of information are

available for other information systems or humans

through shared information model which can be

called integration ontology.

In figure 1 proposition of combining these two

approaches is presented. The most important part of

the Sebi-framework is the development of

integration ontology, which is developed in close

collaboration with business experts to determine

correct concepts and their relationships. As a method

for integration ontology development Sebi-process

(Sarjanoja et al, 2008) is used. Completing Sebi-

process with SOAD the process of business

modelling is included in the framework and more

accurate integration ontology can be developed.

Integration is implemented by connecting concept

models formed from information in separate

information systems to the developed integration

ontology using mappings.

Extension of Sebi-framework with SOAD

approach will contribute to a better quality of

semantic interoperability, because this approach has

more semantic power in comparison with other

methods, to identify and control undesirable

semantic characteristics such as inconsistency and

incompleteness that lowers the quality of data.

Various

concept models

Syntactic level

Integration

ontology

igure 1: An extension of Sebi-framework with SOAD to

Being computation neutral service-oriented

ana

4 ANALYSIS

The primary goal of Service-Oriented Architecture

l system

com

ensure a better quality of semantic interoperability.

lysis facilitates better involvement of

stakeholders without deep technical knowledge in

the area of information system. In the following

chapters the components of this approach are

introduced more closely.

(SOA) is to align the business design with the

information technology (IT) innovations in order to

make both organizational and technical system parts

more effective (Gustas and Gustiené, 2007).

Business and IT solutions can be expressed using

graphical representations of Enterprise Architecture

that provides possibilities to understand and

determine the continual needs for changes.

To understand how and why technica

ponents are useful and fit to the overall

organizational system, then at least three levels of

information system models are necessary to take into

consideration for maintenance of a systematic

change. Three levels are represented in Figure 2.

ICEIS 2009 - International Conference on Enterprise Information Systems

Pragmatic Level

Business-Oriented

Analysis

Semantic Level

Service-Oriented

Analysis

Syntactic Level

Technology-

Oriented Analysis

Computation

Independent

Modeling (CIM)

Computation

Dependent

Modeling (CDM)

Figure 2: Three levels of Enterprise.

Pragmatic level is the businesses-oriented

analysis level, which is the most abstract one. This

level provides motivation behind new business

solutions. Goals stated at this level will be specified

further at semantic level. These specifications drive

and guide the activities at the semantic level.

At semantic level service-oriented analysis is

done. This level has capacity to describe clearly

static and dynamic structures of business processes

across organizational and technical system

boundaries. At this level semantic dependencies are

used for conceptual modelling, which provide

possibility to identify and overcome such

undesirable system characteristics as inconsistency,

incompleteness, redundancy of data, ambiguity and

incoherence (Gustas and Gustiené, 2004; Gustiené,

2003).

At syntactic level technology-oriented analysis is

done. This level defines implementation-oriented

details, which explain the data processing need of a

specific application or software component.

All three levels are interrelated as they define the

same artefact. The framework of three levels

provides the natural view to understand the

modelling artefact as a whole. It provides with a way

for semantic traceability via all three levels and

enables interplay between business needs and

technical solutions.

One of the advantage of the concept of service is

that it can be applied equally well to the

organizational as well as software components

which can be viewed as service requester and

service provider. Enterprise system can be defined

as a set of interacting loosely connected

components, which are able to perform specific

services on request.

Conceptual representations of service

architecture are defined by using one or more

interaction loops between enterprise actors that can

be viewed as organizational or technical

components. Modelling method using service as an

interaction loop or a composition of loops provides

the holistic view of an enterprise as a system.

The core elements of the service-oriented

modelling are actors, communication flows, and

actions. Actors (service requesters or service

providers) are the active elements of an enterprise,

the ones who initiates the actions. The structure of a

service as an interaction loop could be defined by

five phases or steps necessary for modelling of

service structure. They are as following:

1. Identification of interaction flows

2. Identification of actions

3. Identification of transition dependencies

4. Identification of attribute dependencies

5. Semantic integration

Identification of interaction flows and actions

represent intersubjective perspective of the

communication action, which is represented by

interaction dependency link. Identification of

transitions and attribute dependencies define the

objective perspective. It defines the state changes

that objects overcome when actions take place

(Gustas and Gustiené, 2008). Modelling of data can

not be done separately from process. These steps are

important for integration of static and dynamic

aspects, which facilitate reasoning and define the

holistic understanding of enterprise architecture.

Pre-condition

Object Class

Service

Provider

Service request

Service

request

Service

Requester

Intermediate

Object Class

Service response

Service

response

Post-condition

Object Class

Figure 3: Example of .

Service-oriented constructs used for service-

orie

one interaction loop

nted modelling is based on three events:

creation, termination and reclassification events

(Gustiené and Gustas, 2008). Composition of three

types of basic constructs provides possibility to

conceptualize the lifecycle of objects in a service

SEMANTIC FRAMEWORK FOR INFORMATION INTEGRATION - Using Service-oriented Analysis and Design

interaction loop. Objects define the data from

business environment which is necessary to integrate

and which is critical for forming an integrated

ontology. The example is represented in Figure 3.

Intersubjective and objective perspectives are

imp

5 SEBI-FRAMEWORK

The Sebi-framework (Peltomaa et al., 2008) uses

tegration ontology is the most

ucceed in integration ontology

dev

method for

bus

-process is

sta

ontology is physically created with

ont

tes as a link between

diff

6 CONCLUSIONS

Enterprise interoperability becomes a prerequisite

for successful business accomplishment and requires

ortant to distinguish for conceptualization of

organizational as well as technical parts of the

system. Ability to integrate these aspects in one

modelling notation provides possibility to control

and integrate static and dynamic aspects of the

system. Enterprise ontology models should clearly

define the semantic details about the state of

attribute values when creation, termination or

reclassification action takes place.

semantic technologies to enable information sharing

among separate information systems. The developed

framework enables the combination of all three

different approaches if interoperability: mapping-

based, intermediary-based, and query-oriented. As

ontology architecture mixed paradigm with one

global and several local ontologies is used. By using

semantic technologies, an integrated view for

heterogeneous data sources can be provided through

integration ontology.

The definition of in

portant part of using Sebi-framework in the

semantic integration. The most significant results are

achieved, when the integration ontology is broad

enough. Including the whole enterprise in integration

ontology is not possible; the integration ontology

covers one or several domain areas. The integration

ontology can be expanded and further specified

when new information systems are included in the

integration. The integration ontology has to include

all the important concepts, but not be too detailed. It

should not be too simple in order to enable

integration and to provide semantic consistency. On

the other hand too detailed integration ontology

wastes time and resources without providing

additional benefit.

In order to s

elopment the communication gap between

business, domain and IT-experts has to be

eliminated. Pure technical framework is not enough

for achieving the mutual understanding, and

therefore a process for using Sebi-framework is

built. This Sebi-process consists of four sub-

processes: Case Envisioning, Business, Expertise

and IT Domains. In Case Envisioning the basis for

integration ontology development is created. The

foundation of Case Envisioning is on Solution

Envisioning with Capability Cases - approach

(Polikoff et al., 2005). Solution Envisioning

provides means for definition of common

vocabulary between different parties and to make

right technology selection from the constantly

growing mass of available IT solutions.

The process does not specify any

iness process modelling. In semantic framework

presented in figure 1, the Pragmatic level could be

seen as a part of the Case Envisioning process,

where business needs and possibilities are defined.

Semantic level relates closely to the development of

integration ontology and syntactic level is connected

to Business, Expertise and IT Domains where

implementation dependent work is done.

The technical implementation of Sebi

rted by defining the integration ontology. Data in

various information systems is stored in

heterogeneous sources and formats. Using ontology

engineering tools a concept model can be

automatically formed from data sources. This

requires that the data is stored in suitable forms

including common relational database structures,

ontology files, XML-documents and xls-files. If data

is not in suitable form manual data processing is

required.

Integration

ology building tool. Mappings are used to

connect automatically generated concept models and

manually build integration ontology. Mappings

between the concepts in integration ontology and the

concepts in source concept models are done

manually. The purpose of mappings is to connect

concepts which have the same semantics. When

direct correspondence between concepts is not found

the mappings are done using reasoning. The

reasoning may be based on similarity of concepts

and the meaning of concepts.

Integration ontology opera

erent information systems by offering access to

the source information. The information is requested

by executing queries into integration ontology by

using middleware tool. The requested information

can be delimited according to application’s or

person’s needs so the information obtained is just

the information needed.

ICEIS 2009 - International Conference on Enterprise Information Systems

ation sharing among separate

inf

holistic representation of the enterprise

REFERENCES

. Information Integration with

Ontologies: Experiences from an Industrial Showcase,

Berlin and Heidelberg

digm, Springer, Berlin.

of Knowledge Management, e-Commerce and the

Semantic Web, Springer, London, UK.

Gu the Enterprise

Systems. V,

Spain,.

Pel J., 2008.

Pol

Pol n.

e, NJ.

Knowledge and Data

Sar

Enterprises, 12-14

Usc

We s Management.

Concepts, Languages, Architectures, Springer-Verlag

Berlin Heidelberg.

both technical and information integration. This

integration greatly depends how well the

communication among different stakeholders

involved in the information development process

succeeds. Lack of integrated methods and

implementation bias are examples of the problems.

This paper presents a semantic framework for

increasing the quality of enterprise interoperability

means of effective information integration. The

paper contributes with a semantic framework, where

Sebi-framework is extended with SOAD approach.

The advantage of SOAD is that by combining

intesubjective and objective aspects in one

delling notation it facilitates to better

understanding and reasoning about service

architecture across organizational and technical

system boundaries. It combines business data and

business process dimensions and integrates static

and dynamic aspects in one diagram type. Being

computation neutral, service-oriented approach used

for analysis and design is more comprehensible for

business experts.

The Sebi-framework uses semantic technologies

to enable inform

ormation systems. Interoperability between

separate information systems is achieved by

developing a shared information model of the

information.

The presented semantic framework will facilitate

to maintain a

ich is necessary for systematic analysis of service

architectures as well as to enterprise interoperability

where internal and external views are visualized

together.

Alexiev, V. et al., 2005

John Wiley & Sons, Chichester, UK.

Bennet, S., McRobb, S., Farmer, R., 2002. Object-

Oriented System Analysis and Design using UML.

McGrraw-Hill.

Bruijn, J.D., Feier, C., 2005. D4.6.1.1 Report on ontology

mediation for case studies V1. SEKT-project Report

2005-08-05.

Dietz, J.L.G., 2006. Enterprise Ontology Theory and

Methodology, Springer-Verlag

New York.

Dori, C., 2002. Object-Process Methodology: A Holistic

System Para

Gomez-Perez A, Corcho O, Fernandez-Lopez M., 2005.

Ontological Engineering with examples from the areas

Gruber, T.R., 1995. Towards principles for the design of

ontology used for knowledge sharing,. Int. J. Human–

Computer Studies, 43 (5/6), pp.907–928.

stas, R., Gustiené, P., 2004. Towards

Engineering Approach for Information System

Modelling across Organisational and Technical

Boundaries. In Enterprise Information

Kluwer Academic Publisher, Netherlands., pp. 235-252.

stas, R., Gustiené, P., 2007. Service-Oriented

Foundation and Analysis Patterns for Conceptual

Modellling of Information Systems. In 16th

International Conference on Information System

Development, August 29-31, Galway, Ireland.

stas, R., Gustiené, P., 2008. A New method for

Conceptual Modelling of Information Systems. In 17th

International Conference on Information System

Development, August 25-27, Paphos, Cyprus.

Gustiené, P., 2003. On Desirable Qualities of Information

System Specifications. In International Conference on

Concurrent Engineering: Research and Applications,

26-30 July, Madeira, Portugal.

Gustiené, P., Gustas, R. 2008. Introducing Service-

Orientation into System Analysis and Design. In 10th

International Conference on Enterprise Information

Systems, June 12-16, Barcelona,

Noy, N.F. and McGuinness, D.L., 2001. Ontology

Development 101: A Guide to Creating Your First

Ontology, Stanford Knowledge Systems Laboratory

Technical Report KSL-01-05, March 200

Park, J., Ram, S., 2004. Information Systems

Interoperability: What Lies Beneath?. ACM

Transactions on Information Systems, 22(4), pp. 595-632.

tomaa, I., Helaakoski, H., Tuikkanen,

Semantic Interoperability - Information Integration by

Using Ontology Mapping in Industrial Environment.

In 10th International Conference on Enterprise

Information Systems, June 12-16, Barcelona, Spain.

ikoff, I. and Allemang, D. 2003, Semantic Integration

Strategies and Tools, TQ TechnologyBriefing 2003,

TopQuadrant, USA.

lock, J.T., Hodgson, R., 2004. Adaptive informatio

Improving business through semantic interoperability,

grid computing, and enterprise integration, John

Wiley & Sons, Hobok

Ram, S., Park, J., 2004. Semantic Conflict Resolution

Ontology (SCROL): An Ontology for Detecting and

Resolving Data and Schema-Level Semantic Conflicts.

IEEE Transactions on

Engineering, 16(2), pp.189-202.

janoja, E. M., Helaakoski, H., and Peltomaa, I., 2008.

Semantic Interoperability in Industrial Environment.

In Modern Information Technology in the Innovation

Processes of the Industrial

November, Prague, Czech Republic.

hold, M., Gruninger, M., 1996. Ontologies: principles,

methods, and applications. Knowledge Engineering

Review 11(2), pp.93-155.

ske, M., 2007. Business Proces

SEMANTIC FRAMEWORK FOR INFORMATION INTEGRATION - Using Service-oriented Analysis and Design