Model-driven Approach for the Interoperability of Enterprises’ Services

Information Exchange

Catarina Marques-Lucena

, Jo˜ao Sarraipa

, Carlos Agostinho

and Ricardo Jardim-Goncalves

DEE, FCT, Universidade Nova de Lisboa, Caparica, Portugal

Centre of Technology and Systems, CTS, UNINOVA, Caparica, Portugal

Keywords:

Interoperability, Semantic Interoperability, Mappings Representation, Modelling Morphisms, Model Trans-

formations.

Abstract:

Nowadays, enterprises have been pushed up the rate of industrial transformation to high level products and

services. The capability to agilely respond to new market demands became a strategic pillar for enterprises

survival. It leads to the necessity of create mechanisms to allow enterprises to combine forces to compete

jointly in the market, in order to raise their own added value and to become specialist in niche activities. But to

unite forces, enterprises must exchange information/knowledge between them. A semantic problem emerges

when the same representation of a thing can have different meanings todifferent applications and enterprises. It

can be a disabler for information exchange and its interpretation. The Semantic Web Service concept emerges

as a solution to facilitate Web Services description and consequently discovery tasks, unifying the domain

services. Moreover, the data integration between concepts can be achieved trough mappings establishment

between Web Services elements. However, a solution that allows to map Web Service elements to semantic

concepts represented using OWL is still missing. To face this issue, a model-driven approach supported by a

mapping tool is here presented with the aim of facilitate the information exchange trough the establishment of

mappings between enterprises services models and the domain reference lexicon.

1 INTRODUCTION

In the emerging society, competitive markets are be-

coming increasingly dynamic, and consequently com-

plex, with companies not prospering and surviv-

ing through their own individual efforts (Friedman,

2006). Thus, global markets are willing to improve

their competitiveness through collaboration and part-

nerships, motivating companies to look for enhanced

interoperability between systems and applications in

industry. Therefore, enterprises need to be able to dy-

namically adapt themselves to take advantage of mar-

ket opportunities, establishing collaborative business

practices to compete with big enterprises (Agostinho

et al., 2006). Moreover, the ability of an enterprise to

interoperate with others is not only a recognized qual-

ity and advantage to obtain competitiveness in today’s

market, but it also becomes a matter of survival for

many companies.

Interoperability can be deﬁned as the ability of

two or more systems or components to exchange in-

formation and to use the information that has been

exchanged (IEEE, 1990). It means that enterprises

need their systems/products to work with the sur-

rounding systems/products without great efforts to be

interoperable. However, one of the main difﬁcul-

ties regarding the interoperability between systems

and applications is related to the high number of se-

mantic representations of the same segment of reality

(e.g. systems and products) which are not semanti-

cally coincident (even inside the same domain) (Lu-

cena et al., 2013), as a consequence of the hetero-

geneity of communities and enterprises. It results in

a difﬁcult semantic interoperability achievement. Se-

mantic Interoperability, deﬁned as the ability of sys-

tems/components to share and understand informa-

tion at the level of formally deﬁned and mutually ac-

cepted domain concepts (Sø lvberg, 1999), tradition-

ally is achieved through peer-to-peer mappings where

each participant tends to use its own data format and

business rules, handling as many mappings as the

number or partners to interoperate. Here, another in-

teroperability issue emerges, one related to systems

dynamics. Commonly, systems are time-variant, and

even if we are able to ﬁnd a ‘good’ model to describe

it, when facing the dynamics of the environment, such

Marques-Lucena, C., Sarraipa, J., Agostinho, C. and Jardim-Gonçalves, R.

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange.

DOI: 10.5220/0005853207890799

In Proceedings of the 4th International Conference on Model-Driven Engineering and Software Development (MODELSWARD 2016), pages 789-799

ISBN: 978-989-758-168-7

 2016 by SCITEPRESS – Science and Technology Publications, Lda. All r ights reserved

789

model will become obsolete in time. Moreover, a

model is just a representation of how an entity see

the world. It can (and should) be constantly reﬁned

in order to adapt to new requirements. As a conse-

quence, all this dynamics and heterogeneity leads in

most cases, the network to experience interoperabil-

ity problems because if just one of the network mem-

bers adapts to a new requirement, the harmony is bro-

ken, and the network begins experiencing interoper-

ability failure (Agostinho et al., 2011). This is even

more evident in multi-domain networks (e.g. collabo-

rative product design) where information is dispersed

and frequently replicated in many Information Sys-

tems trough Web Services usage.

1.1 Semantic Web Services

Going towards to what was described before, one of

the biggest issue in the Service Oriented Architecture

(SOA) vision is related to the the data heterogene-

ity between inter-operating services. This is because,

typically, enterprise systems are developed over sev-

eral periods of time, by diverse organization and not

necessarily with the same structures or vocabulary

(Nagarajan et al., 2006). As a consequence, a substan-

tial heterogeneity in Web Services elements’ syntax,

structure and semantics is veriﬁed.

Given the pivotal importance of service discov-

ery for service-oriented computing, several attempts

to improve the quality service discovery tasks are cur-

rently being pursued. One of the major efforts in this

direction is promoted by the World Wide Web Con-

sortium (W3C) which strongly advocates the intro-

duction of semantic information in the description of

Web services

(Brogi et al., 2005). Semantics ac-

quires a particular importance to share services in a

Peer to Peer (P2P) system where the lack of common

understanding of the world generated the need of ex-

plicit guidance in the process of discovering available

resources (Arabshian and Schulzrinne, 2007; Schmidt

and Parashar, 2004; Haase et al., 2004).

Currently, services description is expressed by

means of the Web Services Description Language

(WSDL

) by declaring a set of message formats

and their direction (incoming/outgoing) (Brogi et al.,

2005). WSDL, in essence, allows the speciﬁcation of

the syntax of the input ans output messages of a basic

service, as well as other details needed for the invo-

cation of the service (Martin et al., 2005). However,

it can be considered that WSDL do not offer sufﬁ-

cient semantic richness for services to be machine-

processable (Kamaruddin et al., 2012). Human in-

http://www.w3.org/2003/10/swsig-charter

http://www.w3.org/TR/wsdl

tervention is often needed to interpret the meanings

in order to discover, compose, and invoke Web Ser-

vices. Here, the notion of ontology is seen as an effec-

tive way to provide the required speciﬁcations to al-

low the automation of the mentioned tasks (Kamarud-

din et al., 2012). Five projects to support the idea of

ontologies incorporation in Web Services description

are: OntoGov (Tambouris et al., 2004), TERREGOV

(Vicente et al., 2006), SemanticGov (SemanticGov

Consortium, 2007), Access-eGov (Sroga, 2008) and

FUSION (FUSION Consortium, 2008).

1.2 Web Services Interoperability

There are several tools that allow mappings estab-

lishment across several data formats. One example

is the Advanced Mapping and Transformation mod-

ule of talend (talend, 2015). This module uses a

Eclipse-based tooling environment to enable users to

build, tests and maintain mappings across several data

formats, including transformations between java and

complex XML data, JSON, databases, ﬂat ﬁles and

EDI applications. It can be used to integrate applica-

tions, services, and APIs without coding. The SAP

NetWeaver Composition Environment Library(SAP,

2013), allows to users to deﬁne how the data com-

ing from one service operation as input in the ser-

vice ﬂow has to be mapped (deﬁnition of what will

be the output) in the service ﬂow. It allows to directly

deﬁne how the data from one service data structure

is mapped to another data structure. Other applica-

tion tool to consider is IBM’s WebSphere Application

Server V8.5.5 (IBM, 2013). It supports services map-

pings establishment by providing a simple way of per-

forming content-based routing and message transfor-

mation that can be applied to web service messages

being sent from service clients without needing to

make changes to either the service client or service

provider. All this tools are valuable solutions con-

tributing to the deployment of interoperable Web Ser-

vices across platform, applications, and programming

languages. However, the authors didn’t found any in-

dication that this tools are capable to import OWL,

used to deﬁne semantic relationships between con-

cepts. This goes against the need of semantic descrip-

tions usage to facilitate services discovery.

To face this issue, the authors proposed a solution

able to represent connections or mappings between

information systems, or more precisely, between web

services elements and semantic concepts described by

ontologies. The semantic representation of the do-

main is accomplished by a domain reference lexicon

establishment in collaboration with the domain ex-

perts. Thus, the proposed solution to achieve Web

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

790

Services Interoperability is supported by the usage of

a common reference lexicon (domain semantics) to be

the intermediary in the communications between en-

terprises Web Services and to the outside. Then, the

establishment of mappings between enterprises Web

Services and the reference domain lexicon will allow

each of the enterprises to keep its own knowledge and

semantics unchanged, and still able to smoothly inter-

act with its domain.

In this paper, an initial assessment about the need

for semantic descriptions incorporation in Web Ser-

vices and its interoperability is conducted. Based

on this necessity, an approach for the interoperabil-

ity of enterprises information exchange is proposed

in section 2. Then, some key concepts to accomplish

the proposed approach are detailed: Modeling Mor-

phisms to Enable Sustainable Interoperability, Com-

munication Mediator and Model Transformations. In

section 6, the mechanism developed for the banking

and insurance domain, capable of representing con-

nections/mappings between elements of the domains

ontology and Native Business Service (NBS), which

are existing services deployed and exploited within

the context of each service supplier, is presented. Fi-

nally, some conclusions and feature work statements

are presented.

2 APPROACH FOR THE

INTEROPERABILITY OF THE

INFORMATION EXCHANGE

The author’s approach aims to provide a solution (see

Figure 1) able to represent connections or mappings

between enterprises Native Business Services (NBS),

or more precisely, between message elements of ser-

vices and semantic concepts described by ontologies

(reference domain lexicon). NBS are already existing

software installed and exploited within the context of

each supplier. The mappings with a reference lexi-

con can, then, be used to generate Supplier Business

Services (SBS) accordingly to the elements of the do-

main ontology. SBS’s are a set of features that are ex-

posed from each supplier infrastructure and consist in

the externalization of one or more NBSs. The genera-

tion of these SBSs promotes interoperability of enter-

prises applications in the domain allowing integration

across multiple stakeholders (suppliers) and domains

of interest (customers).

The ﬁrst step to materialize this software-based

system is to acquire a platform reference lexicon

together with the domain experts. These are the

main actors of their domain, thus their participation

in the domain glossary (or reference lexicon) would

increase the awareness of the domain terms. Glos-

saries are lists of specialized terms, mostly in alpha-

betic order, that sometimes are unique to a speciﬁc

subject. Each term is composed by its corresponding

description. It includes descriptive comments and ex-

planatory notes, such as deﬁnitions, synonyms, ref-

erences, etc (Sarraipa et al., 2014). Given the rel-

evance of glossaries to better understand domains

lexicon, several works were already presented using

different methods and techniques. While some au-

thors opt for rule-based systems usage to do the text-

mining and glossary building (Muresan and Klavans,

2002), (Westerhout and Monachesi, 2007), others pre-

sented solutions based on machine learning (Fahmi

and Bouma, 2006), and statistical techniques to simu-

late human consensus (Velardi et al., 2006).

Then, a mapping tool that allows to represent

both the domain glossary and the supplier services el-

ements is required. This tool is intended to be used for

administrative purpose (supplier’s manager) and sup-

ports the process of mappings establishment and stor-

age between the platform and suppliers data models.

This mappings will be stored in a speciﬁc ontology -

the Communication Mediators (detailed in section 4).

Finally, the generator is responsible for the gen-

eration of SBSs based on the mappings (represented

in the communication mediator) created between the

NBS and the platform ontology.

3 MODELING MORPHISMS TO

ENABLE SUSTAINABLE

INTEROPERABILITY

Models are used to capture the essential features of

real systems by breaking them down into more man-

ageable parts that are easy to understand and to ma-

nipulate. They are used in systems developmentactiv-

ities to draw blueprints of the systems and to facilitate

communication between different people in the team

at different levels of abstraction. People have differ-

ent views of the system and models can help them to

understand these views in a uniﬁed manner (Abdullah

et al., 2002).

In mathematics, Morphism (MoMo) is an abstrac-

tion of a structure-preserving map between two math-

ematical structures. It can be seen as a function in

set theory, or the connection between domain and co-

domain in category theory (INTEROP NoE, a). Re-

cently, this concept has been gaining momentum ap-

plied to computer science, namely to systems inter-

operability. This new usage of morphisms speciﬁes

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange

791

Figure 1: Approach for the interoperability of the information exchange.

the relations (e.g. mapping, merging, transformation,

etc) between two or more information model speciﬁ-

cations (M as the set of models).

In this context, the research community identiﬁes

two core classes of MoMo: non-altering and model

altering morphisms (INTEROP NoE, a; Agostinho

et al., 2007). In the non-altering morphisms, given

two models (source A and target B), a mapping is cre-

ated relating each element of the source with a cor-

respondent element in the target, leaving both mod-

els intact. In model altering morphisms, the source

model is transformed using a function that applies a

mapping to the source model and outputs the target

model (Delgado et al., 2006). Other relations, such as

the merge operation, can also be classiﬁed as model

altering morphisms, however they are not detailed in

this work.

3.1 Knowledge Enriched Tuple for

Mapping Representations

The research community has developed many propos-

als to morphisms representations (INTEROP NoE, b).

As analyzed in (Agostinho et al., 2011), graph the-

ory has been used in some, although other theories

can be considered to achieve the envisaged goals, e.g.,

set theory (Dauben, 1990), model management (Bern-

stein, 2003), or semantic matching (Sarraipa et al.,

2010). However there is not a single perfect solution

that can be used to achieve all the morphisms goals

at once. Some are ideal for structural issues, oth-

ers for semantics providing good human traceability,

and others are more formal and mathematical based.

For that reason, in this work, is used a 5-tuple map-

ping expression (Mapping Tuple - MapT), presented

in (Agostinho et al., 2011), with the goal of consoli-

date existent approaches to morphisms:

< ID, MElements, KMType, MatchClass, Exp >

(1)

being:

• ID - unique identiﬁer of the MaptT;

• MElements - pair indicating the mapped ele-

ments;

• KMType - stands for Knowledge Mapping Type

and determines the type of mappings represented

in a speciﬁc instance of MapT;

• MatchClass - stands for Match/Mismatch classi-

ﬁcation and depends on the KMType:

– if a = b, the mapping is absolute and

MatchClass = Equal;

– if KMType = conceptual, the mapping is relat-

ing terms/concepts:

{Equal, Naming,Coverage,MoreGeneral,

LessGeneral, Disjoint}

– Otherwise the mapping is either non-existing or

more concrete addressing structural issues

• EXP - stands for mapping expression that relates

and further speciﬁes the previous tuple compo-

nents. Normally, this expression can be trans-

lated to an executable transformation language

(e.g. ATL).

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

792

4 COMMUNICATION MEDIATOR

The Communication Mediator Ontology has been

built up as an extension of the Model Traceability On-

tology deﬁned by (Sarraipa et al., 2007), and it is able

to represent ontology semantic operations, like: (1)

Semantic mismatches; (2) Semantic transformations;

(3) Ontologies mappings; and other ontologies opera-

tions. Thus, the MO is able to log ontology and entity

operations in a way that is possible to trace changes in

all the ontology life cycle. It addresses traceability as

the ability to chronologically interrelate the uniquely

identiﬁable objects in a way that can be processed by

a human or a system. The mapping relations can be

related to a traceability element, in such sense that

a speciﬁc term deﬁned in the reference ontology has

a related one in the organisation member ontology,

making possible a way to trace ontology elements.

This way, the morphisms are modelled with trace-

ability properties in a sense that they enable to store

different versions of information model elements, as

well as mappings between speciﬁc objects deﬁned in

a model or ontology (Sarraipa et al., 2010).

The structure of the MO is described as follows:

the MO has two main classes: ‘Object’ and ‘Mor-

phism’. The ‘Object’ represents any ‘Information-

Model’ (IM), which is the model itself, and ‘Mod-

elElements’ (MElements) (also belonging to the IM)

that can either be classes, properties or instances.

The ‘Morphism’ associates a pair of ‘Objects’ (re-

lated and relating), and classiﬁes their relationship

with a ‘MorphismType’ (MType), ‘KnowledgeMap-

pingType’ (KMType) (if the morphism is a mapping),

and ‘Match/Mismatch’ (MatchClass) class (Sarraipa

et al., 2010).

5 MODEL TRANSFORMATIONS

The models mapping speciﬁcations can be performed

either at a high level of formalization using graphs,

sets, tuples, etc. or at lower levels, i.e. specifying the

mappingsby text. However, in both cases is necessary

to implement them using a transformation language.

Atlas Transformation Language (ATL) is one

of the most used transformation languages, hav-

ing a large user base and being very well docu-

mented(Bzivin et al., 2003). An ATL transformation

is composed by a set of rules (matched rules) that

deﬁne how the source model elements are linked to

the target model elements. These elements can then

be ﬁlled with information from the source model by

called rules (similar to functions in usual object lan-

guages like JAVA) and action blocks (blocks of im-

perativecode which can be used by matched rules and

called rules).

Model transformations are a current practice to

enable interoperability among two organizations or

ecosystems of organizations, each with their own ser-

vice system. With them, one can specify P2P (Peer

to Peer) mappings to translate any data provided from

one sides format speciﬁcation into the other, thus al-

lowing a seamless exchange of information.

When performing a horizontal model transforma-

tion (e.g. converting instances of a model to in-

stances of another model) an explicit or an implicit

mapping of the ‘meta-model’ has to be performed.

Thus, as depicted in 2 the idea that when perform-

ing a transformation morphism at a certain level n,

this transformation has (implicitly or explicitly) to be

n + 1. Once the n + 1 level mapping is complete,

executable languages can be used to implement the

transformation at n level, e.g. using ATL and the

Query/View/Transformation (QVT) (OMG, 2008).

Figure 2: Method for Horizontal Transformation.

This type of transformations are normally static

processes that once deﬁned can be repeated any num-

ber of times achieving the same results. However due

to the constant knowledge change caused by the dy-

namics of the global market, services and models that

regulate enterprise systems, are not. In fact, some re-

searchers have attempted to extrapolate results from a

‘general systems theory’ or ‘complexity theory’ that

could explain the importance of evolution of systems

in all ﬁelds of science (Gharajedaghi, 2011). These

theories view all systems as dynamic, ‘living’ enti-

ties that are goal-oriented and evolve over time, thus,

information systems, services and the mappings that

connect them should be prepared to respond to the

environment dynamics which is in a constant update.

To support this dynamicity, horizontal transforma-

tions should be provided with traceability features,

and mappings stored in a parseable and structured

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange

793

knowledge-base (Agostinho et al., 2011).

Since it is used a knowledge enriched tuple for

mappings representation that are stored on a CM

is possible to keep traceability of model mapping

changes so that readjustments can be easier to man-

age, and data exchange re-established automatically

using the model-driven development paradigm.

6 ISOFIN CLOUD APPLICATION

SCENARIO

The ISOFIN cloud project envisage the establishment

of technical solutions, guidelines and standards to

support the interoperability resolution, facilitating a

seamless sharing of artifacts at the knowledge and

software level that would enable a fast and efﬁcient

creation of new Financial (e.g. bench and Insurance)

products or services. Thus, ISOFIN’s overall aims

are:

• Analyze the problems resulting from the internal-

ization of ﬁnancial applications and identify po-

tential solutions

• Create ways of reducing the time that software de-

velopers take to achieve interoperability between

domain ﬁnancial stakeholders (e.g. bench and in-

surance)

• Standardize the domain semantic level, where no

harmonization exists

• To implement mechanisms to facilitate the gener-

ation of new services in a ubiquitous context that

could easily be discovered and accessed by do-

main customers

Based on the project aims, the expected result is

the development of a platform that speeds up the de-

velopment and deployment of services to be sold to

the clients and users of the ﬁnancial domain. For that,

the steps identiﬁed in the approach presented in sec-

tion 2 for the interoperability of the information ex-

change section must be followed: (1) acquisition of a

platform reference lexicon together with the domain

experts; (2) deﬁnition of mappings between enterprise

NBS and the reference lexicon using a mapping tool;

and (3) Generation of SBSs based on the established

mappings.

6.1 Domain Reference Lexicon

Establishment

It is a fact that when an information system intends to

represent a domain’s knowledge it needs to be aligned

to the community that it represents. Consequently it

is required to have a solution where community mem-

bers could present their knowledge about the domain

and discuss it with their peers. Additionally, such

knowledge must be available and dynamically main-

tained by all the involved actors (Marques-Lucena

et al., 2015).

The solution adopted by the authors for the do-

main reference lexicon establhishement is based on

MENTOR - methodology for enterprise reference on-

tology development (Sarraipa et al., 2010), supported

by an interface accordingly with the works presented

in (Lucena et al., 2014; Marques-Lucena et al., 2015)

to implement the following MENTOR’s steps (1) ter-

minology gathering; (2) glossary building; (3) the-

saurus building. The result is a domain lexicon,

whose semantics are constantly reﬁned trough a spe-

ciﬁc front-end in order to handle the knowledge pro-

vided by the domain experts.

6.2 Mapping Tool

The Mapping tool was previously developed to pro-

vide a graphical means to deﬁne different kinds of

mapping between models (Agostinho et al., 2012).

This tool was developed with the intention to open

LIMM (Language Independent Meta-Model) ﬁles, an

abstract interface on top of information systems.

JGraph has been elected and modiﬁed to read the

input information model ﬁles and store morphisms at

the MO ontology. It is a widely used open source

project for graph visualization and manipulation, sim-

ilar to Microsoft Visio, with good documentation and

several examples. Features include a complete selec-

tion of layouts to automatically position the graph,

many styles of shapes and edges, validation of con-

nections, as well as an undo and redo manager.

Some adjustments were made to enable the inter-

action (mapping deﬁnition) between two different in-

formation models’ graphs, and to become integrated

to the Communication Mediator Ontology. To the in-

tegration with the CM ontology, JENA was used - a

Java API for OWL providing services for model rep-

resentation, parsing, database persistence and query-

ing ontologies.

In the ISOFIN’s application scenario. the map-

ping tool supports the mappings deﬁnition between

NBSs elements (e.g. services and message concepts

names) described in WSDL 1.0 and XSD, and the

ISOFIN’s reference lexicon described in OWL 2.0.

For the purpose, the enabled features are the ones rep-

resented in Figure 3 use case: 1) Open wsdl ﬁle; 2)

save mappings; 3) import mappings; 4) generate wsdl

ﬁle. The import of the platform reference lexicon

is made when the open wsdl ﬁle option is triggered.

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

794

That means, when the user decided to open a wsdl ﬁle

for mappings establishment, both the wsdl and refer-

ence lexicon graphical representations are presented

side by side (see Figure 9).

Figure 3: Mapping Tool Use Case.

In addition to the menu bar, three other areas com-

pose the tool: 1) Mapping text description; 2) Mini-

mized view of the full mapping screen; and 3) Map-

ping screen. The ﬁrst area is where are described the

selected mappings and elements, the second allows to

navigate in the full map and ﬁnally, the third is where

is done the representation of the models and the map-

pings (see Figure 4).

Figure 4: Mapping Tool working area.

Starting from the work presented in (Agostinho

et al., 2012), some changes were made to the mapping

tool to allow the interpretation of WSDL 1.0 (sup-

pliers web services standard language) and OWL 2.0

(platform semantics description language). The most

relevant changes are related to the transformation of

both the input ﬁles into LIMM, so

both the reference

lexicon and NBS wsdl can be opened for mappings

establishment

6.2.1 Language Independent Meta Model

This language enables the abstraction in relation to

technologies and implementation details associated

with the different modeling languages, and thus, en-

large the scope of users involved in a traditional map-

ping deﬁnition activity. Thus, having domain experts

(suppliers) involved in the mapping process increases

the quality of the mappings enabling the interoper-

ability between suppliers and the rest of the platform

participants.

6.2.2 OWL to LIMM Transformation

The transformation of OWL 2.0 to LIMM was al-

ready existing in the context of the European Project

CRESCENDO (CRESCENDO IP, 2009) and ex-

plained in (Agostinho et al., 2012). To enable the

mapping among the OWL meta-model (W3C, 2008)

and the LIMM, one needs to ﬁrstly put the OWL data

in XMI serialization following the OWL meta-model

speciﬁcations. Nevertheless the procedure to do so

is not straightforward as desirable since, in spite of

the inputting OWL model is already XML serialized,

it cannot be directly processed by the ATL toolkit

which needs XMI as an input. The complete process

for accomplishing the language mapping is illustrated

in Figure 5, where the ﬁrst step consists in doing

an injection of the original model to an XML MOF

meta-model speciﬁcation. Following that, the second

preparatory step consists in mapping that XML for-

mat to the reference OWL meta-model which will be

the starting point for the actual θ(OWL, LIMM) lan-

guage mapping (step 3). The, the step 4 takes place

to execute the transformation between models itself

using ATL.

Figure 5: OWL 2.0 to LIMM Transformation (Agostinho

et al., 2012).

6.2.3 WSDL to LIMM Transformation

The transformation of WSDL 1.0 to LIMM follows

the same approach of the transformation of OWL 2.0

to LIMM described before. Since not all the WSDL

model elements are relevant for the described use

case, only the elements depicted using yellow (doted

line) on Figure 6 where transformed. These are the

elements containing the description of the services

name, input and output messages, and the catego-

rization of the messages. To allow the representation

of services input and output messages, the container

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange

795

types were also considered (continuous line in red).

The WSDL ‘types’ element is a container for XML

Schema type deﬁnitions. The type deﬁnitions are ref-

erenced from higher-level message deﬁnitions in or-

der to deﬁne the structural details of the message. To

get more structured representation of the NBS con-

tent, the representation of the services was split in two

parts, namely the types and the services (Figure 7).

Figure 6: WSDL 1.0 Meta-model.

Figure 7: WSDL 1.0 LIMM Graphical representation.

6.3 Mappings Storage

As mentioned before, the mapping tool provides a

graphical means to deﬁne different kind of mappings

while storing them in the Communication Mediator

Ontology. Using the mapping tool developed, it is

possible to open and show mappings between infor-

mation models, namely, the NBS services and the

ISOFIN reference lexicon, represented respectively in

WSDL and OWL (Figure 9).

Since a mapping deﬁnition is a complex and time-

consuming task, the Mapping tool is capable of halt-

ing the process at any time without losing the progress

made so far by the user. This functionality still allows

keeping mappings traceability.

6.4 Generate SBS WSDL File

To mediate information between the ISOFIN suppli-

ers (NBS) and the ISOFIN platform it is needed to

align the semantics of such information. To accom-

plish that, a generator to enable the representation of

NBS services according to the ISOFIN nomenclatures

was developed. This generator consists in a compo-

nent of the mapping tool and its function is

to gen-

erate a WSDL of the supplier services accordingly to

the ISOFIN semantics from the established mappings.

Figure 8: WSDL output of the established mapping.

In Figure 8, the Supplier Business Service re-

sulting of the mapping established in Figure 9 is

presented. In the example provided, the element

codsys of the complex element getXMLApoliceBy-

ModApolConﬁg is replaced by the ISOFIN concept

CodigoDoProduto. In this way, each enterprise is ca-

pable to keep its internal nomenclature (e.g. codsys)

and still be able to interact with its domain using a

common reference lexicon as the intermediary in the

communications established between the enterprise

and the outside.

7 CONCLUSIONS

The aim of this work is to provide interoperability in-

tegration on the information exchanged between do-

main suppliers and a speciﬁc platform. To achieve

that goal, an approach capable to represent connec-

tions or mappings between information systems is

proposed by the authors. The outcome is that sys-

tems information subsystems are able to interact, but

keeping their internal nomenclature thanks to the gen-

eration of Supplier Business Services.

Thus, in the application scenario, a mapping tool

was used to support the establishment and deﬁnition

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

796

Figure 9: Mapping Tool demonstration.

of mappings between the ISOFIN ontology and sup-

plier data models. These mappings are logged ac-

cordingly to a determined speciﬁcation able to facil-

itate further traceability mechanisms implementation

to facilitate maintenance of the established interoper-

ability between the information systems. Such spec-

iﬁcations are represented by a tuple that formalizes

the mappings, which then can be stored as knowledge

elements in the Communication Mediator Ontology.

The stored knowledge elements enable the gen-

eration of SBS (wsdl) web service already compli-

ant with ISOFIN’s reference lexicon. Such new web

services will facilitate the process of sending and re-

ceiving interoperable data between suppliers and the

ISOFIN system.

However, due to the inﬁnite possible model for-

mats (meta-models) to represent any kind of informa-

tion in web services contents composition (e.g. NBS),

it is difﬁcult to create a full interoperability solution.

Thus, despite the fact of the complexity of reaching

a complete automatic or dynamic interoperability it

is concluded that partial solutions could be reached

when it is acknowledged in advance the nature of the

meta-models of the information exchanged and when

there is a speciﬁcation for traceability representation

able to represent all the kinds of mismatches.

These traceability representations support the pro-

posed generator in handling speciﬁc interoperability

situations. In matter of fact this is the speciﬁc ISOFIN

goal to which this work is focused on. It provides

traceability representation able to support some spe-

ciﬁc interoperability establishment solutions imple-

mentation.

ACKNOWLEDGEMENTS

The research leading to these results have received

funding from the European Union 7th Framework

Programme (FP7/2007-2013) under grant agreement

OSMOSE

nr 610905, EC HORIZON2020 Program

under grant agreement nr C2NET 636909

, and also

QREN under agreement 2010/013837 ISOFIN

REFERENCES

Abdullah, M. S., Benest, I., Evans, A., and Kimble, C.

(2002). Knowledge modelling techniques for develop-

ing knowledge management systems. In Third Euro-

pean Conference on Knowledge Management: Trinity

College Dublin, Ireland, page 17.

Agostinho, C., Costa, R., Malo, P., and Jardim-Goncalves,

R. (2006). Product Data integration in the demand of

interoperability in e-Business. In Intelligent Systems,

2006 3rd International IEEE Conference on, pages

807–812.

Agostinho, C., ern, J., and Jardim-Goncalves, R. (2012).

Mda-based interoperability establishment using lan-

guage independent information models. In van Sin-

deren, M., Johnson, P., Xu, X., and Doumeingts,

G., editors, Enterprise Interoperability, volume 122

of Lecture Notes in Business Information Processing,

pages 146–160. Springer Berlin Heidelberg.

Agostinho, C., Sarraipa, J., DAntonio, F., and Jardim-

Gonalves, R. (2007). Enhancing step-based interoper-

abity using model morphisms. In Gonalves, R., Mller,

J., Mertins, K., and Zelm, M., editors, Enterprise In-

teroperability II, pages 817–828. Springer London.

http://www.osmose-project.eu

http://www.c2net-project.eu/

http://isoﬁncloud.i2s.pt/

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange

797

Agostinho, C., Sarraipa, J., Goncalves, D., and Jardim-

Goncalves, R. (2011). Tuple-based semantic and

structural mapping for a sustainable interoperability.

In Camarinha-Matos, L., editor, Technological Inno-

vation for Sustainability, volume 349 of IFIP Ad-

vances in Information and Communication Technol-

ogy, pages 45–56. Springer Berlin Heidelberg.

Arabshian, K. and Schulzrinne, H. (2007). An ontology-

based hierarchical peer-to-peer global service discov-

ery system. Journal of Ubiquitous Computing and In-

telligence, 1(2):133–144.

Bernstein, P. A. (2003). Applying Model Management to

Classical Meta Data Problems. In Proceedings of the

1st Biennial Conference on Innovative Data Systems

Research (CIDR).

Brogi, A., Corﬁni, S., and Popescu, R. (2005).

Composition-oriented service discovery. In

Gschwind, T., Amann, U., and Nierstrasz, O.,

editors, Software Composition, volume 3628 of

Lecture Notes in Computer Science, pages 15–30.

Springer Berlin Heidelberg.

Bzivin, J., Dup, G., Jouault, F., Pitette, G., and Rougui, J. E.

(2003). First experiments with the atl model transfor-

mation language: Transforming xslt into xquery. In

2nd OOPSLA Workshop on Generative Techniques in

the context of Model Driven Architecture.

CRESCENDO IP (2009). Collaborative and robust engi-

neering using simulation capability enabling next de-

sign optimisation (fp7-234344).

Dauben, J. (1990). Georg Cantor: His Mathematics and

Philosophy of the Inﬁnite. History of science. Prince-

ton University Press.

Delgado, M., Agostinho, C., Malo, P., and Jardim-

Goncalves, R. (2006). A framework for step-based

harmonization of conceptual models. In Intelligent

Systems, 2006 3rd International IEEE Conference on,

pages 781–785.

Fahmi, I. and Bouma, G. (2006). Learning to Identify Deﬁ-

nitions using Syntactic Features. Learning to identify

deﬁnitions using syntactic features.

Friedman, T. L. (2006). The World is ﬂat : a brief history

of the twenty-ﬁrst century. Farrar, Straus and Giroux,

New York.

FUSION Consortium (2008). Fusion project website.

Gharajedaghi, J. (2011). Systems Thinking: Managing

Chaos and Complexity: A Platform for Designing

Business Architecture. Elsevier Science.

Haase, P., Agarwal, S., and Sure, Y. (2004). Service-

oriented semantic peer-to-peer systems. In Bussler,

C., Hong, S.-k., Jun, W., Kaschek, R., Kinshuk, Kr-

ishnaswamy, S., Loke, S., Oberle, D., Richards, D.,

Sharma, A., Sure, Y., and Thalheim, B., editors, Web

Information Systems WISE 2004 Workshops, volume

3307 of Lecture Notes in Computer Science, pages

46–57. Springer Berlin Heidelberg.

IBM (2013). Websphere application server v8.5.5.

IEEE (1990). IEEE Standard Glossary of Software Engi-

neering Terminology. Technical report.

INTEROP NoE. Deliverable MoMo. 2-TG MoMo

Roadmap. InterOp.

INTEROP NoE. Problem Analysis and Roadmap - I-VLab

Platform. InterOp.

Kamaruddin, L. A., Shen, J., and Beydoun, G. (2012).

Evaluating usage of wsmo and owl-s in semantic web

services. In Proceedings of the Eighth Asia-Paciﬁc

Conference on Conceptual Modelling - Volume 130,

APCCM ’12, pages 53–58, Darlinghurst, Australia,

Australia. Australian Computer Society, Inc.

Lucena, C., Sarraipa, J. a., and Jardim-Goncalves, R.

(2014). A knowledge management framework to sup-

port online communities creation. In Camarinha-

Matos, L., Barrento, N., and Mendona, R., editors,

Technological Innovation for Collective Awareness

Systems, volume 423 of IFIP Advances in Informa-

tion and Communication Technology, pages 29–36.

Springer Berlin Heidelberg.

Lucena, C., Sarraipa, J. a., and Jardim-Gonc¸alves, R.

(2013). Semantic Adaptation of Knowledge Repre-

sentation Systems. In Camarinha-Matos, L., Tomic,

S., and Grac¸a, P., editors, Technological Innovation

for the Internet of Things, volume 394 of IFIP Ad-

vances in Information and Communication Technol-

ogy, pages 88–95.

Marques-Lucena, C., Agostinho, C., Koussouris, S., and

Sarraipa, J. (2015). Knowledge management frame-

work using wiki-based front-end modules. In Pro-

ceedings of the 17th International Conference on En-

terprise Information Systems, pages 79–86.

Martin, D., Paolucci, M., McIlraith, S., Burstein, M., Mc-

Dermott, D., McGuinness, D., Parsia, B., Payne, T.,

Sabou, M., Solanki, M., Srinivasan, N., and Sycara,

K. (2005). Bringing semantics to web services: The

owl-s approach. In Cardoso, J. and Sheth, A., edi-

tors, Semantic Web Services and Web Process Com-

position, volume 3387 of Lecture Notes in Computer

Science, pages 26–42. Springer Berlin Heidelberg.

Muresan, S. and Klavans, J. (2002). A Method for Automat-

ically Building and Evaluating Dictionary Resources.

Proceedings of the 3rd LREC, pages 231–234.

Nagarajan, M., Verma, K., Sheth, A., Miller, J., and

Lathem, J. (2006). Semantic interoperability of web

services - challenges and experiences. In Web Ser-

vices, 2006. ICWS ’06. International Conference on,

pages 373–382.

OMG (2008). Meta object facility (omg) 2.0

query/view/transformation speciﬁcation.

SAP (2013). Sap netweaver composition environment li-

brary.

Sarraipa, J., Jardim-Goncalves, R., and Steiger-Garcao, A.

(2010). Mentor: an enabler for interoperable intelli-

gent systems. International Journal of General Sys-

tems, 39(5):557–573.

Sarraipa, J., Marques-Lucena, C., Baldiris, S., Fabregat, R.,

and Aciar, S. (2014). The alter-nativa knowledge man-

agement approach. Journal of Intelligent Manufactur-

ing, pages 1–17.

Sarraipa, J., Zouggar, N., Chen, D., and Jardim-Goncalves,

R. (2007). Annotation for enterprise information man-

agement traceability. In ASME 2007 International De-

sign Engineering Technical Conferences and Comput-

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

798

ers and Information in Engineering Conference, pages

893–899. American Society of Mechanical Engineers.

Schmidt, C. and Parashar, M. (2004). A peer-to-peer ap-

proach to web service discovery. World Wide Web,

7(2):211–229.

SemanticGov Consortium (2007). Semanticgov project

website.

Sø lvberg, A. (1999). Data and What They Refer to. In

Goos, G., Hartmanis, J., van Leeuwen, J., Chen, P.,

Akoka, J., Kangassalu, H., and Thalheim, B., editors,

Conceptual Modeling, volume 1565 of Lecture Notes

in Computer Science, pages 211–226. Springer Berlin

Heidelberg.

Sroga, M. (2008). Access-egov-personal assistant of public

services. In Computer Science and Information Tech-

nology, 2008. IMCSIT 2008. International Multicon-

ference on, pages 421–427.

talend (2015). Advanced mapping and transformation.

Tambouris, E., Gorilas, S., Kavadias, G., Apostolou, D.,

Abecker, A., Stojanovic, L., and Mentzas, G. (2004).

Ontology-enabled e-gov service conﬁguration: An

overview of the ontogov project. In Wimmer, M., ed-

itor, Knowledge Management in Electronic Govern-

ment, volume 3035 of Lecture Notes in Computer Sci-

ence, pages 122–127. Springer Berlin Heidelberg.

Velardi, P., Poler, R., and Tom´as, J. (2006). Methodology

for the Deﬁnition of a Glossary in a Collaborative Re-

search Project and its Application to a European Net-

work of Excellence. In Konstantas, D., Bourri`eres,

J.-P., L´eonard, M., and Boudjlida, N., editors, Inter-

operability of Enterprise Software and Applications,

pages 311–322. Springer London.

Vicente, S., Prez, M., Garca, X., Gimeno, A., and Naval, J.

(2006). Terregov: egoverment interoperability on a se-

mantically driven world. In Konstantas, D., Bourrires,

J.-P., Lonard, M., and Boudjlida, N., editors, Inter-

operability of Enterprise Software and Applications,

pages 461–463. Springer London.

W3C (2008). Owl 2 web ontology language mof-based

metamodel.

Westerhout, E. and Monachesi, P. (2007). Extraction

of Dutch deﬁnitory contexts for eLearning purposes.

Proceedings of CLIN, pages 1–15.

Model-driven Approach for the Interoperability of Enterprises’ Services Information Exchange

799