e-swim: Enterprise Semantic Web Implementation Model

Towards a Systematic Approach to Implement the Semantic Web in Enterprises

Reinaldo Ferreira

and Isabel Seruca

2,3

Planeta Virtual, Porto, Portugal

DICT, Universidade Portucalense, Porto, Portugal

Centro Algoritmi, Universidade do Minho, Braga, Portugal

Keywords: Semantic Web, Linked Data, Implementation Model, Enterprise.

Abstract: The adoption of Semantic Web technologies constitutes a promising approach to data structuring and

integration, both for public and private usage. While these technologies have been around for some time,

their adoption is behind overall expectations, particularly in the case of Enterprises. This paper discusses the

challenges faced in implementing Semantic Web technologies in Enterprises and proposes an

Implementation Model that measures and facilitates that implementation. The advantages of using the

model proposed are two-fold: the model serves as a guide for driving the implementation of the Semantic

Web as well as it helps to evaluate the impact of the introduction of the technology.

1 INTRODUCTION

Over the last twenty years, the Web evolved from a

laboratory where it consisted in merely an idea to a

ubiquitous and universal environment, accessible

from basically everywhere, with information from

basically everything. Yet, the “Web of Documents”

is far from being considered a fully structured

version of the Web. The thinking underlying the

Semantic Web, providing a common framework that

allows data to be shared and reused across

application, enterprise and community boundaries

(W3C, 2013), puts forward the introduction to data

access as well as to data structure and meaning, and

enables a more integrated robust and practical

environment. It is believed that the next two decades

will provide enough time to spread the “Web of

Data”, being that the expectation of the scientific

community.

The introduction of the Semantic Web involves

the progressive transformation of the Web based on

hyperlinks between documents, the base for its first

generation, in the Web based on hyperlinks between

data or information, giving place to a Web scale

distributed database (Heath & Bizer, 2011). This

very large database is still in its early days, although

a rather significant number of examples that allow to

evaluate its potential can be found, since some

research efforts and government policies for data

publication have already produced satisfactory

results (Bizer, Heath, Idehen, & Berners-Lee, 2008).

Despite these contributions, Semantic Web

technologies usage within the enterprise community

is still a rather unexplored theme and in an early

adoption phase (Ahmed & Gerhard, 2010; Kuhn,

2010). The reasons for this reduced level of adoption

may point to several difficulties, such as the

homogenization and validation of data sources, the

definition of knowledge rules and borders that allow

to relate data in a uniform way, the analysis of too

complex examples, the availability of low cost

technological capacity to allow its implementation,

the availability of development tools, the recruitment

of experienced professionals, the diffusion of

success stories and the adoption of a paradigm shift

in modelling, design and development (Ahmed &

Gerhard, 2010; Kang et al., 2008; Kuhn, 2010;

Pollock, 2008).

Acknowledging these shortcomings, this paper

presents a model for Semantic Web implementation

in Enterprises with two main goals. The first goal is

to facilitate the introduction of the technology in

organizations with different characteristics and

motivations, acting as a guide and providing a

roadmap for a quicker and more intensive adoption.

The second goal is to evaluate the impact of the

introduction of the technology in the applications

used in these organizations and in the tasks

653

Ferreira R. and Seruca I..

e-swim: Enterprise Semantic Web Implementation Model - Towards a Systematic Approach to Implement the Semantic Web in Enterprises.

DOI: 10.5220/0004972706530658

In Proceedings of the 16th International Conference on Enterprise Information Systems (ICEIS-2014), pages 653-658

ISBN: 978-989-758-028-4

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

performed by their users.

This paper is structured as follows. Section 2

discusses the present and future trends for Semantic

Web implementation. In Section 3 we present the e-

swim, the Semantic Web Implementation Model for

enterprises proposed in this paper. In Section 4 we

outline the research methodology used to validate

the model and include some preliminary findings.

Section 5 concludes with a summary of the project

achievements and directions for future work.

2 THE SEMANTIC WEB:

PRESENT AND FUTURE

TRENDS

The economic dimension is one of the most

important dimensions of the World Wide Web. It

has been so in its first generation, with “long tail”

enterprises approaching the global market of

customers and products, until the targeting of the

smallest segment. It has also been so with the

growth of social networks, and these being explored

as a competitive advantage to connect people to

enterprises, brands and products. And it will

eventually be so in its third generation, with

enterprises actively participating in the construction

of the Semantic Web, the Web of Data. It constitutes

certainly a great and broad research opportunity, in

many subjects, allowing the space for the creation of

the Web Science (Berners-Lee, Hendler, Hall,

Shadbolt, & Weitzner, 2006; Ferreira & Seruca,

2013a).

Semantic Web technology may transform

enterprise software, contribute to the emergence of

new business models and reduce costs in areas like

data integration, master data management and

enterprise information management (Pollock, 2008).

Enterprise internal networks based on Linked Data

principles constitute a particular subset of Semantic

Web technologies that, amongst other benefits, may

substantially reduce information integration costs,

such as in the integration of information about a

product, supplier, materials, legal, market and

finance data or other internal and external data

sources (Janowicz & Hitzler, 2010). However, the

growing adoption of Semantic Web technologies

and Linked Data principles raise the question of

which applications may be developed to take

advantage of this potential. The answer may be

obtained from identifying the areas where these

technologies and principles may constitute a

distinctive contribution, when compared with

traditional technologies (Heath, 2010).

Meanwhile, the compromise between the issues

of computational effort and flexibility tends to

favour the latter, in line with Moore’s law

projection, since additional computing power leads

to less concerns about simplification or optimization.

Simultaneously, the economy has benefited from

technological innovation and became a highly

competitive environment where speed and flexibility

often play a more important role than robustness and

trust. The economic validity of current data

normalization models can therefore be questioned in

comparison to the flexibility and universality

promises of the Semantic Web (Segaran, Evans, &

Taylor, 2009).

The Semantic Web development has been guided

by the way different communities envisage its

evolution, considering their specific areas of

research. One approach is that of semantic

annotation, addressing the large volumes of data

available on the Web and using different techniques

to originate structured data. Another approach is that

of data repositories, starting from pre-defined

structures that are updated and interlinked with

additional structures. Finally, the approach that puts

forward the Semantic Web as an agent platform,

with applications combining different data sources

and, ultimately, executing actions in replacement of

individuals (Domingue, Fensel, & Hendler, 2011).

Additionally, considering the diverse nature of

organizations, it seems reasonable that the enterprise

approach may be different from the academic one.

Enterprises will look for more practical results,

focusing on the short term and higher success rates,

while academic research will typically look for

theoretical results that are more ambitious and

focused on the medium term. However, the fact that

too much divergence may cause a fracture, with

enterprises focusing on a small number of issues and

academia in issues that will never be tackled

(Cardoso, Miller, Su, & Pollock, 2008), suggests to

take a balanced view between both approaches.

The debate over the possible commercial success

of the Semantic Web is exhausted and has been

replaced by the discussion of what changes in

commercial software applications may occur with

the introduction of the technology. Some authors

suggest as possible evolutions the proliferation of

highly distributed applications, agile development

and dynamic integration of legacy applications,

sensor networks and decision support systems

(Domingue et al., 2011).

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

654

3 THE ENTERPRISE SEMANTIC

WEB IMPLEMENTATION

MODEL

Acknowledging the different issues to consider in

the adoption of Semantic Web technologies in

Enterprises, it is envisaged that a model for

Semantic Web implementation in Enterprises will

constitute a useful approach to guide enterprises in

the introduction of the technology as well as in the

evaluation of its’ impact in the applications used and

tasks performed by users in these organizations

(Ferreira & Seruca, 2013b).

The model hereby presented is called “e-swim”,

an acronym for Enterprise Semantic Web

Implementation Model, and is based in four

dimensions, where each dimension considers

different requirements for the technology

implementation and, simultaneously, serves as a

guide to identify the desired continuous evolution

through subsequent steps.

Figure 1 illustrates the e-swim model and its four

dimensions: Adoption, Provenance, Accessibility

and Activities. The following sub-sections describe

the rationale for each dimension.

Figure 1: The e-swim Model.

3.1 Adoption

The purpose of this dimension is to determine the

degree of preparation of the Enterprise to adopt

Semantic Web technologies.

The main technological adoption models for

enterprise application are the Innovation Diffusion

Model and the Technology-Organization-

Environment Model (Oliveira & Martins, 2011).

According to the Innovation Diffusion Model

(Rogers, 1995), organizational innovation is

essentially dependent on Leadership, Organizational

Structure and Openness and is adopted according to

a normal distribution of organizations that includes

Innovators, Early Adopters, Early Majority, Late

Majority and Laggards. The Technology-

Organization-Environment Model proposes three

issues of the enterprise context that influence

technological innovation, namely Technology,

Organization and Environment. By combining both

approaches, the e-swim Model considers the

following features for the dimension:

 Technology, including opportunities of

technology usage in the organization, in this

case, Semantic Web technologies and its

applications;

 Organization, referring to internal organization

relations and including:

o Leadership, attitude towards change from

the top management

o Structure, relations between people in the

organization

 Exterior, about the external framing of the

organization, including:

o Interface, openness of the organization to

the outside

o Environment, players in the space where

the organization is located

These features are summarized in a graphical

form in Figure 2.

Figure 2: The technology Adoption dimension.

The planned intervention in enterprises takes

these features into account and, through observation

and interrogation, tries to quantify the position of

each Enterprise in the path from a lower to a higher

technology adopter.

3.2 Provenance

This dimension of the Implementation Model

considers data provenance as a determinant factor.

Enterprise innovation and competitive advantage

e-swim:EnterpriseSemanticWebImplementationModel-TowardsaSystematicApproachtoImplementtheSemantic

WebinEnterprises

655

depend entirely of its capacity to deal with a

constant and always growing information flow.

Consequently, information integration efforts must

follow that growth. Semantic Web technologies

usage in those integration efforts may increase

substantially the return, reducing integration costs

and increasing subsequent benefits (Janowicz &

Hitzler, 2010). Opening public data to citizens

represents an increasing democratic transparency,

possible due to technological availability. In that

area, the efforts of the American and British

governments, among others, already led to a broad

availability of data sources with wide usage

possibilities. The need to explore these data sources

reveals itself primarily as the possibility to explore

wealth sources (Koumenides, Salvadores, Alani, &

Shadbolt, 2010).

Therefore, the purpose of this dimension, as

illustrated in Figure 3, is to identify those data

sources, according to:

 applications topology, namely Web Sites,

Extranets, Intranets, Web Applications or Web

Services

 location: outside or inside the enterprise

Figure 3: Enterprise Data Provenance.

The planned intervention aims to identify the

data sources in the context of the enterprise

Information Systems and to quantify their respective

usage.

3.3 Accessibility

The common user perception about data available in

the Web is largely influenced by the format of these

data. However, the structure of these data can

influence their subsequent usage. The path to the

universal availability of data, as shown in Figure 4,

was clearly identified by Tim Berners-Lee with a

five stars classification or five steps of evolution

(Berners-Lee, 2006):

 available: just having data available in the Web

 formatted: through proprietary formats

 open: with open formats

 semantic: via semantic standards

 interconnected: with data hyperlinks.

Figure 4: Data Accessibility, adapted from (Berners-Lee,

2006).

The planned intervention will determine the

degree of openness of data sources and consequently

quantify their suitability towards Semantic Web

standards.

3.4 Activities

Information Systems have been classified according

to different approaches, some with a broader scope,

others with more specific purposes (Lopes, Morais,

& Carvalho, 2005). Despite the high number of

efforts, the issue is not exhausted and, in this

particular case, it is important to find a classification

that positions the Web as a base environment, in

alternative or in complement to more traditional

classifications. With several studies related with

Web usage, Tom Heath’s work introduces important

clarifications and a purpose oriented classification

(Heath, 2010). According to this classification, user

activities may be instantiated in the following

categories:

 Locating: look, find

 Exploring: gather, research

 Grazing: navigate, browse, follow

 Monitoring: monitor, check, detect

 Sharing: distribute, collaborate

 Notifying: state, inform, communicate

 Asserting: opinion, suggestion

 Discussing: comment, respond

 Evaluating: assess, analyse

 Arranging: combine, negotiate

 Transacting: transfer, pay

Figure 5 shows the space for Semantic Web

technologies implementation according to the

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

656

degrees of diversity and conceptualization of the

tasks performed.

Figure 5: Activities over Conceptualization and Diversity,

adapted from (Domingue et al., 2011).

The conceptualization and diversity degrees of

the tasks performed influence technology

implementation feasibility. Simple and repetitive

tasks are those naturally already satisfied by

traditional software applications and that will less

benefit from Semantic Web technologies. As task

conceptualization increases, opportunities arise for

the technology, but in the conceptualization

threshold, when tasks are hardly typified and

demand a person’s creative intervention, the effort

for their implementation is simply higher than that

of just executing the tasks. Low diversity tasks are

once again easily supported by traditional software

and in the threshold of diversity it will rarely be

worth systematizing tasks to be supported by the

technology. Hence, the ideal space for Semantic

Web technology implementation will be that of

some conceptualization or diversity that make tasks

too complex or too diverse for traditional software

(Domingue et al., 2011). By positioning the previous

classified activities in this space, the opportunity for

this technology can be visualized.

This dimension addresses the wide area of the

enterprise Information Systems and introduces a fit

to action approach adequate to an Action-Research

effort. The planned intervention will determine what

are the tasks performed that may better incorporate

Semantic Web technologies and quantify their usage

through the Enterprise.

4 MODEL VALIDATION

In order to validate the proposed model, an Action-

Research based intervention was planned in several

enterprises. Previous work conducted identified the

low awareness for Web technologies and lack of

collaboration between enterprises, academia and

government (Ferreira, 2013), which justifies the

need for an immersive approach, where all parts

involved work for a common goal. The approach to

take would have to consider, within the scope of the

proposed implementation model, the need to provide

technological leadership, develop awareness for

Web technologies, facilitate access to training where

appropriate, identify and quantify available

resources, qualify the tasks performed and involve

the participants in a collaborative effort to, whenever

possible, implement change recommendations.

The planned intervention is based on two cycles,

where the first cycle aims to identify

recommendations for each individual enterprise and

the second to share and reapply the most relevant

recommendations with all the participant enterprises.

The first results of this intervention are promising

and challenging, with several recommendations

being currently developed. Some of these

recommendations are clearly adequate for wide

reapplication, benefiting the outcomes of this

research work, while others are more specific to

each enterprise, contributing to the satisfaction of

individual expectations.

5 CONCLUSIONS

Semantic Web constitutes an innovation opportunity

for Enterprises. This research project identified

implementation barriers, such as the organizational

structure, as well as important issues to address,

namely what are the enterprise software applications

that may benefit with the technology introduction.

This paper proposes an Enterprise Semantic Web

Implementation Model (e-swim) based in four

dimensions:

 Adoption

 Provenance

 Accessibility

 Activities

The Implementation Model hereby proposed as

well as the results obtained with the planned

intervention in several enterprises should facilitate

the measurement of the implementation degree of

the technology in an Enterprise and serve as a guide

for the Semantic Web technology implementation by

providing a roadmap for a quicker and more

intensive adoption as well as pointing possible

evolutions in the process.

e-swim:EnterpriseSemanticWebImplementationModel-TowardsaSystematicApproachtoImplementtheSemantic

WebinEnterprises

657

REFERENCES

Ahmed, Z., & Gerhard, D. (2010). Role of Ontology in

Semantic Web Development. Knowledge

Management.

Berners-Lee, T. (2006). Linked Data Desing Principles.

Retrieved July 14, 2012, from http://www.w3.org/

DesignIssues/LinkedData.html.

Berners-Lee, T., Hendler, J., Hall, W., Shadbolt, N., &

Weitzner, D. J. (2006). Computer science. Creating a

science of the Web. Science, 313(5788), 769–771.

Bizer, C., Heath, T., Idehen, K., & Berners-Lee, T. (2008).

Linked data on the web (LDOW2008). In WWW2008

Workshop on Linked Data on the Web (Vol. 2008, pp.

1265–1266). ACM New York, NY, USA.

doi:10.1145/1367497.1367760.

Cardoso, J., Miller, J., Su, J., & Pollock, J. (2008).

Academic and Industrial Research : Do their

Approaches Differ in Adding Semantics to Web

Services ? Lecture Notes in Computer Science,

3387/2005, 14–21. doi:10.1007/978-3-540-30581-1_2.

Domingue, J., Fensel, D., & Hendler, J. A. (Eds.). (2011).

Handbook of Semantic Web Technologies. Berlin,

Heidelberg: Springer Berlin Heidelberg. doi:10.1007/

978-3-540-92913-0.

Ferreira, R. (2013). Applying Semantic Web technologies

to City Tourism information. In 1st International

Conference - Porto as a Tourism Destination. Porto,

Portugal: CEPESE.

Ferreira, R., & Seruca, I. (2013a). A Ciência da Web :

oportunidades de investigação. In 13

Conferência da

Associação Portuguesa de Sistemas de Informação

(pp. 13–31).

Ferreira, R., & Seruca, I. (2013b). Modelo de

Implementação da Web Semântica nas Empresas. In

Á. Rocha, L. P. Reis, M. P. Cota, M. Painho, & M. C.

Neto (Eds.), Actas da 8

Conferência Ibérica de

Sistemas e Tecnologias de Informação (CISTI’2013),

Lisboa, Portugal, 19-22 Junho 2013 (pp. 615–620).

Lisbon: AISTI.

Heath, T. (2010). A taskonomy for the Semantic Web.

Semantic Web Journal, 1, 75–81. doi:10.3233/SW-

2010-0003.

Heath, T., & Bizer, C. (2011). Linked Data: Evolving the

Web into a Global Data Space. In J. Hendler & F. Van

Harmelen (Eds.), Proceedings of the 17th

international conference on World Wide Web WWW

08 (Vol. 1, pp. 1–136). Morgan & Claypool.

doi:10.1145/1367497.1367760.

Janowicz, K., & Hitzler, P. (2010). Creating knowledge

out of interlinked data. Semantic Web, 1, 97–104.

doi:10.3233/SW-2010-0019.

Kang, S. S., Yang, J. Y., Lee, S. K., Gong, K. H., Myung,

J. S., Park, S. C., & Lee, S. G. (2008). An Enterprise

Strategy for Semantic Technology Adoption. In The

5th International Conference on Information

Technology and Applications.

Koumenides, C. L., Salvadores, M., Alani, H., &

Shadbolt, N. R. (2010). Global Integration of Public

Sector Information. In Proceedings of the WebSci10:

Extending the Frontiers of Society On-Line, April 26-

27th, 2010, Raleigh, NC: US.

Kuhn, W. (2010). Modeling vs encoding for the Semantic

Web. Semantic Web Journal, 1, 11–15. doi:10.3233/

SW-2010-0012.

Lopes, F. C., Morais, M. P., & Carvalho, A. J. (2005).

Desenvolvimento de Sistemas de Informação (FCA., p.

209).

Oliveira, T., & Martins, M. (2011). Literature Review of

Information Technology Adoption Models at Firm

Level. The Electronic Journal Information Systems

Evaluation, 14(1), 110–121.

Pollock, J. (2008). A Semantic Web Business Case. W3C.

Retrieved July 14, 2012, from http://www.w3.org/

2001/sw/sweo/public/BusinessCase/BusinessCase.pdf.

Rogers, E. M. (1995). Diffusion of innovations. (M. B.

Salwen & D. W. Stacks, Eds.)An integrated approach

to communication theory and research (Vol. 65, p.

519). Free Press. doi:10.1525/aa.1963.65.5.02a00230.

Segaran, T., Evans, C., & Taylor, J. (2009). Programming

the Semantic Web. (M. Treseler, Ed.) (p. 300).

O’Reilly.

W3C. (2013). W3C Semantic Web Activity Homepage.

Retrieved from http://www.w3.org/2001/sw/

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

658