INFOFLEX: FLEXIBLE AND DISTRIBUTED CONTENT

MANAGEMENT

USING WEB SERVICES AND SEMANTIC WEB TO MANAGE CONTENT

Jesús Villamor-Lugo, Norberto Fernández-García, Luis Sánchez-Fernández, Jesús Arias-Fisteus

Department of Telematic Engineering,Universidad Carlos III de Madrid, Spain

Tomás Nogales-Flores, Antonio Hernández-Pérez, David Rodríguez-Mateos

Department of Librarianship and Information Science, Universidad Carlos III de Madrid, Spain

Key

words: Content Management, Web Service

s, Semantic Web

Abstract: The development of information and communication technologies and

the expansion of the Internet means

that nowadays there are huge amounts of information available via these emergent media. The need to

manage such information, which was in the past stored on paper media, has become apparent in different

fields. A number of content management systems have appeared which aim to achieve this task. Most of

these systems are oriented towards Web publishing on a central site, and they do not support collaboration

among several, distributed sources of managed content. In this paper we present a proposal for an

architecture for the efficient and flexible management of distributed content.

1 INTRODUCTION

At different application domains, there exists the

need to offer a service that permits managing,

producing and offering quality information. The

entities devoted to the production and/or the register

of contents must face several requirements. Some of

them are:

• Content cr

eation

and manipulation

management. A mechanism for the

coordination of related work among different

persons is needed.

• User Pr

ofile Management

. At an environment

in which the contents are created and accessed,

we must be able to define mechanisms to

identify who must create, modify or query some

documents. It may also be necessary to provide

some document review mechanisms, as

previous step for the internal or public

publishing.

• Decision criteria on the

conten

ts to be

published. Two types of factors must be

considered in the decision process: technical

ones (i.e. which contents will be published on

the Web site, links validation, etc) and political

ones (i.e. trust degree on the information,

confidentiality, etc).

• Inte

gration of several fragments in on

document (i.e. for constructing a web page

from several content sources).

• Va

lidation. T

he contents and functionalities

offered by the system must be checked in order

to detect errors or inconsistencies.

In this paper, we show an architecture

owaday

s at development process) for the

management of flexible and distributed contents: the

Infoflex System (Sánchez et al., 2003). It offers the

required facilities for a Content Management System

(CMS). These facilities are currently being

implemented using the following technologies:

• XM

L for the description (or encapsulation,

if tex

t data) of the information. The XML

documents are stored in an XML Database

613

Villamor-Lugo J., Fernández-García N., Sánchez-Fernández L., Arias-Fisteus J., Nogales-Flores T., Hernández-Pérez A. and Rodríguez-Mateos D.

(2004).

INFOFLEX: FLEXIBLE AND DISTRIBUTED CONTENT MANAGEMENT - USING WEB SERVICES AND SEMANTIC WEB TO MANAGE CONTENT.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 613-616

DOI: 10.5220/0002656606130616

 SciTePress

to ease the recovering. Likewise, the use of

XML will permit to export the contents to

different formats (using XSLT style

sheets).

• A light-weight workflow management

system.

• A version management system.

• Web Services (WS), to deal with collaboration

among different distributed CMSs.

• Business Process Execution Language for

Web Services (BPEL4WS) (Andrews et al.,

2003), to define a complex interaction based on

the coordination of different Web Services.

• Semantic Web, to deal with heterogeneity in

data models.

The novelty of Infoflex is the use of Web

Services and Semantic Web to allow a user (human

or informatics system) to retrieve information from

different distributed and heterogeneous management

systems (even when their location is unknown) and

to perform a unique query on several systems.

This paper is organized as follows. Section 2

introduces Web Services and how they are used in

our framework. Section 3 introduces Semantic Web

and its use in Infoflex. Section 4 describes the

proposed working model and architecture.

Concluding remarks complete this paper.

The Web was born with the idea of making

information accessible by humans. But with the

quick Web growth, the amount of data available is

so big, that, to accede it in an efficient manner, new

automatic mechanisms are required. On the other

hand, nowadays the Web is not only information but

also services. For final users, it would be interesting

that software agents could use such services in an

automatic manner, freeing humans of tedious tasks.

2 WEB SERVICES AND ITS

APPLICATION IN INFOFLEX

Web Services (Web Services, 2003) have been

defined as self-describing applications that can

discover and engage other web applications to

complete complex tasks over the Internet. This

technology was designed to ease the cooperation

among remote applications. Web Services

technology is supported by a set of companies (IBM,

HP, SUN, Microsoft, …) and standardization

organisms like the World Wide Web Consortium

(W3C).

The Web Services architecture is a middleware

technology based on XML data formats for encoding

the exchanged messages (SOAP), service definitions

(WSDL), Service registering-and-discovering

facilities (UDDI) and Internet standard protocols (as

HTTP or SMTP).

The interoperability model proposed by Web

Services perfectly fits our distributed contents access

requirements.

Initially, users don’t know the location of the

providers. So Infoflex must allow a user (human or

software application) to query and retrieve

information in several distributed CMSs in the

desired domain (news agencies, content syndication

sites, etc). In our framework, CMSs provide access

to their functionalities through Web Services. This

way, clients use UDDI repositories to look for them.

The choice of Web Services as communication

mechanism provides two advantages:

• Interoperability and platform

independence, because Web Services

provide a standard text-based

communication mechanism.

• Data format compatibility, because CMSs

typically store their data as XML

documents.

3 THE SEMANTIC WEB AND ITS

APPLICATION IN INFOFLEX

This desired degree of automation could only be

achieved if we allow machines (and not only

humans) to understand resources and contents. This

requires giving meaning (semantics) to data: we

need a Semantic Web (Berners-Lee, 2001). With this

purpose the Semantic Web Activity Group

(Semantic Web, 2003) of W3C began its activities in

2001. Two are the main development areas:

• Metadata: the objective is the development of

mechanisms that allow the description of

contents, represented in web publishing formats

like HTML or XML, in a machine-readable

format. Perhaps the most important

development in this field is the RDF (Resource

Description Framework) recommendation

(RDF, 2003).

• Ontologies: if we want contents to be

interpreted by machines, it is not enough to

provide machine-readable descriptions: we also

need machine-understandable descriptions.

ICEIS 2004 - DATABASES AND INFORMATION SYSTEMS INTEGRATION

614

Ontologies are used with this purpose. An

ontology is a model of a knowledge domain,

that is, a set of concepts that are typically used

in that domain and a set of relations among this

concepts. These models are formal enough to

allow machines to reason about them. This way,

inferences can be made automatically and

knowledge could be improved. When we want

to describe a resource using metadata, we do not

do it using natural language but the set of

concepts included in the ontologies. So,

machines could understand this metadata and

the contents described by them. In this field the

most important contribution of the W3C is

OWL (Ontology Web Language) (OWL, 2003).

This XML application is currently under

development.

Infoflex uses these technologies to face the

problem of accessing heterogeneous sources of

information. A domain can be modelled with many

different ontologies. To allow interoperation, we

propose to use a general and consensual ontology.

This ontology will be written in OWL. It will

describe all the concepts and relations of interest.

Every CMS must develope its own ontology, which

is used to describe its contents. These CMS-specific

ontologies must use the general ontology as a

reference in one of the following ways:

The specific ontology can be just a subset of the

general ontology.

• The specific ontology can be used for

translating between the general ontology and

another internal (and typically unknown)

ontology used by the CMS to describe its

information.

• This idea of using ontologies to deal with

heterogeneity of information sources has been

previously addressed in other papers like

(Brisaboa, 2002).

To use a general ontology, it must be accessible

for all CMSs, so we propose to store this ontology in

a global service. This service is called Query Service

(QS) and it will be implemented as a Web Service.

All CMSs must access it, query the general

ontology, and register their specific ontologies.

Clients must also query this QS because they

typically do not know the CMS data-model and

query language. So, they must query the QS, using a

commonly-accepted language and data-model (the

general ontology). As a result, they obtain a specific

query adapted to the CMS data-model and language.

This way, heterogeneity in data models and

formats is hidden to final clients. Next section

explains this process in detail.

4 ARCHITECTURE AND

WORKING MODEL

In this section we will introduce the proposed

working model of our Infoflex framework.

The main actors involved in the full process are:

• Clients, who request information using the

general query language. This language is not

specified yet. It will be based on XML

(XQuery, RDF, or similar).

• Query Servers, which translate queries from

the model represented in the general ontology

(used by the Client to make his queries) to the

CMS information model.

• Content Management Systems, which store

and manage the requested information.

• UDDI repositories, where all the Web Services

involved on this scenario are registered.

Figure 1: Architecture and Working Model

In figure 1, we can see the steps which are

necessary to obtain information from one provider.

Without loss of generality, we suppose in our

exposition that only one component of each kind is

available.

Basically, the steps are:

INFOFLEX: FLEXIBLE AND DISTRIBUTED CONTENT MANAGEMENT - USING WEB SERVICES AND

SEMANTIC WEB TO MANAGE CONTENT

615

1. The Query Server registers two Web Services

into a UDDI repository. One WS (WSc) is for

Client access (to make queries). The other one

(WSp) is for CMS access (to register its own

information model ontology in the Query

Server)

2. The CMS access the UDDI looking for the

Query Server. It obtains a reference to the WSp

of this Query Server.

3. The CMS registers its ontology in the Query

Server, whose reference has been obtained in

the previous step.

4. The CMS generates dinamically a BPEL4WS

document. This document specifies the steps

(WS invocations) that Clients must follow to

query it. Two WS invocations are necessary at

least: one for translating the query (operation

done by the Query Server) and the other one to

achieve the information (querying a Content

Provider WS). The BPEL4WS document also

contains information on how to compose the

results of such invocations (use the responses of

one service as the parameters of the other one)

and how to react when a unexpected situation

occurs (for example, the Query Server is

unreachable). This document must be obtained

by the Client to enact it and access the CMS, so

this Provider must publish its BPEL4WS. In our

proposal, the document is retrieved using a WS,

which is registered by the CMS in a UDDI.

5. The Client access the UDDI looking for the

CMS. It obtains the reference of its WS.

6. The Client invokes the Web Service of the

CMS. It obtains as the result the BPEL4WS

document. The Client executes this document to

retrieve the desired information.

7. During the execution, the Client writes a query,

using concepts in the general ontology. The

BPEL4WS execution engine uses this query (as

is indicated in the BPEL4WS document) to

invoke the QS, which translates it to the CMS

model.

8. The obtained traslated query is automatically

used by the BPEL4WS engine to invoke the

CMS and retrieve the desired information (if it

is avaliable).

5 CONCLUSIONS

In this paper we have presented a proposal of an

architecture for the management of contents in a

flexible and distributed manner. This proposal is

based on advanced Web technologies such as XML,

Web Services or Semantic Web. We expect for this

system to be useful at environments like business to

business portals, news agencies or content

syndication sites.

ACKNOWLEDGMENTS

This work has been partially financed by the

“Ministerio de Ciencia y Tecnología de España”

(Science and Technology Ministry of Spain) through

the TIC2003-07208 "InfoFlex" Project.

REFERENCES

Berners-Lee, T., 2001; Hendler, J.; Lassila, O. “The

Semantic Web”. "Scientific American". Vol. 284, nº 5.

2001. pp. 34-43.

Brisaboa, N., 2002; Penabad, M.; Places, A.; Rodríguez,

F. "Ontologies for Database Federation".

"UPGRADE". Vol. III, n. 3. 2002. pp. 52-61.

World Wide Web Consortium. Web Services Activity.

Consulted at: http://www.w3.org/2002/ws/ (28-10-

2003).

World Wide Web Consortium. Semantic Web Activity.

Consulted at: http://www.w3.org/2001/sw/ (28-10-

2003).

World Wide Web Consortium. Resource Description

Framework (RDF). Consulted at:

http://www.w3.org/RDF/ (28-10-2003).

World Wide Web Consortium. OWL Web Ontology

Language Reference. Consulted at:

http://www.w3.org/TR/owl-ref/ (28-10-2003).

Sánchez L., Villamor J., Arias J., Fernández N., Nogales

T., Hernández A., Rodríguez, D., 2003 “Gestión

Flexible de Contenidos Distribuidos”. Boletín de RedIris,

diciembre 2003-enero 2004, Vol. 22.

Andrews T., Curbera F., Dholakia H., et al., 2003;

“Business Process Execution Language for Web

Services. Version 1.1 Specification.” Available at:

http://www-106.ibm.com/developerworks/webservices/

library/ws-bpel (28-10-2003)

N. Guarino (ed.), 1998; “Formal Ontology in Information

Systems”. Proceedings of FOIS’98. Amsterdam, IOS

Press, pp. 3-15.

Documentum, Enterprise Content Management. Consulted

at: http://www.documentum.com/ (28-10-2003)

Microsoft Content Management Server. Avaliable at :

http://www.microsoft.com/cmserver/ (28-10-2003)

Zhou N., Meliksetian D., Weitzman L., et al. 2001; “XML

Content Management: Challenges and Solutions”.

Proceedings of XMLEurope 2001.

ICEIS 2004 - DATABASES AND INFORMATION SYSTEMS INTEGRATION

616