Sign up & Download
Sign in

Automatically Grounding Semantically-Enriched Conceptual Models to Concrete Web Services

by Eran Toch, Avigdor Gal, Dov Dori
Conceptual Modeling–ER 2005 (2005)

Abstract

The paper provides a conceptual framework for designing and executing business processes using semantic Web services. We envision a world in which a designer defines a virtual Web service as part of a business process, while requiring the system to seek actual Web services that match the specifications of the designer and can be invoked whenever the virtual Web service is activated. Taking a conceptual modeling approach, the relationships between ontology concepts and syntactic Web services are identified. We then propose a generic algorithm for ranking top-K Web services in a decreasing order of their benefit vis-á-vis the semantic Web service. We conclude with an extention of the framework to handle uncertainty as a result of concept mismatch and the desired properties of a schema matching algorithm to support Web service identification.

Cite this document (BETA)

Available from www.springerlink.com
Page 1
hidden

Automatically Grounding Semantically-Enriched Conceptual Models to Concrete Web Services

Automatically Grounding Semantically-enriched
Conceptual Models to Concrete Web Services
Eran Toch, Avigdor Gal and Dov Dori
Technion – Israel Institute of Technology
Technion City, Haifa 32000, Israel
Abstract. The paper provides a conceptual framework for designing and exe-
cuting business processes using semantic Web services. We envision a world in
which a designer defines a “virtual“ Web service as part of a business process,
while requiring the system to seek actual Web services that match the specifi-
cations of the designer and can be invoked whenever the virtual Web service
is activated. Taking a conceptual modeling approach, the relationships between
ontology concepts and syntactic Web services are identified. We then propose a
generic algorithm for ranking top-K Web services in a decreasing order of their
benefit vis-a´-vis the semantic Web service. We conclude with an extention of the
framework to handle uncertainty as a result of concept mismatch and the desired
properties of a schema matching algorithm to support Web service identification.
1 Introduction
Web services allow universal connectivity and interoperability of applications and ser-
vices, using well-accepted standards as UDDI, WSDL, and SOAP. Current Web ser-
vice standards focus on syntactic, operational details for implementation and execution
rather than semantic capabilities description. A recent development enables the spec-
ification of semantic Web services. The semantic Web [5] aims to extend the World-
Wide-Web by representing data on the Web in a meaningful and machine-interpretable
form. The semantic Web is based on a set of languages that provide well-defined se-
mantics and enable the markup of complex taxonomic and relations between entities on
the Web. Ontologies, commonly defined as specifications of a conceptualization, [20]
serve as the key mechanism for the semantic Web by allowing concepts to be globally
defined and referenced. A leading language for ontology modeling for the semantic
Web is the Web Ontology language (OWL) [9], providing a semantic markup for the
definition of concept classes, relationships among them, and their instances. Several
methods for annotating Web services with semantic metadata have been proposed. One
of the most prominent methods is OWL-S [3]. Based on OWL, it provides an ontology
for Web services, enabling a description of the service’s profile, process model and its
grounding - a mapping to the syntactic definition of the concrete Web service.
We envision a world in which some Web services have semantic descriptions, while
others are only syntactically defined (using WSDL, for example). In particular, design-
ers can define a “virtual” Web service as part of their business processes. An execution
engine is required to look for actual Web services that match the specifications of the
Page 2
hidden
designer and can be invoked whenever the virtual Web service is activated. The design
of semantic Web services can be an iterative process, starting from rough design, and
gradually refine the design based on feedback from some mechanism that grounds the
semantic Web service to some existing Web services.
It is the aim of this paper to provide a framework for a model-driven design using
semantic Web services. Taking a conceptual modeling approach, relationships between
ontology concepts and syntactic Web services are identified. We then propose a generic
algorithm for ranking top-K Web services in a decreasing order of their benefit vis-a´-vis
the semantic Web service. We conclude with a discussion on extending the framework
to handle uncertainty that stems from concept mismatch and the desired properties of a
schema matching algorithm to support Web service identification.
The main contribution of this work is twofold. At the conceptual level, we introduce
a method for designing business processes as a composite set of Web services. At the
algorithmic level, we provide a generic algorithm for ranking concrete Web services
with respect to their suitability in fitting a semantic Web service description, in effect
offering a model-driven approach for service-oriented computing. The semantic Web
service serves as a conceptual model. Rather than generating a code out of the model,
the model is implemented by locating and invoking existing services. It is worth noting
that the concrete Web services are not necessarily annotated with semantic meta-data,
and may be described as WSDL documents, reflecting the current state of affairs. Fi-
nally, we discuss the characterization of requirements for a schema matching algorithm
should satisfy to qualify for interfacing with the Semantic Web.
The rest of the paper is organized as follows. Section 2 presents the model and
formally defines the problem. The use of ontologies in ranking Web services is given
in Section 3, followed by an algorithm for the matching process (Section 4). Section 5
discusses an extension to support semantic heterogeneity. Section 6 contains a related
work. The paper concludes with a summary and future work (Section 7).
2 Model and Problem Definition
In this section, we provide a formal definition of the two main elements of our model,
namely Web services (Section 2.1) and Semantic Web services (Section 2.2). We con-
clude with a formal introduction of the problem at hand (Section 2.3).
2.1 Web Services
Web services are loosely coupled software components, published and invoked across
the Web. Several XML-based standards ensure the regulation of discovery and the in-
teraction of Web services. In particular, UDDI allows Web services to be discovered
through a keywords search. A Web Services Description Language (WSDL) document
describes the interface and communication protocol of Web services. In this paper, we
use restricted WSDL definition, ignoring namespaces, faults handling, and communi-
cation issues. Therefore, a Web service is a quadruple,WS = (T,M,O,A), where:
– T is a finite set of types. A type can be primitive (e.g., integer) or complex, de-
scribed by an XML schema.

Sign up today - FREE

Mendeley saves you time finding and organizing research. Learn more

  • All your research in one place
  • Add and import papers easily
  • Access it anywhere, anytime

Start using Mendeley in seconds!

Already have an account? Sign in

Readership Statistics

5 Readers on Mendeley
by Discipline
 
 
by Academic Status
 
40% Researcher (at an Academic Institution)
 
20% Professor
 
20% Associate Professor
by Country
 
40% France
 
20% Germany
 
20% Spain