An Advanced Ontology-Based Approach for Semantic Service Identification

Resumo

Service identification is a key challenge in the migration of legacy systems toward a Service-Oriented Architecture (SOA). Most existing approaches have addressed this issue at the source code level. In contrast, this work proposes a model-driven solution for identifying services at the model level. The proposed approach is based on the construction of an OWL ontology that captures the knowledge embedded in a UML class diagram, including both entities and their associated operations. Using a set of  transformation rules, the approach makes explicit several properties and relationships between classes that are not directly visible in the original class diagram. Although the transformation into an OWL ontology provides a rich semantic representation of system knowledge, an important challenge remains: enabling the effective exploitation of the identified concepts and operations within concrete application contexts.

To overcome this limitation, the approach is extended to support service identification directly from the resulting OWL ontology. The extracted services provide operational interfaces for the modeled concepts and relationships, facilitating their integration into software systems. This process enables structured and reusable access to ontological knowledge, supports the management of complex operations, and improves system maintainability by allowing services to evolve independently while preserving semantic coherence.