Validating enterprise architecture using ontology-based approach: a case study of student internship programme

As the practice of Enterprise Architecture (EA) diversifies, the schematic management of its objects, semantics and relationship continues to be complex. Given that EA provides support for the IT environment by simulating alignment between dynamic business architectures, heterogeneous application systems and incongruent technologies, the need to ensure validation of conceptualized EA models has become also critical. As a relatively new discipline, its disparate and widespread methodologies makes it even more challenging to adopt a generic approach in which models can be verified sequel to the unavailability of unified EA modelling languages able to describe a wide range of Information Technology domains. This paper seeks to present an approach for addressing this challenge through the use of ontologies and queries based on constraints specified in the model's motivation taxonomy. The paper is experimental research-based and grounds its hypothesis on initial model created using the ArchiMate modelling language. By transforming its conceptual metamodel into a model instance, a process which can be achieved irrespective of the modeling language used in the design of the EA, it represents extracted triples as resource description framework schema (RDFS) by mapping the model artefacts directly into classes and slots using a more conventional web ontology language. The generated RDF is then queried using Protocol and RDF Query Language (SPARQL) adopting the Behavior Driven Development (BDD) concept. A case study of the Student Internship Program (SIP) is deployed to translate information from business needs to IT solutions encapsulating a view of abstraction of the EA. The paper also proposes an implementation of the approach using an open source platform that allows construction of domain models and knowledge-based applications with ontologies and is a contribution towards effective validation of EA through taxonomy decomposition, systematic generation of testable EA artifacts, creation of structural triples of model elements and alignment of motivational goals to business behavior specifications