Improving it service management using an ontology-based and model-driven approach

Résumé

Best practice frameworks, focused on the integration of business and Information Technology (IT), help organizations create and share effective IT Service Management (ITSM) processes. However, service management guidelines and models are commonly specified using natural language or graphical representations, both lacking the computational semantics needed to enable their automated validation, simulation or execution. This thesis proposes Onto-ITIL, an ontological and model-driven approach that captures the best practices provided by the IT Infrastructure Library (ITIL) framework. OntoITIL aims to help domain experts to model and implement ITSM processes avoiding semantic ambiguities, uncertainties and contradictions. Formalizing ITSM processes in terms of ITIL is as a first step towards bridging the current gap between business and IT. Onto-ITIL has been defined using the Web Ontology Language (OWL) and has been enriched with a set of rules defined using the Semantic Web Rule Language (SWRL) to provide semantic constraints and knowledge inference. The definition of a set of queries based on the Semantic Query-Enhanced Web Rule Language (SQWRL) enables the retrieval of knowledge from OWL and inferred by the SWRL rules. Onto-ITIL not only enables the formal specification of ITSM processes, but also to share, reuse, and interchange these specifications by automated means using e-business frameworks such as ebXML. Adopting a Model-driven Engineering (MDE) approach, a Domain Specific Language (DSL), based on Onto-ITIL, is used in order to implement workflow-based information systems that underpin ITSM Systems (ITSMSs). The resulting high-level models enriched with ontological knowledge are defined just in terms of the business logic, without any architectural or platform-specific consideration. That is, according to the OMG's four-layered architecture, the ontology-based workflow models could be placed at a Computation Independent Model (CIM) level. In summary, the approach presented in this thesis aims: (i) to formalize the knowledge associated to ITSMSs in terms of ontologies that gather ITIL best practices; (ii) to model the semantics of the activities associated to ITSM processes in terms of workflows; (iii) to automatically generate the high-level requirements models of the information systems needed to support these processes; and (iv) from the latter, to obtain lower-level models (and eventually code) by means of automated model transformations. The proposed approach has been validated using a real case study from a Spanish IT service provider.