An Ontology-Based and Model-Driven Approach for Designing IT Service Management Systems

Currently, few projects applying a Model-Driven Engineering (MDE) approach start from high-level requirements models defined exclusively in terms of domain knowledge and business logic. Ontology Engineering (OE) aims to formalize and make explicit the knowledge related to a particular domain. In this vein, this paper presents a modeling approach, formalized in ontological terms, for defining high-level requirements models of software systems that provide support for the implementation of Information Technology Service Management Systems (ITSMSs). This approach allows for: (1) formalizing the knowledge associated to the ITSM processes contained in an ITSMS; (2) modeling the semantics of the activities associated to these processes in terms of workflows; (3) automatically generating the high-level requirements models of the workflow-based software systems needed to support (part of) the ITSM processes; and (4) from the latter, obtaining lower-level models (and eventually code) by means of automated model transformations. A real case study describing the use of this proposal to model an Incident Management System is also included to demonstrate the feasibility and the benefits of the proposed approach. DOI: 10.4018/jssmet.2011040104 66 International Journal of Service Science, Management, Engineering, and Technology, 2(2), 65-81, April-June 2011 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. have increased their expectations related to the IT department, and now they need IT to support their business processes in a strategic way. That is, organizations are aware of the closer relationship and convergence between business and IT. In this vein, IT Service Management (ITSM) provides a set of specialized organizational capabilities and a professional practice, supported by an extensive body of knowledge, experience and skills for providing value to customers in the form of IT services (OGC, 2007). The implementation of any IT serviceoriented software system requires performing a number of different steps in order to produce all the required artifacts (either internal or deliverable). Based on the notion that a software system is a representation of another system (i.e., the real-world), the first step is to formalize the domain concepts and the relationships between them (i.e., the ontology), in order to obtain a common vocabulary agreed by all the stakeholders involved in a given project for requirements elicitation. In addition, apart from the domain concepts, additional rules, constraints and semantics are required in order to avoid semantic ambiguities, uncertainties and contradictions. The Web Ontology Language (OWL) (Smith, Welty, & McGuinness, 2004), the de facto standard for ontology representation, enables the definition of rules, constraints and semantics in terms of logic based domain concepts. The importance of an investigation of the issues involved in the IT service-oriented requirements analysis is also remarked by Lichtenstein, Nguyen & Hunter (2004). However, in spite of the efforts of the Software Engineering (SE) community to define new intuitive and powerful techniques, there is still an open gap regarding the automated and seamless integration of domain aspects (i.e., the business view) into the software development process. The emerging Model-Driven Engineering (MDE) paradigm offers a promising solution to cope with this limitation. MDE addresses the inability of third-generation languages to cope with increasing software complexity, allowing us to describe domain concepts effectively (Schmidt, 2006; Gašević & Hatala, 2010). Model Driven Engineering (MDE) is a software and system construction approach based on high-level abstract modeling. All the relevant information in a project is stored in models based on well-defined languages and development is then carried out as a sequence of model transformations. The MDE term was first proposed by Kent (2002) but it is derived from the OMG’s Model Driven Architecture (MDA) initiative (OMG, 2003). As shown in Figure 1, MDA defines a particular MDE process aimed at separating the business logic from the technological platforms. Thus, organizations can use MDA to meet the integration challenges posed by new platforms, while preserving their investments in existing business logic. MDA proposes three modeling levels, namely (ordered from highest to lowest levels of abstraction): Computation Independent Model (CIM), Platform Independent Model (PIM), and Platform Specific Model (PSM). Different Model-to-Model (M2M) transformations among these abstraction levels can be defined either top-down or bottom-up. Commonly, each CIM (model gathering high-level business requirements, sometimes called a domain model) is transformed into one or more PIMs (platform-independent architectural models). Similarly, each PIM is transformed into one or more PSMs (one for each target platform). PSMs are commonly very low level models, enabling the definition of a direct Model-to-Text (M2T) transformation for automatically generating the final system implementation, including code, documentation, etc. At the model layer (M1), CIMs are commonly high-level business models that represent the high-level requirements for the system to build (M0). A high-level requirement is focused on the actual stakeholders problems and needs and describes the characteristics of the domain of the systems (that is, what is needed, but not how this is to be implemented) (Olivé, 2007). Therefore, CIMs help in bridging the gap between the conceptual level mainly performed by domain experts and the implementation level performed by the designers and developInternational Journal of Service Science, Management, Engineering, and Technology, 2(2), 65-81, April-June 2011 67 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. ers of the artifacts that together satisfy the domain requirements (OMG, 2003). However, most actual MDA projects are focused on PIMto-PSM transformations or PSM to code transformations, and they rarely get a comprehensive understanding of the problem domain. In this paper, we present a formal ontology-based and model-driven approach for the semantic enrichment of workflow models in the context of Information Technology Service Management Systems (ITSMSs). This approach aims to enable the creation of enriched machineprocessable workflow models, which can be later processed and reused by other workflow modeling tools. In addition, our proposal aims at supporting a set of (automated) model transformations from the ontology-based workflow models to CIM models describing the high-level requirements of the software systems supporting the activities defined as part of the workflows in an ITSM process. An ITSM process is defined as “a structured set of activities designed to accomplish a specific objective” (WfMC, 1999). For this purpose, we use the workflow ontology defined by Valiente, García-Barriocanal, and Sicilia (2011) which allows us to formally describe ITSM processes in terms of workflow models. This ontology uses OWL, combined with the Semantic Web Rule Language (SWRL) (Horrocks, Patel-Schneider, Boley, Tabet, Grosof, & Dean, 2004), the latter used to specify additional constraints and infer new knowledge. Then, we transform the resulting OWL workflows into Business Process Model and Notation (BPMN) models (OMG, 2010a), which can be further processed and manipulated using the BPMN modeler included as part of the Eclipse SOA Tools Platform (STP) project (Eclipse, 2006). To do so, an XML Metadata Interchange (XMI) file (OMG, 2007) extracted from an OWL workflow is transformed into a XMI file that represents a BPMN model using an Extensible Stylesheet Language Transformations (XSLT) script (Clark, 1999). The main objective of the research presented in this paper is to translate ITSM process models (expressed using natural language or informal graphical representations) into formal and comprehensive representations of the Information Technology Infrastructure Library (ITIL) (http://www.itil-officialsite.com/) processes described in terms of the workflow ontology, so that they can be used as the basis Figure 1. MDA and the OMG’s four-layers metamodeling pyramid as depicted in (VicenteChicote & Alonso, 2007) 68 International Journal of Service Science, Management, Engineering, and Technology, 2(2), 65-81, April-June 2011 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. for modeling and designing the software systems that underpin an ITSMS. As a proof of concept, we started a pilot project with a Spanish IT service provider (the Information and Communication Technology Department – ICTD – of a Spanish company) interested in improving the quality of the services they were delivering to their customers in order to obtain an optimal level of customer satisfaction and to become more competitive and efficient. The rest of the paper is organized as follows. Section 2 covers the background. Then, Section 3 presents the proposed approach for implementing workflow-based ITSM processes contained in an ITSMS, exemplifying it with a case study regarding an Incident Management process based on the ITIL framework. Section 4 reviews related work. Finally, Section 5 draws some conclusions and outlines some future work.