This chapter introduces QSE, the Qualitative Service Elicitation method. It applies qualitative research procedures in service elicitation. Service engineering practice lacks lightweight methods to identify service candidates in projects with tight schedules. QSE provides a systematic method to analyze requirement material in service-oriented systems development with a feasible effort. QSE uses the procedures of the grounded theory research method to elicit service candidates from business process descriptions and business use case descriptions. The chapter describes the method with examples and a case study.
TopThe Challenge Of Service Elicitation
The service oriented approach differs fundamentally from the conventional development paradigms in the key concept of dynamically accessible services. The scope and performance of services are under constant development to support an increasing number of consumers. Components and objects do not provide this kind of run-time flexibility. Likewise, traditional requirement engineering practices do not support service composition nor do they encourage the identification of reusable services. (Papazoglou, Traverso, Dustdar, Leymann, & Kramer, 2006; Van Nuffel, 2007; Zimmermann, Schlimm, Waller, & Pestel, 2005)
Papazoglou et al. (2006) have listed the main challenges of the service-oriented engineering domain in their research roadmap. Novel approaches are required in service engineering to address the current challenges and to provide sound methods that allow enterprises to design and deploy services more efficiently while adapting to the changes matching the rate and pace of the business.
The QSE approach addresses some of the challenges identified by Papazoglou et al. (2006). For example, QSE supports the refinement of service compositions and links the compositions to service candidates identified in the projects. Similarly, QSE provides practical means to build an enterprise level service catalogue, which can be used in gap analysis. Additionally, the catalogue provides a ground for refining the right granularity of the services. The method itself does not provide automation in the analysis, but provides systematic procedures for the analysis, thus helping to reduce human errors. To enable systematic analysis, we have taken ingredients from research methodology. We believe service elicitation by nature much resembles qualitative research.
The identification of services has been studied for some time and various methods already exist, but they focus on specific areas and the elicitation of specific types of services. A survey by Ramollari et al. (2007) lists ten different methods with varying coverage of the SOA project life cycle. Arsanjani (2005) classifies the SOA approaches into six categories: business process driven, tool-based MDA, wrap legacy, componentized legacy, data driven and message driven approaches.