Samenvatting

In the past 10 years an important change has occurred in how to develop, integrate and reuse information systems. A new paradigm called Service Oriented Architecture (SOA) has emerged that is based on the development, deployment and reuse of (web) services which can easily be assembled in different ways allowing organizations to quickly adapt to changing business needs (Cox & Kreger, 2005).
However while SOA has a large potential for business one of the most complex issues in any SOA project is to define the right granularity of the services. The quest for the right granularity is complex if the promises of flexibility and reusability must be obtained. Decisions about the level of service granularity are typically made using various heuristics (Feuerlicht & Wijayaweera, 2007) and service designers usually make a best guess as to how services will be used (Stevens, 2002). A method for service definition is important in environments where a lot of services are available and the set of available services changes over time.
The question how services should share organizational principles or how services should be modeled to obtain flexibility and be able to adopt organizational changes is not answered yet. The quest for service granularity has been addressed in many articles (De Jong & Dietz, 2010), (Feuerlicht & Wijayaweera, 2007), (Rosen, 2007), (Papazoglou & Van den Heuvel, 2006), (Foody, 2005), (Sims, 2005) but none of these sources answers the question how to define service granularity properly; neither do they provide some kind of concrete guidelines. Most sources just mention the importance of finding a right service granularity.
Finding an optimum in service granularity is the main theme of this research. This optimum must be independent of the context and applicable for all organizations. Researching the main theme is directed by the concepts and several determinants that are valid for service granularity and the utilization of the Design & Engineering Methodology for Organizations (DEMO). This objective leads to the central research question:
How to define the right granularity for services such that the determinants of granularity are balanced with the proposed business benefits of flexibility and reusability using DEMO's informational construction modeling from a service-oriented perspective?
In this research an extensive review of literature sources on service granularity and DEMO has been performed. The topic has been explored from different angles to get a complete picture of the service granularity environment.
To test how to define the right granularity of services two explorative case studies and one improving case study at organizations using the positivist case study research method have been conducted. The explorative case studies resulted in a conceptual model that reflected the process and activities that have been performed. The objective of the improving case study was to test and improve the conceptual model from the explorative case studies. The applied research approach followed the action research cycle of (Baskerville & Wood-Harper, 1996). In the action taking phase the collaborative modeling approach is applied. In total three iterations of the action research cycle were conducted in which the conceptual
Finding an optimum in service granularity
Page 8 of 205
framework is extended by including the results of each step. This is also known as the spiral towards understanding (Carroll & Swatman, 2000). The findings of the improved case study have been validated by an expert panel.
The process of the improving case study was aimed at exploring the concept of the granularity of services and by collaborative modeling we tried to find an optimum. The improving case study showed that the right granularity of services is not achieved out of the box. We have seen that optimizing service granularity is a process of several iterations and that reaching an optimum is something subjective. The participants of the improving case study agreed that the end result is an optimum model of information services, because of the achieved consensus about the results that could be substantiated whenever necessary and is directive to further design software.
After the improving case study the results were validated by an expert panel. The expert panel concluded that DEMO is applicable in modeling the essence of the informational organization. The DEMO infological construction model supports the modeling of services in a fashion that services could be optimized to their environment. By extending the information services of DEMO with the activities that are stated in the conceptual model, the participants concluded that a coherent and relevant set of activities is provided that contributes to optimizing service granularity.