A New Semi-Automated Method for Service Identification

Abstract

Service identification plays a key role in the design of service-oriented systems. There are non-model-based and model-based methods for extracting services from business processes. These methods suggest a set of mostly descriptive solutions that do not pay sufficient attention to service design guidelines and the conceptual relations between tasks. The challenge is to develop an algorithm to automatically identify services from business processes to simplify the analysis and reduce the gap between information technology and business needs. In this paper, we develop a semi-automated service identification method that addresses this gap. This method incorporates the Goal, Data, and Business Process Models (BPM) to identify services based on related tasks, shared data, and business requirements. It advances previous methods by simultaneously considering both semantic and structural relations between tasks which permits better and more accurate identification of services. Moreover, the proposed method considers the principles of service design such as internal cohesion of service methods, loose coupling of services, and reusability of the identified services.