Data-centric Workflow Engineering Research

The importance of effective Business Process Management (BPM) increases as the needs for better insight, understanding and efficiency for business operations increases. In principle, workflow management systems (WfMS) offer promising perspectives to cope with the challenge of BPM. Workflows, as a typical process modeling approach, often emphasize the sequencing of activities (i.e., control flows), but ignore the informational perspective or treat it only within the context of single activities. This incomplete view of the process adds substantial conceptual complexity to models of business operations and processes, making them hard to understand. Monitoring processes and retrieving history data are also complex in current WfMS.During recent years, an artifact-centric approach of coupling control and data emerged in the practice of business process design. Artifacts are business-relevant objects that are created, evolved, and (typically) archived as they pass through business process. They combine both data aspects and process aspects into a holistic unit, and serve as the basic building blocks from which models of business operations and processes are constructed. The approach enables a natural modularity and componentization of business operations and varying levels of abstraction. However, the lack of guideline in the definition, test, deployment of the artifact-centric model and the missing of a well-designed, effective workflow engine to support the execution of such model slow down the business companies’shifting from classic workflow models to the artifact-centric model.This paper is an initial attempt in the engineering area for the artifact-centric model. This paper presents a guideline for defining, developing, testing and deploying data-centric process definition. In order to express this idea better, this paper then proposes a new artifact-centric process model called ArtiNet which can fully and easily support the guidelines. On every aspect, this paper digs into ArtiNet to show its advantages. Then a new process engine called ArtiMT is developed for executing the new model. ArtiMT has a micro kernel for scheduling and several loose coupled components such as data controller, task performer. This makes the engine more robust.Finally, an experiment is designed to test the validation and practicality of the model and the engine. The result of the experiment and the feedback from real business worker show that ArtiNet and ArtiMT are quite promising. They have the potential for providing business world a more easy-to-use, robust and flexible process management environment.