Flows and views for scalable scientific process integration

Workflow technology has recently been employed in scientific applications because of their ever-increasing complexities across multiple organizations, institutes, research labs, or units over the Internet and Intranet. In this paper, we propose a methodology for the decomposition of complex scientific process requirements into different types of elementary flows such as control, data, exception, semantics, and security. Based on that, we can determine the subset of each type of flows (i.e., flow views) necessary and the related requirements for the interactions with each type of collaboration partners in the process integration. These subsets collectively constitute a process view, based on which interactions can be systematically designed, integrated and managed in a scalable way. We show with a case study in a scientific research environment to demonstrate our approach. We further illustrate how these flows can be implemented with various contemporary Web services technologies.