Supporting Change Impact Analysis for Service Oriented Business Applications

Business applications encode various business processes within an organization. Business process specification languages such as BPEL (Business Process Execution Language) are commonly used to integrate various services in order to automate business processes within an organization. To remain competitive edge, managers frequently modify their processes. Determining the cost of modifying a business process is not trivial since the changes to the business process have to account for source code changes in various services. In this paper, we propose an approach to estimating the cost of a business process change in a service oriented business application. The approach applies change impact analysis techniques to business process specifications, and source code. The approach generates an initial change impact set from business process components. These components are then mapped to the corresponding source code entities. These code entities act as seeds for traditional source code impact analysis. Using code dependencies, such as call and inheritance relations, we derive a metric to capture the complexity of particular business process changes. Managers can then use this metric to gauge the cost and resources needed to implement changes in their business processes without having to study the code. We demonstrated the feasibility of our approach using an experiment on an open source service oriented business application.