A Pattern-Oriented Impact Analysis Approach for Mashups

Mashups have gained popularity among the end users for its convenient building and sharing fashions. Inevitably, a mashup is vulnerable to break down caused by its integrated components' change, such as upgrade and inaccessibility. This paper focuses on the change impact analysis for mashups. We distill the data dependencies of mashups from pattern-oriented perspective, and abstract the mashup creation to a dependency graph. With the graph and the influence probability we defined, the impact of the components on the mashup can be calculated quantitatively. The approach is implemented in our mashup environment -Mashroom. The case study is also discussed.