A feature-driven variability-enabled approach to adaptive service compositions.

Service compositions are widely used to construct complex applications. Due to the frequent changes of environment and requirements, service compositions need to be adaptable enough. In this work, we propose a feature-driven variability-enabled adaptive service composition approach to systematically treat the variability in the full life-cycle of service compositions. Specifically, the feature model is introduced to represent common and variable requirements and drive the variability design of service compositions. An abstract service composition model is used to define the variable business process. Rules and algorithms are then defined to transform the feature model to the abstract service composition model, from which different process instances are derived on demand to meet different requirements. We have developed a prototype tool to facilitate and automate our approach as much as possible. Finally, a case study is conducted to demonstrate the proposed approach and validate its effectiveness and efficiency.