Flexible Service Compositions

Our collaboration with Barbara began in the DISCORSO and MAIS projects, which pioneered an integrated framework for designing and managing flexible and adaptive service-based business processes. Anticipating the shift from static Web services to dynamic, service-oriented architectures well before the cloud era, the framework introduced mechanisms for runtime adaptation, QoS-aware service selection, and continuous self-validation. Its design environment combined an extended BPMN model with WebML for human activities. Automated activities were specified through abstract services annotated with functional and QoS requirements. The corresponding BPEL fragments were automatically generated, and the runtime components, including an extended UDDI registry, QoS matchmakers and a Service Selector, mapped abstract specifications to optimized concrete services using mixed-integer linear programming. This foundational work has since evolved into broader research on scalable optimization, dynamic software product lines, advanced monitoring and exception handling, situational computing, and multi-layered service architectures, culminating in new perspectives for AI-driven adaptation.