Method and Models to Enable Optimal Automated Service Composition

Abstract

Since the development of the Service Oriented Architecture concept, business analysts and system developers have looked forward to the day when they could reconfigure applications to adapt to new business by combining services in new ways to adapt to changing business needs. Technologies such as the Web Services Description Language and Business Process Modeling Notation (BPMN) provide key building blocks but are not sufficient to enable run-time reconfiguration of services. To enable this functionality, this research develops Druid, a framework for dynamically composing web services into executable processes based on a business process model defined using the BPMN modeling language. To support this framework, this research develops a service description modeling language, extensions to the BPMN language, and a formal model for composing services based on a business model. This research also develops a Quality of Service (QoS) model used for calculating the optimal service composition.