Dynamic web services composition

• Authors: Liangzhao Zeng
• Venue: Dynamic web services composition
• Year: 2003

With the proliferation of the Internet and the wide acceptance of e-commerce, an increasing number of distributed and heterogeneous Web services are being offered. These Web services are commonly composed and coordinated to execute business processes, using process-based approaches. Current process modelling provides adequate support for static business processes. However, it has severe limitations on composing Web services for e-business. There are two major requirements for Web service composition. First, frequently changing business conditions, imply that business processes need to be composed and refined on the fly. Second, the volatile and dynamic nature of Web environments also leads to the dynamic availability of Web service offerings. In this work, we propose a dynamic Web service composition framework. We mainly address three issues: (1) Dynamic creation of process schemas via runtime business rules inference. Instead of statically defining process schemas to accommodate an explosive number of possibilities, we advocate a rule-directed approach to dynamically generate and execute composite services. As a result, end users can focus on the business goals to be achieved and the business policies that should be followed, without worrying about detailed description of control and data flow constraints. (2) Quality-driven and dynamic service selection. Instead of fixed association of tasks to Web services at build time, we have devised a quality-driven service selection mechanism that makes use of service quality information, inter-task constraints in business processes as well as preferences set by the requester. In this approach, quality constraints and preferences are assigned to composite services rather than to individual tasks within a composite service. Service selection is then formulated as an optimization problem and a linear programming method is used to compute optimal service execution plans for composite services. (3) Adaptive Service Composition. We propose an adaptive service composition approach to handle both component and unexpected exceptions. In our approach, composite services continuously monitor the behavior of their components and adapt themselves to appropriately react to run-time exceptions. At the same time, composite services also continuously check the consistencies between their process schemas and the business processes real world, and modify process schemas if changes occur in business processes.