Run-time Optimization of Composite Web Services with Response Time Commitments

In this chapter we address dynamic decision mechanisms for composite web services. We represent the composite web-service as a (sequential) workflow of tasks. For each task within this workflow, a number of third-party service alternatives may be available. We assume that the third-party service (task) alternatives offer the same functionality at different price-quality levels. Before a task in theworkflowwill be executed, a service alternative has to be selected that implements the task functionality. Decisions are represented in a decision strategy which defines decisions for all tasks in the workflow. Our goal is to find a dynamic strategy that maximizes the expected benefit for the composite service providers. For each task, the serviceselection decision may be based on information about: observed response times in the current workflow, sub-service costs, response time characteristics of the alternatives, and end-to-end response time objectives with the corresponding rewards and violation penalties. We propose an approach, based on dynamic programming, to determine the optimal, dynamic selection policy. Numerical examples show significant potential gain in expected benefits using the dynamic approach compared to other, non-dynamic approaches.