Research on Cost-Driven Services Composition in an Uncertain Environment

Abstract

In recent years, increasing numbers of researchers have concentrated on service workflow to support cross-domain software development. However, the uncertain characteristics of the Internet impose high risks on service workflow reliability. The risk of failure caused by unavailable services may increase costs when using service workflow-based applications. Thus, it is necessary to consider the non-functional factors, such as service cost and reliability. In this paper, we propose a cost-driven services composition approach for enterprise workflows that employs formal verification to recommend appropriate services for abstract workflows. The services composition is measured quantitatively to ensure that the configuration to service the workflow solution has the best performance, high reliability and low cost. First, this solution introduces a service search approach based on an inverted index, and the service recommendation method is based on an improved Pearson formula. Next, the solution returns a minimum set of candidate services for constructing a workflow instance. Second, the service and workflow models are defined to formalize the behaviour of service composition; this is considered to be a verification model. Third, transformation rules are provided to change BPEL4WS into a verification model, and PCTL (Probabilistic Computation Tree Logic) formulae are used to specify the reliability and cost-related properties. The quantitative verification method checks each possible plan for service composition using probabilistic model checking. Finally, the results of a series of experiments show that our approach is effective in generating an optimal service workflow.