Performance Prediction and Analysis of Quality of Services for Cross-Organizational Workflows

Service-Oriented Architecture (SOA) promotes the combination of workflow and service composition technology, and it provides important technical supports for cross-organizational workflow applications. This paper proposes an analysis and prediction model based on time series using Particle Swarm Optimization based Back Propagation Neural Network (PSO-BPNN) model, to predict the dynamic performance of workflow systems. When the predicted value out of the preset range, we analyze the issues according to data of Quality of Service (QoS) detected at runtime, to find why cause service performance failure, which suggests more suitable recovery strategies for service composition. The results of simulation experiment have validated the effectiveness of the proposed approach.