Software Performance Prediction Approach Based on Stochastic Process Algebra

Performance is one of the inherent attributes of software system.It is mainly decided by the selection of software architecture and designing information.Whether the performance demands are meet can be evaluated by model-based performance prediction technologies.So do the capability projecting and the fine-tuning.The interaction process of J2EE was modeled using Performance Evaluation Process Algebra(PEPA).Remote invocation and container processing details are represented in the model,and major performance factors of distributed J2EE application are reflected.A COTS-oriented performance prediction approach based on PEPA models was proposed.It can help to improve responsibility and scalability besides predicting system performance.The approach was verified by a J2EE application case study.