A Performance Prediction Model for Google App Engine

Cloud computing environments are becoming popular as platforms for enterprise information systems. However, in public PaaS environments, performance prediction is one of the obstacles to migrate into the cloud, since only a little performance information on the platforms is available. In addition, the structure of the platforms is not opened to general public. This paper proposes a modeling and simulation based framework to predict the cloud performance. As a modeling and simulation tool, we use the UPPAAL model checker, which expresses the models in the form of timed automata. The framework is build focusing on the application structure, which consists of a series of cloud APIs. The platforms are simply regarded as a mechanism to produce the probabilistic process delay. The paper uses Google App Engine (GAE) as a platform, however the approach can be applied to any other PaaS type cloud environments.