A Deployment-Oriented Performance Modeling and Evaluation Method for SaaS Application Based on LQN

The deployment scheme of SaaS application has an important influence on the performance of SaaS application. Studying the impact of deployment scheme on the performance of SaaS application could contribute to performance improvement of application. However, there are few methods that can directly and efficiently model and analyze the relationship between deployment scheme and the performance of SaaS application. In this paper, a deployment-oriented method for modeling and evaluating the performance of SaaS application based on LQN is proposed. The performance modeling method can explicitly model both the SaaS application and the VM resources in multiple deployment schemes, and the performance metrics can be solved based on the model. The deployment-oriented performance evaluation method can evaluate the performance of services in SaaS application based on response time, which is used to further analyze the impact of the service-VM relationship and service-service relationship on the performance of services, and the analysis conclusions can provide useful help for adjusting the deployment scheme and improving the performance of SaaS application. Finally, some experiments are carried out to show the effectiveness of the proposed method and the analysis conclusions.