Framework for monitoring and testing web application scalability on the cloud

Abstract

By allowing resources to be acquired on-demand and in variable amounts, cloud computing provides an appealing environment for deploying pilot projects and for performance testing of Web applications and services. However, setting up cloud environments for performance testing still requires a significant amount of manual effort. To aid performance engineers in this task, we developed a framework that integrates several common benchmarking and monitoring tools. The framework helps performance engineers to test applications under various configurations and loads. Furthermore, the framework supports dynamic server allocation based on incoming load using a response-time-aware heuristics. We validated the framework by deploying and stress-testing the MediaWiki application. An experimental evaluation was conducted aimed at comparing the response-time-aware heuristics against Amazon Auto-Scale.