Modellus: Automated Modeling of Complex Data Center Applications

The rising complexity of distributed server applications in enterprise data centers has made the tasks of modeling and analyzing their behavior increasingly difficult. This paper presents Modellus, a novel system for automated modeling of complex data center applications using statistical methods from data mining and machine learning. Modellus can automatically derive models to predict the resource usage of an application and the workload it triggers; these models can be composed to capture multiple dependencies between interacting applications.