Extracting Realistic User Behavior Models

Workloads play a central role in assessing software qualities, like performance and privacy. They are characterized by intensity and user behavior patterns. Combining multiple intensities and behaviors are used to create workload profiles which, among others, sevaluate software design, predict of system utilization. The central challenge for workload profiles is their fit to real workloads and in particular the match to specific behaviors. This is especially relevant for understanding and identifying specific user groups and support workload composition by operators. In this paper, we address the identification of such realistic user behaviors by utilizing domain specific attributes, report on our evaluation of the fitness of behavior clustering approaches, and discuss our setup to evaluate further clustering approaches.