Statistics and Machine Learning Techniques for Real End-User Experience

Real End-User Experience (RUE) is a monitoring approach that aims to measure the end-user experience by providing information on availability, response time, and reliability of the real used IT services. The response time of each user transaction is measured by an analysis of the network communication flows. Several performance metrics get archived to monitor RUE over time. An abstract, generalized view of performance over time is of advantage before digging into data. We explored how advanced statistics and machine learning techniques can be used as effective tools to bring registered data to the desired level of abstraction. The resulting high-level visualizations help to get the big picture of what is happening within a network under investigation and improve our understanding of application performance