System Runs Analysis with Process Mining

Abstract

Information systems (IS) produce numerous traces and logs at runtime. In the context of SOA-based (service-oriented architecture) IS, these logs contain details about sequences of process and service calls. Modern application monitoring and error tracking tools provide only rather straightforward log search and filtering functionality. However, "clever" analysis of the logs is highly useful, since it can provide valuable insights into the system architecture, interaction of business domains and services. Here we took runs event logs (trace data) of a big booking system and discovered architectural guidelines violations and common anti-patterns. We applied mature process mining techniques for discovery and analysis of these logs. The aims of process mining are to discover, analyze, and improve processes on the basis of IS behavior recorded as event logs. In several specific examples, we show successful applications of process mining to system runtime analysis and motivate further research in this area.The article is published in the authors' wording.

Using Process Mining to Leverage the Development of a Family of Mobile Applications

“…Leemans et al even constructed a ProM plug-in -Statechart Workbench [24] that can support users of this approach. In 2015, Shershakov and Rubin published a paper [25] on how to analyse real-life software executions and what can be discovered from event logs of such executions. Later, the team of collaborators of Sergey Shershakov developed several techniques to synthesise UML activity diagrams and other types of visual model for service-oriented and component-based systems [26], [27].…”