Magedanz, Felix (2011) Dynamic analysis of .NET applications for architecture-based model extraction and test generation. (Diploma thesis), Kiel University, Kiel, Germany, 132 pp.

Preview

Text Magedanz2011DynamicAnalysisOfNETApplicationsForArchitectureBasedModelExtractionAndTestGeneration.pdf - Published Version Download (6MB) | Preview

Abstract

Extending the Kieker framework to allow comprehensive dynamic analysis of applications developed with Microsoft's .NET framework is the fundamental approach taken by this thesis. Furthermore, the practicality and robustness of the presented extension---which we call Kieker.NET---had to be proved not only by controlled experiments in the lab, but also by an in-depth case study on a real-world software system. We achieved the first goal by developing a .NET-based Kieker framework that does not just simply co-exist with its Java-counterpart, but rather cooperates with it. We employed a .NET/Java bridging technology to make Kieker's core available to .NET, and ported only as few Java classes as possible. Our monitoring probes are based on Postsharp, a powerful AOP implementation for the .NET framework. With running examples throughout the development stages of our framework, as well as a comprehensive overhead analysis conducted with a micro-benchmark, we showed that Kieker.NET is capable of performing almost all tasks in the .NET programming environment that can be done with Kieker's original monitoring components for Javabased applications. An extensive case study on a complex real-world software system provided by a major German finance institute, including multiple extracted architecture- and trace-based models and a macro-benchmark, substantiated our findings from the lab studies.

Actions (login required)

View Item