A graph-based approach to capture software behavior in architecture
T.G.C. Ipskamp · Utrecht University Repository (Utrecht University) · 2018
Very Large Software Systems (VLSSs) grow over time and cause the actual Software Architecture to deviate from the intended Software Architecture. Often over time, it becomes less clear what the actual Software Architecture looks like. To alleviate this problem, many Software Architecture Reconstruction (SAR) techniques exist. However due to changing programming techniques, software systems have become more dynamic in nature. This dynamic nature introduces new problems when using SAR techniques, since not all dependencies can be found by analyzing static artifacts and documents. To solve this problem, this thesis proposes a new method to record and analyze Software Behavior at runtime to identify these dynamic dependencies. This method combines both existing static techniques and new runtime techniques to gather comprehensive data and aims to visualize this data in new Software Behavior visualizations. These visualizations are aimed at supporting Software Architects in reasoning about their software system.