Dynamic Analysis for the Reconstruction of System Behavior Models

The analysis of large software systems is often a difficult and time consuming task because of the sheer complexity of such systems, and the limited access software engineers have to the system´s source code. In this respect, it is important to be able to analyze the system based on the events that can be collected while the system performs its specified operations. Such analysis, can take the form of reconstructing system behavioral models, and identifying common operational patterns that may assist software engineers deduce important information on the properties and characteristics of the system being analyzed. In this paper, we present an analysis technique that allows for the hierarchical reconstruction of sequence diagrams and the identification of common event patterns from system traces. The technique has been applied for the analysis of different operations in the Session Initiation Protocol.