Software Clustering Using Dynamic Analysis and Static Dependencies

Decomposing a software system into smaller, more manageable clusters is a common approach to support the comprehension of large systems. In recent years, researchers have focused on clustering techniques to perform such architectural decomposition, with the most predominant clustering techniques relying on the static analysis of source code. We argue that these structural relationships are not sufficient for software clustering due to the increased complexity and behavioural aspects found in software systems. In this paper, we present a novel software clustering approach that combines dynamic and static analysis to identify component clusters. We introduce a two-phase clustering technique using software features to build a core skeleton decomposition and structural information to enrich these clusters. A case study is presented to evaluate the applicability and effectiveness of our approach.

• Type: Conference
• Year: 2009
• Where: CSMR
• Authors: Chiragkumar Patel, Abdelwahab Hamou-Lhadj, Juergen Rilling
• Keywords: Clustering Technique, CSMR 2009, Predominant Clustering Techniques, Software Clustering Approach