Dynamic structure measurement for distributed software

With the advent of network technologies and the ultra-fast increasing of computing ability, the distributed architecture has become a necessity for the majority of software systems. However, it is difficult for current architecture measurements to evaluate distributed systems, such as cohesion and coupling. Most current methods focus on the relations among various classes or packages but barely consider the structure at component level, which has a serious impact on change impact analysis, fault diagnosis, or other maintenance activities. In this paper, we propose a dynamic structure measurement for distributed software. The intra-component and inter-component dependencies are introduced into a Calling Network model to further represent distributed software. More importantly, based on the Kieker monitoring framework, the measurement methods are proposed and implemented for distributed software. Two structural quality attributes cohesion factor of component (CHC) and coupling factor of component (CPC) are measured. Finally, case studies are conducted on two open-source distributed systems: RSS Reader Recipes and the distributed version of iBATIS JPetStore. By applying the proposed methods and comparing with the existing ones, the features of CHC and CPC can be assessed and observed for distributed software.