Abstract

Modern architectural principles, like the microservice approach, introduce new challenges to the design and evolution of distributed systems. The diversity and number of components in a system landscape, as well as composite interactions between the components, causes higher overall complexity of the system. This paper presents a new approach for identification of problem areas and design issues within distributed systems. The proposed framework combines traditional software metrics and introduces new architecture metrics. A reasonable combination of metrics is used to spot most critical components of a system and by this, predict post-release defects. The method is successfully evaluated on one industrial distributed system and showed its potential. However, further test data and evaluations are needed.

@inproceedings{...,
author = {...},
title = {...},
booktitle = {...},
year = {...},
isbn = {...},
pages = {...},
location = {...},
doi = {...},
publisher = {...},
address = {...},
}