An Architecture Smell Knowledge Base for Managing Architecture Technical Debt

Abstract

Many software projects suffer from architecture erosion and archi-tecture technical debt. One challenge is to identify affected parts and prioritize them for refactoring. Architecture smells are indica-tors of potential architecture technical debt, but architecture smells are ambiguous and their impact is not always clear. To address this, we have built a knowledge base that improves understanding of architecture smells and identifies violated software design prin-ciples and affected quality attributes. The design principles help our understanding of what causes architecture smells, while the impaired quality attributes represent the consequences. We con-ducted a systematic literature review to identify these relations and built an architecture smell ontology. This ontology provides a knowledge base that architects can use to prioritize the smells according to the project's individual quality goals.