Architecture-based software testing: Systematic literature review

Context:: Structural insights derived from software architecture can be leveraged to design targeted test cases that align with component interactions, dependencies, and architectural constraints. Architecture-based software testing (ABT) explicitly and intentionally utilizes software architecture to improve the effectiveness and efficiency of testing. Objectives:: The current scope and maturity of ABT remain unclear, particularly regarding how architecture styles/patterns, tactics, and views are utilized in testing. This systematic literature review assesses the current state of ABT and identifies directions for future research. Methods:: This paper defines ABT and discusses the testing implications of architecture design decisions, including architecture styles/patterns, architecture tactics, and architecture views. Based on these testing implications, we formulate research questions and conduct a systematic review involving study selection, quality assessment, data extraction, and qualitative synthesis. Results:: Based on the defined study selection criteria and quality assessment criteria, 41 primary studies were selected and analyzed. The results show that architecture views are most frequently used in ABT, mainly to construct intermediate test models, derive test cases, support test selection, and guide testing decisions. Architecture styles/patterns are used more selectively to identify style-specific testing challenges or to guide individual testing techniques, whereas architecture tactics are rarely explicitly used in test design. Conclusion: This study identifies three key directions for advancing ABT: (1) formulating reusable and architecture-consistent test strategies that incorporate the structural constraints and behavioral semantics defined by architecture styles/patterns, and tactics; (2) establishing architecture-driven coverage criteria together with traceable and operationalizable derivation procedures that enable rigorous adequacy and completeness assessment; and (3) developing test methods that systematically and purposefully leverage multiple, complementary architecture views in accordance with test objectives and system characteristics.