A Pattern Language for the Evolution of Component-based Software Architectures

Abstract

Architecture-centric software evolution enables change in a system’s structure and behaviour while maintaining a global view of the software to address evolution-centric trade-offs. The existing solutions for architectural maintenance and evolution fall short of exploiting generic and reusable expertise to address recurring evolution problems. We present a pattern language as a collection of interconnected change patterns that enable reuse-driven and consistent evolution of component-based software architectures. Pattern interconnections represent possible relationships among patterns (such as variants or related patterns) in the language. In general, we introduce architecture change mining (pattern language development) as a complementary and integrated phase to facilitate reuse-driven architecture change execution (pattern language application). We evaluate the language applicability to support pattern-driven reuse in architecture evolution of a payment system case study. We also analyse the precision and recall factor as a measure of selecting the most appropriate pattern(s) from the language collection. The pattern language itself continuously evolves with an incremental discovery of new patterns from change logs over time