Abstract
High-tech companies are struggling today with the maintenance of legacy software. Legacy software is vital to many organizations as it contains the important business logic. To facilitate maintenance of legacy software, a comprehensive understanding of the software’s behavior is essential. In terms of component-based software engineering, it is necessary to completely understand the behavior of components in relation to their interfaces, i.e., their interface protocols, and to preserve this behavior during the maintenance activities of the components. For this purpose, we present an approach to infer the interface protocols of software components from the behavioral models of those components, learned by a blackbox technique called active (automata) learning. To validate the learned results, we applied our approach to the software components developed with model-based engineering so that equivalence can be checked between the learned models and the reference models, ensuring the behavioral relations are preserved. Experimenting with components having reference models and performing equivalence checking builds confidence that applying active learning technique to reverse engineer legacy software components, for which no reference models are available, will also yield correct results. To apply our approach in practice, we present an automated framework for conducting active learning on a large set of components and deriving their interface protocols. Using the framework, we validated our methodology by applying active learning on 202 industrial software components, out of which, interface protocols could be successfully derived for 156 components within our given time bound of 1 h for each component.
Highlights
Large-scale software systems are inherently complex, with complexity caused by a large number of constituent components and the interactions between them [54]
We use the terms active learning result for models obtained by learning the implementation of software components and derived interface protocols for interface protocols inferred from active learning results
8.7.1 Number of components learned. As both the theoretical bounds for learning and testing algorithms are influenced more by the number of states than by the number of inputs [16,26], we grouped the components according to the number of states for the analysis of our results
Summary
Large-scale software systems are inherently complex, with complexity caused by a large number of constituent components and the interactions between them [54]. ASML, Veldhoven, The Netherlands of evolving requirements, emerging technology trends and hardware changes [38]. To deal with this ever-increasing complexity of high-tech software systems, different software development methodologies emerged with the goal of raising the abstraction level of software development. We choose Mealy machines to represent the results of active learning and derived interface protocols. This is because Mealy machines provide the notion of inputs and outputs; they are good representations for reactive systems. Definition 3 An interface protocol is a tuple IP = S, Σ, , I , →, s , where – S is a set of states, – Σ is a set of input actions, – is a set of output actions, – →⊆ S × Σ × × S is a transition relation and – s ∈ S is the initial state.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
