Recombinant and evolutionary high-level software architectures: Design, integration, evolution and analysis

点击翻译按钮获取中文摘要

One of the fundamental goals of software engineering is to enable predictable and modular construction and evolution of software systems by assembling software architecture components. However, due to the increasing complexity, distributed, and dynamic nature of current information systems, undetected integration failures from internal violations may be ultimately revealed only as costly accidents that often cause undesirable system crashes. In addition, interface and configuration issues frequently arise from misunderstanding, incompatibilities, or differences resulting from independent development practices, heterogeneous environments, or different vendors. As a result, since maintaining and changing this class of (Web-based) distributed and dynamic software systems usually requires the examination and modification of components and component interactions, system integration becomes difficult to maintain as system components are extended or upgraded. The main goal of this project is to define and evaluate a new recombinant and evolutionary software architecture approach, and supporting methods and tools, for component-based design, integration, evolution, and analysis. The approach is based on formal abstract models of distributed event-based recombinant components, extended component interfaces, concern and process-based evolutionary extensions, and integration and evolution contracts that can provide a more expressive definition of component-based system design representations and an enhanced analysis of the interaction of architectural components, their extended functionality and behavior, and their compositions. In a general way, the results of this research can impact individuals or (Canadian) organizations that are struggling to produce large, complex (dynamic) systems and applications, and can help push the drive to improve productivity, software quality, reduce maintenance overheads, and impact on the establishment of a marketplace of contract-based software components. In addition, we are convinced that providing a non-ad-hoc and comprehensive component integration and evolution theory and its automated support will radically improve software engineering efforts.