CAREER: Architecture-Based Support for Software Application Families

提案号：ccr -9985441标题：职业：基于架构的软件应用程序家族支持主要研究者：Nenad Medvidovic， USCABSTRACT----------------软件架构研究通过关注粗粒度软件组件的组成，已经产生了许多开发复杂应用程序的技术。然而，现有的基于体系结构的方法通常将系统建模为“独一无二的”；此外，它们未能确保将体系结构属性一致地转移到系统的实现中。实现软件架构的全部潜力需要一种更现实的方法，在这种方法中，系统作为应用程序家族的成员进行建模和实现。这个项目关注于支持软件应用程序家族的基于体系结构的开发的技术的设计、实现和经验评估。研究了三个一般性问题：(1)应用程序族体系结构建模的符号；(2)将家族架构细化为其实现的机制；(3)软件连接器的性质——促进组件交互的服务——以及它们在一个家族中跨应用程序控制非功能属性（例如，性能、安全性）方面的作用。这项研究将以系统为中心的观点来增强当前以模型为中心的体系结构关注，并为异构性和重用提供原则性支持，这是开发应用程序家族的两个基石。通过让学生了解基于体系结构和基于应用程序家族的开发对基本软件工程原理的影响，该研究还将作为软件工程课程现代化的工具

Proposal Number: CCR-9985441Title: CAREER: Architecture-Based Support for Software Application FamiliesPrincipal Investigator: Nenad Medvidovic, USCABSTRACT----------------Software architecture research has yielded numerous techniques fordeveloping complex applications by focusing on the composition ofcoarse-grain software components. However, existing architecture-basedapproaches have typically modeled systems as "one of a kind;" additionally,they have failed to ensure consistent transfer of architectural propertiesto a system's implementation. Realizing the full potential of softwarearchitectures requires a more realistic approach, in which systems aremodeled and implemented as members of an application family. This projectfocuses on the design, implementation, and empirical evaluation oftechniques for supporting architecture-based development of softwareapplication families. Three general issues are studied: (1) notations formodeling application family architectures; (2) mechanisms for refiningfamily architectures into their implementations; and (3) nature of softwareconnectors--services facilitating component interactions--and their role incontrolling non-functional properties (e.g., performance, security) acrossapplications in a family. This research will augment the currentmodel-centric focus on architectures with a system-centric perspective andprovide principled support for heterogeneity and reuse, two cornerstones ofdeveloping application families. The research will also serve as a vehiclefor modernizing the software engineering curriculum, by exposing students tothe impact of architecture-based and application family-based development onfundamental software engineering principles