CSR-SMA: Collaborative Proposal: A Model-Driven Performance Analysis Framework for Distributed, Performance-Sensitive Software Systems

本项目通过开展以下协同研究活动，重点研究设计时验证DPSS软件架构和流程的创新方法：(1)性能分析的概念框架，该框架将应用基于随机奖励网（SRNs）的分析/数值技术和仿真技术进行DPSS系统架构的设计时性能分析。该活动将开发一个理论框架，为中间件的各个构建块及其组合构建可配置和可组合的性能分析模型。(2)用于全系统性能分析的模型驱动生成技术，这将消除易出错和繁琐的DPSS系统架构和流程的临时开发。(3)面向方面的可变性解决方案，它将开发新的面向方面的建模工具，帮助模块化由可配置中间件构建块引起的可变性。然后，这些工具将把这些构件编织到现有的性能模型中，以专门化块及其组成。

This project focuses on innovative methods for design-time validation of DPSS software architectures and processes by conducting research on the following synergistic research activities: (1) Conceptual framework for performance analysis, which will apply Stochastic Reward Nets (SRNs)- based analytical/numerical techniques and simulation techniques for design-time performance analysis of DPSS system architectures. This activity will develop a theoretical framework to build configurable and composable performance analysis models of individual building blocks of middleware and their composition. (2) Model-driven generative technologies for whole-system performance analysis, which will eliminate error-prone and tedious ad hoc development of DPSS systems architectures and processes. (3) Aspect-oriented variability resolution, which will develop novel aspect-oriented modeling tools that help modularize the variability incurred by the configurable middleware building blocks. These tools will then weave in these artifacts into the existing performance models to specialize the blocks and their composition.