Approaching the Model-Driven Generation of Feedback to Remove Software Performance Flaws
Approaching the Model-Driven Generation of Feedback to Remove Software Performance Flaws
复制标题
接近模型驱动的反馈生成以消除软件性能缺陷
DOI：
--
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Catia Trubiani
中科院分区：
文献类型：
--
作者：
V. Cortellessa;A. Marco;Romina Eramo;A. Pierantonio;Catia Trubiani
解释性能分析的结果和提供反馈的软件模型，以克服性能缺陷的问题可能是在软件性能工程领域最关键的开放问题。这一步中的自动化将有助于将性能验证作为软件生命周期中的集成活动引入，而不会显著影响软件开发人员的日常实践。在本文中，我们接近的问题与模型驱动的技术，我们建立一个通用的解决方案。基于性能反模式的概念，即软件建模中导致性能缺陷的不良做法，我们引入了元模型和转换，可以支持缺陷检测和解决的整个过程。我们提出的方法是符号独立的，可以嵌入到任何（现有的或未来的）具体的建模符号，通过使用编织模型和自动生成的模型转换。最后，我们讨论了从这项工作中打开的问题和未来的成就，在这一领域的手感谢模型驱动技术。
The problem of interpreting results of performance analysis and providing feedback on software models to overcome performance flaws is probably the most critical open issue in the field of software performance engineering. Automation in this step would help to introduce performance validation as an integrated activity in the software lifecycle, without dramatically affecting the daily practices of software developers. In this paper we approach the problem with model-driven techniques, on which we build a general solution. Basing on the concept of performance antipatterns, that are bad practices in software modeling leading to performance flaws, we introduce metamodels and transformations that can support the whole process of flaw detection and solution. The approach that we propose is notation-independent and can be embedded in any (existing or future) concrete modeling notation by using weaving models and automatically generated model transformations. Finally, we discuss the issues opened from this work and the future achievements that are at the hand in this domain thanks to model-driven techniques.