CSR --- SMA: Predictive Testing of System Software

CSR --- SMA：系统软件的预测测试

• 批准号: 0720906
• 负责人: Koushik Sen
• 金额: $ 35万
• 依托单位: University of California-Berkeley
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2007
• 资助国家: 美国
• 起止时间: 2007-09-01 至 2011-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0720906&HistoricalAwards=false
• 关键词: CSR; SMA; Predictive; Testing; System

Today's software systems suffer from poor reliability and security,with software errors costing the U.S. economy upwards of $60 billionannually. This situation is likely to get worse, as the complexity ofsoftware systems increases without a matching increase in theeffectiveness of software quality tools and techniques. The situationis particularly challenging for concurrent systems, which posesignificantly higher difficulties for software quality tools, yet arebecoming more widespread with the growing use of multicore machines.Static analysis software-quality tools are very precise but do notscale well to large codes. Testing is easy to use but for goodcoverage requires extensive test suites. We propose to bridge the gapbetween ad hoc testing and static analysis by combining them in a newscalable technique called predictive testing. Predictive testing usesstatic program analysis to maximize the effectiveness of a giventest-suite for finding in the testing stage bugs that could manifestin real production runs. Although predictive testing tools use complexstatic analysis and automated theorem proving techniques internally,all of this complexity is hidden from the user by a testing usagemodel. For this reason, we expect that such tools can be easilyintegrated into existing software engineering processes and will beusable even by unsophisticated developers.

今天的软件系统的可靠性和安全性都很差，软件错误每年给美国经济造成的损失超过600亿美元。 这种情况可能会变得更糟，因为软件系统的复杂性增加了，而软件质量工具和技术的有效性却没有相应的增加。这种情况对于并发系统来说尤其具有挑战性，这给软件质量工具带来了更大的困难，但随着多核机器的日益普及，软件质量工具的静态分析变得越来越普遍。静态分析软件质量工具非常精确，但不能很好地扩展到大型代码。 测试很容易使用，但是为了良好的覆盖率，需要大量的测试套件。 我们建议将特设测试和静态分析结合在一个新的可扩展的技术，称为预测测试，以弥合两者之间的差距。 预测测试使用静态程序分析来最大化给定测试套件的有效性，以便在测试阶段发现可能在真实的生产运行中表现出来的错误。虽然预测测试工具在内部使用复杂的静态分析和自动化定理证明技术，但所有这些复杂性都被测试使用模型隐藏起来。出于这个原因，我们期望这样的工具可以很容易地集成到现有的软件工程过程中，甚至可以被不成熟的开发人员使用。