Introducing fault-based combinatorial testing to web services
Introducing fault-based combinatorial testing to web services
复制标题
DOI：
10.1109/secon.2010.5453905
复制
发表时间：
2010-03
期刊：
影响因子：
--
通讯作者：
K. Z. Watkins
中科院分区：
文献类型：
--
作者：
K. Z. Watkins
组合测试被认为对于发现软件故障是有效的。它也很高效，因为它使测试数量相对较少。然而，似乎很少有研究将组合测试视为 Web 服务的测试方法。通常通过向网络注入引起故障的数据扰动来测试它们。看看组合测试是否可以利用组合测试的好处来补充现有的扰动可能是值得的。本文提出的方法称为基于组合故障的测试。这种类型的测试将现有的基于故障的测试技术（例如故障注入）与组合测试相结合，以尝试查找 Web 服务中不同强度的故障。基于组合故障的测试使用故障注入，并通过仅关注基于故障的组合来帮助减少组合爆炸问题。这就提出了以下研究问题：是否有一种方法可以在源代码不可用的情况下利用 Web 服务组合测试的优势，同时最大限度地减少组合爆炸的可能性？基于组合故障的测试看起来非常有希望回答这个问题。作为一个副作用，它可能提供一种方法来确定测试 Web 服务的最大交互强度。
Combinatorial testing is considered effective for finding software faults. It is also efficient, since it keeps the number of tests relatively small. However, there seems to be very little research that considers combinatorial testing as a testing approach for web services. They are commonly tested by injecting fault-causing data perturbations into the network. It may be worthwhile to see if combinatorial testing can complement existing perturbations with the benefits of combinatorial testing. The approach proposed in this paper is called combinatorial fault-based testing. This type of testing combines existing fault-based testing techniques, such as fault injection, with combinatorial testing to attempt to find faults of varying strength within a web service. Combinatorial fault-based testing uses fault injection and helps reduce the problem with combinatorial explosion by focusing solely on fault-based combinations. This raises the following research question: Is there a way to take advantage of the benefits of combinatorial testing for web services assuming that source code will not be available, while minimizing the possibility of a combinatorial explosion? Combinatorial fault-based testing looks very promising for answering this question. As a side effect, it could potentially offer a way to determine the maximum strength of interactions to test for web services.