Automated Testing of Web Services and Service-Oriented Computing

随着包括亚马逊、苹果、谷歌、IBM、微软和雅虎在内的主要技术参与者最近对面向服务的计算的投资，面向服务的计算的驱动力显而易见。虽然这些新系统仍然与传统的信息技术系统和现代的基于Web的应用程序密切相关，但它们提出了一些独特的挑战，需要创新的软件工程方法。面向服务的计算系统中的独特测试挑战包括Web服务端点的验证和跨多个异构节点的系统状态的验证。测试传统软件系统的非功能性质量属性是一项具有挑战性的工作，对于面向服务的计算也是如此。虽然服务之间的安全基础设施已经建立，但需要开发策略、工具和测试技术来减少Web服务中的缺陷数量，以产生安全的面向服务的计算系统。本研究工作的目标是产生一个低成本的自动化测试解决方案，用于面向服务的计算。这一目标将通过以下方式实现：（i）从WSDL规范中为上下文无关的语法输入定义创建自动提取机制;（ii）基于Web服务和HTTP服务器日志的运行时跟踪开发自动逆向工程输入约束求解器;（iii）开发自动缺陷检测以帮助整体测试解决方案;以及（iv）为现有的企业服务总线基础设施开发一个不变验证框架。研究结果预计将有显着的影响，强大的Web服务和下一代面向服务的计算基础设施的发展，以及更广泛的Web工程和软件测试领域。

With recent investments in service-oriented computing by major technological players including Amazon, Apple, Google, IBM, Microsoft, and Yahoo, the drive towards service-oriented computing is evident. While these novel systems remain closely tied to traditional information technology systems and modern web-based applications, they pose a number of unique challenges for which innovative software engineering methodologies are required. Unique testing challenges within service-oriented computing systems include the verification of the web-service endpoints and the verification of the state of the system across multiple heterogeneous nodes. Testing the non-functional quality attributes of traditional software systems is a challenging endeavour, and the same holds true for service-oriented computing. Although infrastructure for security between services is established, development strategies, tools and testing techniques that decrease the number of defects within web-services are required to produce secure service-oriented computing systems. The objective of this research endeavour is produce a low-cost automated testing solution for use in service-oriented computing. This goal is to be realized through: (i) the creation of an automated extraction mechanism for a context-free grammatical input definition from a WSDL specification; (ii) the development of an automated reverse-engineering input-constraints solver based on the runtime traces of a web-service and HTTP server logs; (iii) the development of automated defect detection to aid the overall testing solution; and (iv) the development of an invariant-validation framework for existing Enterprise Service Bus infrastructure. The outcome of the research is expected to have significant implications for the development of robust web-services and the next generation of service-oriented computing infrastructure, as well as for the broader web engineering and software testing fields.