Towards cost-effective load testing of platform-applications

平台应用程序是利用超大规模（ULS）平台提供的各种服务的应用程序。这类应用的例子包括黑莓或Android手机应用，以及基于b谷歌App Engine开发的应用，或者利用Evernote基础架构的应用。这样的应用程序必须准备好应对非常大的用户群。然而，这类应用程序的开发人员在为大型用户群进行计划和测试方面经验有限。此外，开发人员执行此类测试的资源有限。本研究项目的目标是提高此类平台应用程序的质量。我们将开发工具和流程，以便在这些应用程序部署到ULS平台之前对其进行验证。这样的验证将确保平台应用程序提供良好的用户体验，并且这些应用程序不会对其运行的ULS平台产生负面影响。提出的研究计划将以经济有效的方式大大提高平台应用程序的质量和可扩展性。这些应用程序中的许多都是由中小型企业（SME）开发的，它们执行此类深入测试的资源有限，研究预算也有限。这些应用程序经常在没有经过大规模测试的情况下被部署。这种未经测试的部署对中小企业和ULS平台的运营商产生了负面影响。例如，设计不良的平台应用程序的快速增长通常会影响同一平台上的其他应用程序。工业（黑莓）和学术（SAIL）合作伙伴在ULS平台的性能分析和验证方面都是世界公认的领导者，在优化和测试此类平台的性能方面拥有超过十年的经验。该研究建立在他们丰富的经验基础上，并将在软件性能工程、负载测试、经验软件工程以及基本的通信和演示技巧方面培训9名hqp。

Platform-applications are applications that make use of the various services provided by Ultra-Large-Scale (ULS) platforms. Examples of such applications include BlackBerry or Android mobile applications, and applications that are developed on top of Google's App Engine, or ones that make use of Evernote's infrastructure. Such applications must be ready to cope with a very large user base. However, developers of such applications have limited experience in planning and testing for large user bases. Moreover, the developers have limited resources to perform such testing. The goal of this research project is to improve the quality of such platform-applications. We will develop tools and processes to enable the verification of such applications before they are deployed on top of ULS platforms. Such verification will ensure that the platform-applications provide a good user experience and that the applications do not negatively impact the ULS platform on top of which they run. The proposed research program will lead to large improvements in the quality and scalability of platform-applications in a cost-effective manner. Many of these applications are developed by Small to Medium Enterprises (SME) with limited resources to perform such type of in-depth testing and with limited research budgets. All too often such applications are deployed without ever being tested at scale. Such un-tested deployments negatively impact the SME and the operators of ULS platforms. For example, the rapid growth of a badly-designed platform-application often impacts other applications on the same platform. Both the industrial (BlackBerry) and academic (SAIL) partners are world-recognized leaders in the performance analysis and verification of ULS platforms with over a decade of experience working on optimizing and testing the performance of such platforms. The research builds on their extensive experience and will train nine HQPs in software performance engineering, load testing, empirical software engineering, and essential communication and presentation skills.