The Somerset Laboratory: An Open Facility for Experimentation in Safety-Critical Systems

萨默塞特实验室是弗吉尼亚大学正在开发的一个设施，重点关注安全关键软件工程的实验研究。 从长远来看，实验室的目的是建立安全关键系统的实证研究作为软件工程的一个子领域。 该实验室通过互联网向世界各地的研究人员提供以下内容： (1) 与少量安全关键应用程序相关的一套全面的工件，尽可能准确、全面地描述应用程序。有足够的工件并且工件具有足够的质量，研究人员可以进行大量的实验，这些实验将与他们可以直接访问实际应用程序时可以执行的实验非常相似。 (2) 使用可用应用程序进行的研究项目的结果存储库。该存储库包含贡献的工件、工具、研究论文、数据文件和联系信息。除了允许传播结果之外，该存储库还通过提醒研究人员其他人的状态、计划和工作来促进合作研究。 (3) 用于准确比较安全关键系统的软件工程技术的工具。利用与实验室中安全关键应用相关的工件，研究人员能够评估工具和技术，然后比较和对比不同的工具和技术。 在这种情况下，比较尤其重要，因为数据是通过将它们应用到完全相同的应用程序来获得的。 通过使用来自实际安全关键应用的大量材料进行实验，研究人员了解了此类应用的复杂细节。 处理这些细节有助于确保所进行的研究的相关性和实用价值。此外，一旦设计出解决方案概念或方法，就可以通过将其应用于一个或多个安全关键应用示例来对其进行评估（至少部分地），从而提供可信的评估。该实验室选择支持的第一个安全关键应用是核反应堆。该反应堆是一种低功率研究反应堆，采用传统的机电控制系统。 尽管它是一个小型反应堆，但它对对基于软件的控制和安全系统感兴趣的研究人员提出了重大挑战。正在为实验室开发的工件包括反应堆物理结构的详细信息、必须执行的安全规则、物理设备关键元件的数字化图像、所有反应堆传感器的样本痕迹、反应堆模拟器和完整的接线图。 该规划拨款的目标是获得研究界对实验室可能感兴趣程度的评估，并制定实验室实施的详细计划。 ***

The Somerset Laboratory is a facility being developed at the University of Virginia that provides a focus for experimental research in safety-critical software engineering. Over the long term, the intent of the laboratory is to establish empirical research with safety- critical systems as a sub-field of software engineering. The laboratory provides the following to researchers World wide, via the Internet: (1) A comprehensive set of artifacts associated with a small number of safety-critical applications that, to the extent possible, describe the applications accurately and comprehensively. There are sufficient artifacts and the artifacts are of sufficient quality that researchers can perform substantial experiments which will be very similar to those that they could be perform if they had direct access to the actual applications. (2) A repository of the results of research projects undertaken using the available applications. This repository contains contributed artifacts, tools, research papers, data files, and contact information. As well as allowing dissemination of results, the repository facilitates cooperative research by alerting researchers to the status, plans, and work of others. (3) A vehicle for accurate comparison of software engineering techniques for safetycritical systems. Using the artifacts associated with the safety-critical applications in the laboratory, researchers are able to evaluate tools and techniques and then compare and contrast different tools and techniques. The comparisons are especially significant in this case because data is obtained by applying them to precisely the same applications. By carrying out experiments using extensive material from actual safety-critical applications, researchers are presented with the complex details of such applications. Dealing with these details helps to ensure the relevance and practical value of the research that is undertaken. In addition, once a solution concept or approach is devised, it can be evaluated (at least in part) by applying it to one or more sample safetycritical applications thus providing a believable assessment. The first safety-critical application that the laboratory has chosen to support is a nuclear reactor. The reactor is a lowpower, research reactor that employs a conventional electromechanical control system. Though a small reactor, it presents significant challenges to researchers interested in softwarebased control and safety systems. The artifacts being developed for the laboratory include details of the reactor's physical structure, the safety rules that must be enforced, digitized images of key elements of the physical equipment, sample traces of all the reactor's sensors, a reactor simulator, and complete wiring diagrams. The goal of this planning grant is to obtain an assessment of the likely level of interest in the laboratory by the research community and to develop a detailed plan for the implementation of the laboratory. ***