SHF: Small: Evolving Safety Cases in Agile Development Environments

在安全关键领域运行的软件不仅必须支持其预期的功能，而且必须确保使用安全。交付此类系统需要进行严格和系统的危害分析，以识别和减轻潜在危害。在部署之前，通常构建一个安全案例，为系统安全性提供声明、证据和论证。传统上，安全关键型系统是通过精心控制的过程来设计的，这些过程强调详细的规划、预先设计和质量保证。这导致了一种被称为“大冻结”的现象，即引入新功能的成本、努力和难度变得非常昂贵。因此，许多在安全关键领域运营的组织正在采用更敏捷的方法，在这种方法中，软件以更短的发布周期交付。与此同时，传统上没有在安全领域工作的组织正在越来越多地构建网络物理系统，例如工厂车间机器人、无人驾驶飞机系统和医疗设备，但通常没有知识或工具来支持适当的危害分析和安全保证。这两种趋势——从过程谱的两端出现——指向了开发安全关键型软件的一种新方法，这种方法包含了安全关键型开发的严谨性，同时受益于敏捷解决方案可能带来的更多增量、更快的交付周期。研究团队将提供一个智能的解决方案，用于在敏捷的安全关键项目环境中创建、发展和使用跟踪链接。新颖的软件工件森林分析（SAFA）方法将帮助在敏捷环境中工作的验证者、验证者、安全分析师和其他项目利益相关者理解和分析变更对现有安全案例的影响，评估当前系统的安全性，并相应地发展安全案例。该研究将提供一个过程工作流，用于指导开发人员完成创建跟踪链接的任务，用于在敏捷项目中自动创建和发展跟踪链接的可跟踪性解决方案，用于可视化系统如何减轻已识别的危险的交互式解决方案，以及用于支持变更影响分析和安全保证用例维护的技术。该研究直接解决了在安全关键项目中采用敏捷过程以解决“大冻结”问题的新兴工业挑战。通过该项目交付的算法、工具和流程预计将对工业产生重大影响。在整个项目中，工业合作伙伴将积极参与测试，推动创新和实用的解决方案。通过让处于学术生涯各个阶段的代表性不足的学生参与具有挑战性的研究项目，将为扩大计算机领域的参与提供机会。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Software operating in safety-critical domains must not only support its intended functionality but must also be assuredly safe for use. Delivering such systems requires a rigorous and systematic hazard analysis to identify and mitigate potential hazards. Prior to deployment, a safety case is often constructed that provides claims, evidence, and arguments for system safety. Safety-critical systems have traditionally been engineered using carefully controlled processes which emphasize detailed planning, upfront design, and quality assurance. This has led to the phenomenon referred to as the 'big freeze' in which the cost, effort, and difficulty of introducing new functionality becomes prohibitively expensive. As a result, many organizations operating in safety-critical domains are adopting more agile approaches in which software is delivered on shorter release cycles. At the same time, organizations that have not traditionally worked in the safety domain are increasingly building Cyber-Physical Systems, such as factory-floor robots, unmanned aerial systems and medical devices, often without the knowledge or tools to support appropriate hazard analysis and safety assurance. These two trends -- emerging from opposite ends of the process spectrum -- point to a new way of developing safety-critical software, one which embraces the rigor of safety-critical development while benefiting from the more incremental, faster delivery cycles made possible by agile solutions.The research team will deliver an intelligent solution for creating, evolving, and using trace links within agile safety-critical project environments. The novel Software Artifact Forest Analysis (SAFA) approach will aid validators, verifiers, safety analysts, and other project stakeholders working in an agile environment to understand and analyze the impact of change upon an existing safety case, to assess the safety of the current system, and to evolve the safety case accordingly. The research will deliver a process workflow for guiding developers through the task of creating trace links, traceability solutions for automating the creation and evolution of trace links in an agile project, interactive solutions for visualizing how the system mitigates identified hazards, and techniques for supporting change impact analysis and maintenance of safety-assurance cases. The research directly addresses the emerging industrial challenge of adopting agile processes in safety-critical projects in order to address the `big freeze' problem. The algorithms, tools, and processes delivered through the project are expected to have significant industrial impact. This will be aided by the proactive engagement of industrial partners in test-driving novel and practical solutions produced throughout this project. Opportunities will be provided for broadening participation in computing by engaging underrepresented students at all stages of their academic careers in challenging research projects.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.