TWC: Small: Scalable Hybrid Attack Graph Modeling and Analysis

网络物理系统（cps）运行着几乎所有的社会关键基础设施（如能源、交通和医疗）。在执行关键任务功能时，cps表现出混合（离散和连续）行为，因为它们使用数字技术来控制和监视物理过程。CPS安全分析特别具有挑战性，因为攻击者可以利用系统的数字元素（例如，网络）、系统的物理元素或某些组合中的各种漏洞。该项目正在开发一个数学和计算框架，用于对大型复杂的cps进行建模和分析，以捕获其漏洞和由此产生的攻击路径（攻击者可能用于破坏系统的步骤）。该项目正在开发混合攻击图（hag），作为表示cps中的安全属性和复合暴露的数学形式。hag反映了暴露的功能视图，捕获由于在网络或物理域中执行漏洞利用而通过cps进行的状态转换。因此，它们提供了全面记录CPS攻击面的潜力。研究人员正在设计HAG生成算法，应用智能搜索和并行化策略，并创建一套基于web的工具，以应对大规模CPS攻击面建模的计算负担。该项目正在开发一系列基于马尔可夫过程、经典可达性和其他技术的分析方法。为了提供一个实验平台来评估项目的工具和技术，研究人员正在建造一个由网络控制的机器人车辆组成的CPS测试平台。测试平台还将提供一个竞争激烈的学习环境，以有趣和吸引人的方式向学生讲授CPS安全原理。

Cyber-physical systems (CPSs) operate nearly all of society's critical infrastructures (e.g., energy, transportation and medicine). In performing mission critical functions, CPSs exhibit hybrid (both discrete and continuous) behavior as they use digital technology to control and monitor physical processes. CPS security analysis is particularly challenging because an attacker can make use of a wide variety of vulnerabilities in the digital elements of the system (e.g., the network), the physical elements of the system, or some combination. This project is developing a mathematical and computational framework for modeling and analyzing large, complex CPSs to capture their vulnerabilities and the resulting attack paths (steps an attacker might use to disrupt the system).This project is developing hybrid attack graphs (HAGs) as a mathematical formalism for representing security properties and compound exposures in CPSs. HAGs reflect a functional view of exposures, capturing state transitions over CPSs due to the execution of exploits in either the cyber or physical domains. As such, they offer the potential to comprehensively document a CPS attack surface. The researchers are designing HAG generation algorithms that apply intelligent search and parallelization strategies and creating a suite of web-based tools to cope with the computational burdens of large-scale CPS attack surface modeling. The project is developing an array of analytical methods, refined based on Markov Processes, classic reachability, and other techniques. To provide an experimentation platform for evaluating the project's tools and techniques, the researchers are building a CPS test bed comprised of network-controlled robotic vehicles. The testbed will also provide a competitive learning environment in which to teach students about CPS security principles in a fun and engaging manner.