Collaborative Research: SaTC: CORE: Small: Hyperproperty-based Enforcement of Information-flow Security

对于处理大量用户敏感信息的软件系统来说，执行保密或完整性等安全策略是至关重要的，即使是短暂的违反这些策略也可能导致泄露或破坏高度敏感的信息，危及安全，或中断重要的公共或社会服务。运行时监视和执行安全策略仍然是保护软件系统的基本技术之一。许多重要的安全策略（如信息流策略）不能表示为单个执行的属性，但可以表示为多个执行轨迹上的属性，这些属性称为超属性。受一组丰富的现实世界应用程序的启发，该项目的总体目标是为超属性开发有效的运行时实施技术，从而能够强大地保护系统免受各种类型的网络攻击和信息泄露。该项目将在信息流属性的系统运行时执行中带来范式转变，组织在三个研究重点中，将开发新的监视设计、开源工具和执行严格的评估。第一个重点将是开发用于执行超属性的黑盒和灰盒预测监视器。这个推力还将研究不同的输入模型，特别是监视器可以操纵系统输入以强制执行超属性的封闭系统和监视器不能干扰输入的反应系统，因为它是由不可控的环境产生的。第二个要点将集中于开发组合的运行时监视器。该项目将设计一个组合监视基础设施，其中系统的各个部分由不同的监视器保护，因为现代系统由来自相互不信任的来源的异构组件组成。第三个研究重点是通过深入研究研究人员的理论发现的应用，致力于严格的评估。此工作将应用前两个重点的结果来实现监视器，以在实际应用程序上强制执行信息流属性和不可延展性信息流。与现有的方法相比，这个项目的结果预计有几个优点：它们将更加通用，因为它们可以处理时间逻辑HyperLTL的丰富片段，并且具有可转移性、可组合性和低开销。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。

Enforcing security policies such as secrecy or integrity is crucial for software systems that handle large volumes of users’ sensitive information, where even a short transient violation of these policies may result in leaking or damaging highly sensitive information, compromising safety, or interruption of vital public or social services. Runtime monitoring and enforcement of security policies remains one of the foundational techniques for securing software systems. Many important security policies such as information-flow policies cannot be expressed as properties of individual executions, but can be expressed as properties on multiple execution traces, which are called hyperproperties. Motivated by a rich set of real-world applications, the overarching objective of this project is to develop effective runtime enforcement techniques for hyperproperties that enable strong protection of systems against various types of cyberattacks and information leaks. This project will bring about a paradigm shift in systematic runtime enforcement of information-flow properties, organized within three research thrusts that will develop novel monitor designs, open-source tools, and perform rigorous evaluation. The first thrust will develop black-box and grey-box predictive monitors for enforcing hyperproperties. This thrust will also investigate different input models, in particular, closed systems where the monitor can manipulate the input to the system to enforce a hyperproperty and reactive systems where the monitor cannot interfere with the input, as it is generated by an uncontrollable environment. The second thrust will focus on developing runtime monitors that are compositional. The project will design a composed monitor infrastructure where parts of the system are guarded by different monitors, as modern systems are composed of heterogeneous components from mutually distrusting sources. The third research thrust is dedicated to rigorous evaluation by delving into applications of the researchers' theoretical findings. This effort will apply results from the first two thrusts to implement monitors to enforce information flow properties and nonmalleable information flow on real world applications. The results of this project are expected to have several advantages as compared to the existing methods: they will be more general, as they can deal with a rich fragment of the temporal logic HyperLTL and will be transferable, compositional, and low-overhead.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.