InTrans: Modular Security on an Open Cloud

这个项目探索了两种工具组合所产生的有趣可能性：加密软件，只要机器是隔离的，就可以将任何计算任务分发到几台机器上，并提供强大的安全保证；多提供商云数据中心，为任何租户提供租用由不同组织管理的多台隔离机器的能力。该项目的新颖之处在于优化用于多提供者数据中心的加密安全多方计算（MPC），利用这些数据中心已经提供的安全属性。该项目的影响是开源软件包，它为采用多提供商方法的云数据中心（如大规模开放云（MOC））提供了引人注目的应用程序。两个工业合作伙伴本田研究院和红帽对InTrans项目的研究非常感兴趣，它建立在云安全前沿模块化方法项目中所做的工作的基础上。更详细地说，这个跨学科项目包括三组任务，它们跨越密码学和分布式系统。首先，研究人员探索改进和优化MPC算法设计的机会，以满足我们观察到的行业合作伙伴和其他人对特定场景的巨大兴趣。其次，研究人员调整和优化系统软件和现有的MOC工具，以改进MPC的网络化和分布式计算步骤。第三，研究人员对MOC的安全相关系统进行了通用可组合性（UC）安全分析，以确定其对MPC的贡献程度。这三个任务是相互依赖的：例如，MOC的UC模型可以为新的MPC算法的设计提供见解，该算法将软件与部分可信的硬件相结合，并验证云是否满足MPC所需的先决条件（例如节点之间的隔离）。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。

This project explores the intriguing possibilities that result from the combination of two tools: cryptographic software that distributes any computing task over several machines with strong security guarantees as long as the machines are isolated, and a multi-provider cloud datacenter that offers to any tenant the ability to rent multiple isolated machines that are administered by different organizations. The project's novelties are to optimize cryptographically secure multi-party computation (MPC) for use in a multi-provider datacenter, leveraging the security properties already provided by such datacenters. The project's impacts are open-source software packages that provide a compelling application for cloud datacenters like the Mass Open Cloud (MOC) that adopt a multi provider approach. The research in this InTrans project is of strong interest to two industrial partners, Honda Research Institutes and Red Hat, and it builds on top of the work performed in the Modular Approach to Cloud Security Frontier project.In more detail, this cross-disciplinary project encompasses three sets of tasks that span cryptography and distributed systems. First, the investigators explore opportunities to improve and optimize the algorithmic design of MPC toward specific scenarios for which we have observed large interest by our industry partners and others. Second, the investigators adapt and optimize system software and existing MOC tools to improve MPC's networking and distributed computing steps. Third, the investigators perform a universal composability (UC) security analysis of the MOC's security-relevant systems in order to determine the extent to which can contribute toward MPC. These three tasks are co-dependent: for instance, a UC model of the MOC can provide insights into the design of new MPC algorithms that combine software with partially trusted hardware and to validate whether the cloud meets the pre-conditions required by MPC (such as isolation between nodes).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.