NSFSaTC-BSF: TWC: Small: Practical Succinct Proof Systems without Trusted Setup

简洁的非交互式证明是一个强大的密码构建块，在安全云外包和物联网、医疗和金融应用等领域有许多有前途的应用，这些领域需要对隐私敏感数据进行计算。简洁证明的现有实现都依赖于可信的设置阶段，通常称为预处理。这个预处理阶段可能是有问题的，因为如果在预处理期间生成的秘密陷门被破坏，则系统将失去所有安全保证。更糟糕的是，预处理通常不是一次性的操作，必须对每个应用程序重复进行，或者更糟糕的是，对同一应用程序的不同输入大小重复进行。该项目通过探索没有可信设置的简洁证明系统来研究规避这一主要缺点的方法。这个项目创建了一个名为Lunaria的编程框架和编译器工具链，它可以将用高级语言编写的通用程序编译成简洁的零知识证明，而无需可信的设置。Lunaria旨在以非平凡的方式将联合收割机代数级、编译器级和实现级的改进结合起来，并在最先进的理论结果的直接实现上实现数量级的加速。预期的影响包括新的开源框架，用于简洁的证明，而无需可信的设置，从而使研究和行业能够获得简洁证明的力量。

Succinct non-interactive proofs are a powerful cryptographic building block with many promising applications in secure cloud outsourcing and in domains such as Internet of Things, medical, and financial applications where computation over privacy sensitive data is desired. Existing implementations of succinct proofs all rely on a trusted setup phase, commonly known as preprocessing. This preprocessing phase can be problematic since the system loses all security guarantees if the secret trapdoor generated during preprocessing is compromised. To make matters worse, the preprocessing is often not a one-time operation, and must be repeated for each application or, even worse, for different input sizes of the very same application.The project investigates ways to circumvent this major drawback by exploring succinct proof systems without trusted setup. This project creates a programming framework and compiler toolchain called Lunaria which compiles general programs, written in high-level languages, into succinct zero-knowledge proofs without trusted setup. Lunaria aims to combine algebraic-, compiler-, and implementation-level improvements in non-trivial ways, and enable orders of magnitude speedup over a straightforward implementation of state-of-the-art theoretical results. Expected impacts include new, open-source frameworks for succinct proofs without trusted setup, thus giving research and industry access to the power of succinct proofs.