As cloud computing becomes more popular, research has focused on usable solutions to the problem of verifiable computation (VC), where a computationally weak device (Verifier) outsources a program execution to a powerful server (Prover) and receives guarantees that the execution was performed faithfully. A Prover can further demonstrate knowledge of a secret input that causes the Verifier’s program to satisfy certain assertions, without ever revealing which input was used. State-of-the-art Zero-Knowledge Proofs of Knowledge (ZKPK) methods encode a computation using arithmetic circuits and preserve the privacy of Prover’s inputs while attesting the integrity of program execution. Nevertheless, developing, debugging, and optimizing programs as circuits remains a daunting task, as most users are unfamiliar with this programming paradigm. In this work, we present Zilch, a framework that accelerates and simplifies the deployment of VC and ZKPK for any application transparently , i.e., without the need of trusted setup. Zilch uses traditional instruction sequences rather than static arithmetic circuits that would need to be regenerated for each different computation. Towards that end, we have implemented Z MIPS: a MIPS-like processor model that allows verifying each instruction independently and compose a proof for the execution of the target application. To foster usability, Zilch incorporates a novel cross-compiler from an object-oriented Java-like language tailored to ZKPK and optimized our Z MIPS model, as well as a powerful API that enables integration of ZKPK within existing C/C++ programs. In our experiments, we demonstrate the flexibility of Zilch using two real-life applications, and evaluate Prover and Verifier performance on a variety of benchmarks.

中文翻译：

---

Zilch：部署透明零知识证明的框架

随着云计算变得越来越流行，研究集中在可验证计算 (VC) 问题的可用解决方案上，其中计算能力较弱的设备 (Verifier) 将程序执行外包给强大的服务器 (Prover) 并获得执行已执行的保证忠实。证明者可以进一步证明秘密输入的知识，该秘密输入导致验证者的程序满足某些断言，而无需透露使用了哪个输入。最先进的零知识知识证明 (ZKPK) 方法使用算术电路对计算进行编码，并在证明程序执行完整性的同时保护证明者输入的隐私。然而，开发、调试和优化程序作为电路仍然是一项艰巨的任务，因为大多数用户不熟悉这种编程范式。在这项工作中，我们提出了 Zilch，这是一个框架，可以加速和简化 VC 和 ZKPK 为任何应用程序的部署透明地 ，即不需要可信设置。Zilch 使用传统的指令序列，而不是需要为每个不同的计算重新生成的静态算术电路。为此，我们实施了 Z MIPS：一种类似 MIPS 的处理器模型，允许独立验证每条指令并为目标应用程序的执行编写证明。为了提高可用性，Zilch 结合了一种新颖的交叉编译器，该交叉编译器来自一种为 ZKPK 量身定制的面向对象的类 Java 语言，并优化了我们的 Z MIPS 模型，以及一个强大的 API，可以将 ZKPK 集成到现有 C/C++ 程序中。在我们的实验中，我们使用两个实际应用程序展示了 Zilch 的灵活性，并在各种基准测试中评估了 Prover 和 Verifier 的性能。