Authors: Mohammad Hashemi, Shahin Tajik, Fatemeh Ganji
Affiliation: Worcester Polytechnic Institute (WPI), USA
Garblet is a secure multi-party computation (MPC) framework designed for protecting chiplet-based systems. It leverages Garbled Circuits (GC) and Oblivious Transfer (OT) to enable efficient and secure computation in the presence of potentially compromised chiplets. The framework is implemented on an AMD/Xilinx UltraScale+ multi-chip module and introduces:
- A customized hardware OT module to reduce communication costs.
- An optimized evaluator engine for efficient garbled circuit evaluation.
- A novel circuit decomposition technique enabling parallel processing across multiple chiplets.
- Secure computation while minimizing overhead compared to traditional server-client MPC setups.
- Chiplet-Based Secure Computation: Secure function evaluation between chiplets, preventing malicious chiplets from leaking sensitive information.
- Efficient Garbled Circuits Implementation: Reduces communication complexity using a dedicated hardware OT module.
- Parallel Execution: A circuit decomposition method enables multiple chiplets to compute in parallel, accelerating execution time.
- Hardware Security Isolation: Separates security-critical tasks from general computations for enhanced security.
- Paper PDF: The research paper presenting Garblet.
- Implementation Details: Source code, experimental setup, and evaluation metrics.
- Benchmark Results: Performance comparisons with traditional MPC implementations.
If you use Garblet in your research, please cite:
@article{hashemi2025garblet,
title={Garblet: Multi-party Computation for Protecting Chiplet-based Systems},
author={Hashemi, Mohammad and Tajik, Shahin and Ganji, Fatemeh},
journal={Cryptology ePrint Archive},
year={2025}
}
For any questions or collaborations, feel free to reach out to:
📧 [email protected] | [email protected]
This repository is under construction and is dedicated to advancing secure computation for chiplet-based architectures. Contributions and discussions are welcome!