Search papers, labs, and topics across Lattice.
This paper introduces the distributed function accelerator (DFA), a hardware solution designed to optimize function secret sharing (FSS) by enhancing the efficiency of distributed point function (DPF) generation and evaluation. By integrating a high-throughput fixed-function engine for AES-based pseudorandom number generation with a lightweight programmable unit, DFA significantly reduces end-to-end latency, communication overhead, and energy consumption in both untrusted and trusted modes. The results demonstrate a remarkable 10X reduction in latency and up to 20X reduction in communication for secure inference, alongside substantial improvements in throughput and energy savings for private information retrieval (PIR).
Achieving up to 20X reductions in communication overhead and 10X latency cuts, DFA revolutionizes the efficiency of function secret sharing in privacy-preserving systems.
Function secret sharing (FSS) is a core building block for privacy-preserving systems such as secure inference and private information retrieval (PIR), but incurs significant overhead in key generation, communication, and data movement.We present the distributed function accelerator (DFA), a hardware accelerator that targets the dominant primitive in FSS: distributed point function (DPF) generation and evaluation. DFA combines a high-throughput fixed-function engine for AES-based pseudorandom number generation with a lightweight programmable unit for protocol-specific logic. In untrusted mode, DFA serves as a pure accelerator for DPF evaluation, improving throughput and energy efficiency without changing the protocol. In trusted mode, it further enables local, on-the-fly key generation, eliminating key distribution, and reducing storage and data movement overheads. Across representative workloads, DFA achieves a reduction of 10X end-to-end latency, a reduction of up to 20X communication and more than 5X energy savings for secure inference; and an improvement of 5X throughput and 10X energy reduction for PIR, with modest hardware cost.
