DNA-HHE: Dual-mode Near-network Accelerator for Hybrid Homomorphic Encryption on the Edge

TL;DR

The paper addresses the high latency and ciphertext expansion of FHE-based PPOC on edge devices by proposing a dual-mode HHE accelerator, DNA-HHE, that supports edge-side RNS-CKKS and Rubato SE within a single compact architecture tightly coupled to the NIC. It introduces a DSP-efficient, multi-field-adaptive BFU and out-of-order task scheduling to maximize edge performance while maintaining small area overhead, enabling flexible switching between computation modes. A near-network packaging capability reduces ciphertext transmission latency, achieving measurable end-to-end speedups in both ASIC and FPGA implementations and outperforming prior single-mode accelerators. The work clarifies when each mode is advantageous and provides design guidelines for dual-mode HHE on resource-constrained edges, with practical implications for privacy-preserving edge computing and PPOC pipelines.

Abstract

Fully homomorphic encryption (FHE) schemes like RNS-CKKS enable privacy-preserving outsourced computation (PPOC) but suffer from high computational latency and ciphertext expansion, especially on the resource-constrained edge side. Hybrid Homomorphic Encryption (HHE) mitigates these issues on the edge side by replacing HE with lightweight symmetric encryption for plaintext encryption, such as the Rubato cipher for the HHE variant of RNS-CKKS, yet it introduces transciphering overhead on the cloud. The respective strengths and limitations of FHE and HHE call for a dual-mode HHE solution with flexible algorithm switching ability. This paper presents DNA-HHE, the first dual-mode HHE accelerator with near-network coupling for edge devices. DNA-HHE supports both edge-side RNS-CKKS and Rubato within a unified architecture driven by flexible custom instructions. To realize a compact implementation for the edge side, we propose a DSP-efficient modular reduction design, a compact multi-field-adaptive butterfly unit, and parallel scheduling schemes of Rubato with a high degree of resource sharing. DNA-HHE is designed with network protocol packaging and transmission capacities and directly coupled to the network interface controller, achieving reduced overall latency of edge-side PPOC by 1.09$\times$ to 1.56$\times$. Our evaluations on the ASIC and FPGA platforms demonstrate that DNA-HHE outperforms the state-of-the-art single-mode designs in both edge-side RNS-CKKS and symmetric cipher with better computation latency and area efficiency, while offering dual-mode functionality.

Figures (14)

• Figure 1: End-to-end workflow of traditional FHE and Hybrid HE scheme in PPOC with edge computing paradigm.
• Figure 2: Dataflow of edge-side RNS-CKKS scheme.
• Figure 3: The round function of Rubato. Symbols $\bigodot$ and $\bigoplus$ respectively represent the PM-Mul and PW-Add over a finite field $\mathbb{Z}_t$
• Figure 4: High-level architecture of edge-side FHE accelerator and dataflow. The red path illustrates the steps involved in the m-to-ct conversion and network packets sending, and the green path shows the steps for network packets receiving and the ct-to-m conversion. Thicker lines with darker colors represent larger data transmission volumes.
• Figure 5: Overview of the high-level hardware architecture of DNA-HHE.