A 10.33 μJ/encryption Homomorphic Encryption Engine in 28nm CMOS with 4096-degree 109-bit Polynomials for Resource-Constrained IoT Clients

• Main Authors: Das, Siddharth, Hagen, McKenzie Van Der, Patil, Swarali, Erbagci, Cagri, Lucia, Brandon, Mai, Ken
• Format: Conference Proceeding
• Language: English
• Subjects: Accelerator; Homomorphic Encryption; Internet of Things; Machine learning; Memory management; Number Theoretic Transform; Performance evaluation; Silicon; Solid state circuits; Voltage
• ISSN: 2643-1319
• DOI: 10.1109/ESSCIRC59616.2023.10268762

Homomorphic Encryption (HE) is used to protect sensitive client data during offloaded compute on a potentially untrusted server. Unfortunately, the computational intensity of HE operations quickly depletes the limited resources on IoT clients. Thus, we present an energy-efficient silicon implementation of encryption/decryption in the Brakerski-Fan Vercauteren HE scheme. To support meaningful applications, including several machine learning workloads, we optimize for fixed parameters N =4096 and log_{2}q =109 through pipelining, multi-level parallelism, and efficient memory accesses. At an energy-optimal operating point of 60MHz and 0. 64V our chip, fabricated in a 2Snm bulk planar CMOS process with an accelerator core area of 1.69 m\mathrm{m}^{2}, consumes 10.33 \muJ/encryption. Ultimately, this work enables IoT clients to participate in privacy-preserving offloaded compute using client-aided Homomorphic Encryption.