Sean Choi ^1* Disha Patel ¹ Diman Zad Tootaghaj ² Lianjie Cao ² Faraz Ahmed ² Puneet Sharma ²

• ¹ Santa Clara University, Santa Clara, United States
• ² Hewlett Packard Enterprise, Alpharetta, Georgia, United States

Federated Learning (FL) has emerged as a promising paradigm for secure distributed machine learning model training across multiple clients or devices, enabling model training without having to share data across the clients. Yet, recent studies revealed that FL can be vulnerable to data leakage and reconstruction attacks even if the data itself is never shared with another client. Thus, to resolve such vulnerability and improve privacy of all clients, a class of techniques, called privacy-preserving FL, incorporates encryption techniques, such as Homomorphic Encryption (HE), to encrypt and fully protect model information from being exposed to other parties. A downside to this approach is that encryption schemes like HE are very compute-intensive, often causing inefficient and excessive use of client CPU resources that can be used for other uses.To alleviate this issue, this work introduces a novel approach by leveraging Smart Network Interface Cards (SmartNICs) to offload compute-intensive HE operations of privacy-preserving FL. By employing SmartNICs as hardware accelerators, we enable efficient computation of HE while saving CPU cycles and other server resources for more critical tasks. In addition, by offloading encryption from the host to another device, the details of encryption remains secure even if the host is compromised, ultimately improving security of the entire FL system.Given such benefits, this paper presents a FL system named FedNIC that implements the above approach, with an in-depth description of the architecture, implementation, and performance evaluations. Our experimental results demonstrate a more secure FL system with no loss in model accuracy and up to 25% in reduced host CPU cycle, but with roughly 46% increase in total training time, showing the feasibility and trade offs of utilizing SmartNICs as an encryption offload device in federated learning scenarios. Finally, we illustrate promising future work and potential optimizations for a more secure and privacy-preserving federated learning system.1 Choi et al.