{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T16:41:28Z","timestamp":1778344888659,"version":"3.51.4"},"reference-count":56,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,10,21]],"date-time":"2024-10-21T00:00:00Z","timestamp":1729468800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Sci."],"abstract":"<jats:p>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. However, recent studies revealed that FL could be vulnerable to data leakage and reconstruction attacks even if the data itself are never shared with another client. Thus, to resolve such vulnerability and improve the 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 study 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 remain secure even if the host is compromised, ultimately improving the security of the entire FL system. Given such benefits, this paper presents an 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 a roughly 46% increase in total training time, showing the feasibility and tradeoffs of utilizing SmartNICs as an encryption offload device in federated learning scenarios. Finally, we illustrate promising future study and potential optimizations for a more secure and privacy-preserving federated learning system.<\/jats:p>","DOI":"10.3389\/fcomp.2024.1465352","type":"journal-article","created":{"date-parts":[[2024,10,21]],"date-time":"2024-10-21T12:12:16Z","timestamp":1729512736000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":11,"title":["FedNIC: enhancing privacy-preserving federated learning via homomorphic encryption offload on SmartNIC"],"prefix":"10.3389","volume":"6","author":[{"given":"Sean","family":"Choi","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Disha","family":"Patel","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Diman","family":"Zad Tootaghaj","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lianjie","family":"Cao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Faraz","family":"Ahmed","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Puneet","family":"Sharma","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2024,10,21]]},"reference":[{"key":"B1","unstructured":"AbadiM.\n            AgarwalA.\n            BarhamP.\n            BrevdoE.\n            ChenZ.\n            CitroC.\n          TensorFlow: Large-scale machine learning on heterogeneous systems2015"},{"key":"B2","doi-asserted-by":"crossref","first-page":"882","DOI":"10.1109\/HPCA56546.2023.10070953","article-title":"\u201cFab: an fpga-based accelerator for bootstrappable fully homomorphic encryption,\u201d","volume-title":"2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)","author":"Agrawal","year":"2023"},{"key":"B3","unstructured":"Server Processor Specifications2024"},{"key":"B4","unstructured":"BadawiA. A.\n            BatesJ.\n            BergamaschiF.\n            CousinsD. B.\n            ErabelliS.\n            GeniseN.\n          Openfhe: Open-source fully homomorphic encryption library. Cryptology ePrint Archive, Paper 2022\/9152022"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2104.0315","article-title":"Tenseal: a library for encrypted tensor operations using homomorphic encryption","author":"Benaissa","year":"2021","journal-title":"arXiv"},{"key":"B6","article-title":"Protection against reconstruction and its applications in private federated learning","author":"Bhowmick","year":"2019","journal-title":"ArXiv"},{"key":"B7","article-title":"\u201cMachine learning with adversaries: byzantine tolerant gradient descent,\u201d","author":"Blanchard","year":"2017","journal-title":"Advances in Neural Information Processing Systems, Vol. 30"},{"key":"B8","doi-asserted-by":"crossref","first-page":"1175","DOI":"10.1145\/3133956.3133982","article-title":"\u201cPractical secure aggregation for privacy-preserving machine learning,\u201d","volume-title":"Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security","author":"Bonawitz","year":"2017"},{"key":"B9","doi-asserted-by":"publisher","first-page":"868","DOI":"10.1007\/978-3-642-32009-5_50","article-title":"\u201cFully homomorphic encryption without modulus switching from classical gapsvp,\u201d","author":"Brakerski","year":"2012","journal-title":"Proceedings of the 32nd Annual Cryptology Conference on Advances in Cryptology"},{"key":"B10","first-page":"309","article-title":"\u201c(leveled) fully homomorphic encryption without bootstrapping,\u201d","volume-title":"Proceedings of the 3rd Innovations in Theoretical Computer Science Conference, ITCS '12","author":"Brakerski","year":"2012"},{"key":"B11","article-title":"Leaf: A benchmark for federated settings","author":"Caldas","year":"2019","journal-title":"ArXiv"},{"key":"B12","doi-asserted-by":"crossref","first-page":"125","DOI":"10.1145\/3369583.3392686","article-title":"\u201cTifl: a tier-based federated learning system,\u201d","volume-title":"Proceedings of the 29th international symposium on high-performance parallel and distributed computing","author":"Chai","year":"2020"},{"key":"B13","doi-asserted-by":"crossref","first-page":"409","DOI":"10.1007\/978-3-319-70694-8_15","article-title":"\u201cHomomorphic encryption for arithmetic of approximate numbers,\u201d","volume-title":"Advances in Cryptology-ASIACRYPT 2017: 23rd International Conference on the Theory and Applications of Cryptology and Information Security, Hong Kong, China, December 3-7, 2017, Proceedings, Part I 23","author":"Cheon","year":"2017"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1910.02578","article-title":"Differential privacy-enabled federated learning for sensitive health data","author":"Choudhury","year":"2019","journal-title":"arXiv"},{"key":"B15","doi-asserted-by":"publisher","first-page":"40","DOI":"10.1109\/MM.2023.3260186","article-title":"The amd 400-g adaptive smartnic system on chip: a technology preview","volume":"43","author":"Dastidar","year":"2023","journal-title":"IEEE Micro"},{"key":"B16","doi-asserted-by":"publisher","first-page":"617","DOI":"10.1007\/978-3-662-46800-5_24","article-title":"\u201cFhew: bootstrapping homomorphic encryption in less than a second,\u201d","author":"Ducas","year":"2015"},{"key":"B17","doi-asserted-by":"publisher","first-page":"94","DOI":"10.3390\/fi13040094","article-title":"Privacy preserving machine learning with homomorphic encryption and federated learning","volume":"13","author":"Fang","year":"2021","journal-title":"Future Internet"},{"key":"B18","first-page":"51","article-title":"\u201cAzure accelerated networking: smartnics in the public cloud,\u201d","volume-title":"Proceedings of the 15th USENIX Conference on Networked Systems Design and Implementation, NSDI'18","author":"Firestone","year":"2018"},{"key":"B19","doi-asserted-by":"publisher","first-page":"338","DOI":"10.1049\/icp.2022.0366","article-title":"\u201cHandling tactful data in cloud using pkg encryption technique,\u201d","author":"Gajarla","year":"2021","journal-title":"4th Smart Cities Symposium (SCS 2021), Vol"},{"key":"B20","first-page":"829","article-title":"In-network aggregation for shared machine learning clusters","volume":"3","author":"Gebara","year":"2021","journal-title":"Proc. Mach. Learn. Syst"},{"key":"B21","first-page":"16937","article-title":"\u201cInverting gradients - how easy is it to break privacy in federated learning?\u201d","volume-title":"Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS'20","author":"Geiping","year":"2020"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1145\/3637528.3671545","article-title":"Fedsecurity: Benchmarking attacks and defenses in federated learning and federated llms","author":"Han","year":"2024","journal-title":"ArXiv"},{"key":"B23","doi-asserted-by":"crossref","first-page":"10011","DOI":"10.1109\/CVPR52688.2022.00978","article-title":"\u201cGradvit: gradient inversion of vision transformersm,\u201d","volume-title":"2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Hatamizadeh","year":"2022"},{"key":"B24","article-title":"Fedml: A research library and benchmark for federated machine learning","author":"He","year":"2020","journal-title":"ArXiv"},{"key":"B25","doi-asserted-by":"crossref","first-page":"11","DOI":"10.1145\/3474366.3486923","article-title":"\u201cPyfhel: python for homomorphic encryption libraries,\u201d","volume-title":"Proceedings of the 9th on Workshop on Encrypted Computing &Applied Homomorphic Cryptography, WAHC '21","author":"Ibarrondo","year":"2021"},{"key":"B26","unstructured":"Tensorflow federated2020"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2109.00675","article-title":"Flashe: additively symmetric homomorphic encryption for cross-silo federated learning","author":"Jiang","year":"2021","journal-title":"arXiv"},{"key":"B28","article-title":"Fedml-he: An efficient homomorphic-encryption-based privacy-preserving federated learning system","author":"Jin","year":"2023","journal-title":"ArXiv"},{"key":"B29","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/SVCC56964.2023.10165427","article-title":"\u201cBlocknic: smartnic assisted blockchain,\u201d","volume-title":"2023 Silicon Valley Cybersecurity Conference (SVCC)","author":"Kapoor","year":"2023"},{"key":"B30","doi-asserted-by":"publisher","first-page":"7754765","DOI":"10.1155\/2023\/7754765","article-title":"Preserving resource handiness and exigency-based migration algorithm (PRH-EM) for energy efficient federated cloud management systems","volume":"2023","author":"Karthikeyan","year":"2023","journal-title":"Mob. Inf. Syst"},{"key":"B31","first-page":"1097","article-title":"\u201cImagenet classification with deep convolutional neural networks,\u201d","author":"Krizhevsky","year":"2012","journal-title":"Advances in Neural Information Processing Systems, Vol. 25"},{"key":"B32","first-page":"741","article-title":"\u201c$ATP$: in-network aggregation for multi-tenant learning,\u201d","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Lao","year":"2021"},{"key":"B33","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/HPEC55821.2022.9926325","article-title":"\u201cProcessing particle data flows with smartnics,\u201d","volume-title":"2022 IEEE High Performance Extreme Computing Conference (HPEC)","author":"Liu","year":"2022"},{"key":"B34","article-title":"Ibm federated learning: an enterprise framework white paper v0.1","author":"Ludwig","year":"2020","journal-title":"ArXiv"},{"key":"B35","first-page":"1273","article-title":"\u201cCommunication-efficient learning of deep networks from decentralized data,\u201d","volume-title":"Artificial intelligence and statistics","author":"McMahan","year":"2017"},{"key":"B36","doi-asserted-by":"crossref","first-page":"142","DOI":"10.1109\/HOST45689.2020.9300288","article-title":"\u201cCPU and GPU accelerated fully homomorphic encryption,\u201d","volume-title":"2020 IEEE International Symposium on Hardware Oriented Security and Trust (HOST)","author":"Morshed","year":"2020"},{"key":"B37","unstructured":"Nvidia doca software framework2023"},{"key":"B38","unstructured":"NVDIA Bluefield Networking Platform2024"},{"key":"B39","doi-asserted-by":"publisher","first-page":"84168","DOI":"10.1109\/ACCESS.2023.3265583","article-title":"Homomorphic encryption on gpu","volume":"11","author":"\u00d6zcan","year":"2023","journal-title":"IEEE Access"},{"key":"B40","doi-asserted-by":"publisher","first-page":"223","DOI":"10.1007\/3-540-48910-X_16","article-title":"\u201cPublic-key cryptosystems based on composite degree residuosity classes,\u201d","author":"Paillier","year":"1999","journal-title":"Advances in Cryptology"},{"key":"B41","doi-asserted-by":"publisher","first-page":"840","DOI":"10.5555\/3540261.3540326","article-title":"Sageflow: robust federated learning against both stragglers and adversaries","volume":"34","author":"Park","year":"2021","journal-title":"Adv. Neural Inf. Process. Syst"},{"key":"B42","article-title":"Pytorch: An imperative style, high-performance deep learning library","author":"Paszke","year":"2019","journal-title":"ArXiv"},{"key":"B43","doi-asserted-by":"crossref","first-page":"200","DOI":"10.1109\/BCCA58897.2023.10338942","article-title":"\u201cSmartnic-powered multi-threaded proof of work,\u201d","volume-title":"2023 Fifth International Conference on Blockchain Computing and Applications (BCCA)","author":"Patel","year":"2023"},{"key":"B44","first-page":"2825","article-title":"Scikit-learn: machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res"},{"key":"B45","article-title":"Adaptive federated optimization","author":"Reddi","year":"2021","journal-title":"ArXiv"},{"key":"B46","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/WF-IoT54382.2022.10152214","article-title":"\u201cOn the feasibility of homomorphic encryption for internet of things,\u201d","volume-title":"2022 IEEE 8th World Forum on Internet of Things (WF-IoT)","author":"Reddy","year":"2022"},{"key":"B47","first-page":"785","article-title":"\u201cScaling distributed machine learning with {In-Network} aggregation,\u201d","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Sapio","year":"2021"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.1109\/CANDARW60564.2023.00035","article-title":"Performance improvement of federated learning server using smart NIC","author":"Shibahara","year":"2023","journal-title":"ArXiv"},{"key":"B49","doi-asserted-by":"publisher","first-page":"100406","DOI":"10.1016\/j.eij.2023.100406","article-title":"Federated learning optimization: a computational blockchain process with offloading analysis to enhance security","volume":"24","author":"Shitharth","year":"2023","journal-title":"Egypt. Inf. J"},{"key":"B50","doi-asserted-by":"publisher","first-page":"694","DOI":"10.48550\/arXiv.2109.14236","article-title":"Lightsecagg: a lightweight and versatile design for secure aggregation in federated learning","volume":"4","author":"So","year":"2022","journal-title":"Proc. Mach. Learn. Syst"},{"key":"B51","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3338501.3357370","article-title":"\u201cA hybrid approach to privacy-preserving federated learning,\u201d","volume-title":"Proceedings of the 12th ACM workshop on artificial intelligence and security","author":"Truex","year":"2019"},{"key":"B52","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1812.00535","article-title":"Beyond inferring class representatives: user-level privacy leakage from federated learning","author":"Wang","year":"2018","journal-title":"arXiv"},{"key":"B53","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1155\/2022\/6759875","article-title":"Fogdedupe: a fog-centric deduplication approach using multi-key homomorphic encryption technique","volume":"2022","author":"Yoosuf","year":"2022","journal-title":"J. Sens"},{"key":"B54","first-page":"81","article-title":"\u201cA smartnic-based secure aggregation scheme for federated learning,\u201d","volume-title":"The 3rd International Conference on Computer Engineering and Intelligent Control","author":"Zang","year":"2022"},{"key":"B55","first-page":"493","article-title":"\u201c{BatchCrypt}: efficient homomorphic encryption for {Cross-Silo} federated learning,\u201d","volume-title":"2020 USENIX annual technical conference (USENIX ATC 20)","author":"Zhang","year":"2020"},{"key":"B56","first-page":"14774","article-title":"\u201cDeep leakage from gradients,\u201d","volume-title":"Advances in Neural Information Processing Systems, Vol. 32","author":"Zhu","year":"2019"}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2024.1465352\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,21]],"date-time":"2024-10-21T12:12:32Z","timestamp":1729512752000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2024.1465352\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,21]]},"references-count":56,"alternative-id":["10.3389\/fcomp.2024.1465352"],"URL":"https:\/\/doi.org\/10.3389\/fcomp.2024.1465352","relation":{},"ISSN":["2624-9898"],"issn-type":[{"value":"2624-9898","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,10,21]]},"article-number":"1465352"}}