EAGER: Accelerating Science Across Borders with Next-Generation SmartNICs/DPUs

This project enhances international exploratory research by strategically deploying next-generation SmartNICs, specifically Data Processing Units (DPUs), at four existing international FABRIC sites, which builds upon FABRIC Across Borders (FAB). DPUs are an emerging technology that places programmable, high-performance processors directly on the network interface, which can accelerate data processing by orders of magnitude by offloading data-centric tasks like networking, storage, and security from traditional CPUs and GPUs. DPUs also reduce time-to-insight for data-intensive scientific experiments across various domains and offer strong workload isolation for researchers, enabling secure networking and encrypted storage experimentation. By integrating SmartNICs into the FABRIC ecosystem, this project will accelerate exploratory data-intensive scientific research across multiple fields, including high-energy physics, federated learning, and wireless networking between the USA and the international sites. Scientific discovery in fields such as high-energy physics, smart cities, and AI-driven applications increasingly depends on the rapid movement and analysis of vast amounts of data. Traditional networking approaches often limit the pace of discovery, as server CPUs become bottlenecks for processing high-volume data streams. SmartNICs—programmable network cards equipped with dedicated processors—can offload and accelerate tasks such as encryption, data filtering, storage management, and machine learning inference. By deploying SmartNICs on FABRIC's international sites, new classes of experiments for exploratory research could be pursued. The deployment of SmartNICs across FABRIC's international nodes will broaden access to advanced networking technologies for a global community of researchers, including students and early-career scientists. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.