Estimated reading time: 7 minutes
By Daphne McCanse, Strategic Research Communications Consultant
Many faculty, researchers, and students desire to use cloud resources but often lack funding to access them.
Here is what they may not know: There are federally funded options available at no cost.
The National Science Foundation and its partners support a broad set of cloud and cloud-like resources for research and education, including options for accessing commercial cloud resources. From virtual machines and classroom coding notebooks to large GPU workloads and network experiments, all of the resources below are available at no cost for eligible U.S. use cases.
Check out what each resource does, who qualifies, and where to start, so campus IT and research computing teams can point faculty to the right fit.
Resources for Specific Use Cases
What is it: Jetstream2 is a set of NSF-funded cloud services offering on-demand virtual machines, GPUs, Jupyter Notebooks, and familiar cloud-style workflows without a commercial bill.
Who’s it for: Researchers who need flexible, interactive computing, and instructors running courses, workshops, and science gateways.
Where to start: Request Jetstream2 resources as part of an NSF ACCESS or National AI Research Resource (NAIRR) allocation.
What is it: The OSPool is a national high-throughput computing resource supported by NSF funding and contributed capacity from campus computing clusters across the country. It is a highly scalable option for work that spans across many independent tasks, and for software without license restrictions on the number of simultaneous runs.
Who’s it for: Researchers and educators affiliated with projects at U.S.-based academic, government, or nonprofit institutions. It is especially useful for those whose work includes many independent tasks (AI inference runs, parameter sweeps, simulations, and image or genomic/text analysis).
Where to start: Register through the OSG Portal. A research computing facilitator will follow up to arrange a short consultation and personalized start-up guidance. Though not required, OSPool resources may also be requested as part of an NSF ACCESS or NAIRR resource allocation (OSG’s OSPool is listed under the Partnership to Advance Throughput Computing).
What is it: NRP is a distributed, Kubernetes-based platform built around the Nautilus cluster. NRP provides free access to GPUs, storage, and large language models for U.S. researchers.
Who’s it for: Researchers comfortable with containers, especially for AI and machine learning work. Browser-based development and hosted JupyterHub environments lower the barrier for newer users.
Where to start: Create a free account through the NRP Nautilus portal. Then join an existing namespace or request one before deploying workloads. Optionally, NRP resources may also be requested as part of an NSF ACCESS or NAIRR resource allocation.
What is it: OSN is an NSF-funded, distributed storage system. It offers S3-compatible object storage for sharing and moving research data between collaborators and endpoints.
Who’s it for: Researchers and projects that need to store, publish, or exchange medium-to- large-scale datasets with collaborators.
Where to start: Request OSN storage as part of an NSF ACCESS allocation.
What is it: CloudBank is an NSF-funded service that provides no-cost allocations for major commercial cloud platforms (AWS, Azure, IBM, and Google Cloud), as well as tools and services for account setup, cost monitoring, workload management, training, and consulting.
Who’s it for: U.S. researchers who want to use commercial cloud resources. NSF awardees with commercial cloud expenses in their budgets can also use CloudBank to manage their spending.
Where to start: Request CloudBank resources as part of an NSF ACCESS or NAIRR allocation. See CloudBank’s Get Access page for instructions for researchers and educators.
What is it: Chameleon is an NSF-funded, reconfigurable testbed for computer science systems research and teaching. It offers reservable bare-metal servers with root and console access, a range of hardware (including CPUs, GPUs, and FPGAs) and architectures, and configurable network and storage hierarchies.
Who’s it for: Computer systems researchers and educators who need deep control of hardware configuration, operating systems, networks, cloud and edge testing, power management, or security. Faculty and research staff at U.S. institutions can lead projects. Students join a project led by a research advisor or instructor.
Where to start: Create a free account by signing in with your institution’s credentials at the Chameleon portal. Then request principal investigator (PI) eligibility and create a project, or ask to join an existing one. Chameleon is not available through ACCESS or NAIRR; it has its own allocation process.
What is it: FABRIC is an NSF-funded, nationwide programmable network testbed with compute, storage, and high-speed links that users configure themselves.
Who’s it for: Researchers and educators building or studying computer networks, distributed systems, and cybersecurity, including instructors who want hands-on lab environments.
Where to start: Join a project led by a faculty PI, or create one if you are faculty. Optionally, FABRIC resources may also be requested as part of an NSF ACCESS resource allocation.
Programs That Connect Researchers to Multiple Resources