San Diego State University has received $5 million from the U.S. National Science Foundation Office of Advanced Cyberinfrastructure to help build a national system that lets researchers compute on sensitive data without ever exposing it. The award funds a five-year project called PARAPET, short for Prototype Architecture for Research Advances using Privacy Enhancing Technologies, which will give scientists in healthcare, finance, internet security and the social sciences a shared testbed for working with medical records, financial information and census statistics while keeping that data locked down.
The cooperative agreement, numbered 2537035, is listed as ongoing and has a total value of $4,999,963, according to federal grant records reviewed by HigherGov. San Diego State partnered with Arizona State University and the University of Utah on the project, as detailed in a release from EurekAlert!. NSF solicitation language describes a Category II award track for deploying forward-looking computing prototypes and testbeds for national research communities rather than funding standard high-performance hardware purchases.
How PARAPET Is Supposed to Work
PARAPET will lean on three technical approaches — differential privacy, homomorphic encryption and federated learning — to let approved users on secured networks work directly on encrypted data without decrypting it first. The system is designed to train AI models without exposing the private data underneath them, using local model updates instead of raw data to build out a larger model, and it's meant to protect information even if an individual device is compromised. Robert Beverly, director of SDSU's Cybersecurity Center for Research and Education and a professor of computer science, said privacy barriers currently impede researchers' ability to work together on data, according to the EurekAlert! release. SDSU assistant professor of computer science Joann Chen added that an attacker who managed to get hold of encrypted data should be unable to do much with it.
The project is also built with an eye toward the cryptographic threats of the future. PARAPET could help protect against future quantum computing attacks, a concern that gained federal urgency after the National Institute of Standards and Technology finalized its first three post-quantum cryptography standards, FIPS 203, 204 and 205, in August 2024 to guard against the kind of attacks a sufficiently powerful quantum computer running Shor's algorithm could unleash on classical encryption like RSA, according to NIST.
A Leadership Team With Federal and Institutional Reach
The team overseeing PARAPET pairs deep knowledge of the funding agency itself with the compliance networks needed to get universities on board. Manish Parashar, who will help manage PARAPET's hardware, is Chief AI Officer for the University of Utah and executive director of its Scientific Computing and Imaging Institute — and previously served as Office Director for NSF's own Office of Advanced Cyberinfrastructure from 2018 to 2023, according to his Wikipedia entry.
Carolyn Ellis, director of Arizona State University's Research Technology Office and a project co-principal investigator, will ensure PARAPET meets evolving regulatory and policy requirements and will lead efforts to recruit users nationwide to test the system, per the EurekAlert! release. Ellis co-founded the Regulated Research Community of Practice in December 2021, an NSF-supported network connecting more than 350 higher education institutions on data compliance issues, according to University of Florida News, which notes the network helps academic research programs implement frameworks like NIST 800-171 cost-effectively. Arizona State will contribute that institutional perspective through the same network, and Ellis said this month that deploying privacy-enhancing technologies requires evaluating how the underlying cryptographic capabilities fit within institutional policies, research workflows and formal data-sharing agreements, according to ASU News.
Where the Servers Will Actually Live
Mike Farley, SDSU's chief technology research officer, oversees the on-campus data center where PARAPET's servers will be housed. That physical infrastructure sits inside a broader institutional push: SDSU's Cybersecurity Center for Research and Education, which Beverly directs, already spans six colleges and administrative divisions across the university, connecting computer science, engineering, business, homeland security and IT leadership, according to San Diego State University. The university says it intends to build on those existing interdisciplinary cyber initiatives and its cybersecurity and data privacy curricula as PARAPET gets underway.
Participating institutions are expected to retain greater control over their own sensitive research data once the system is running, and the broader regulatory backdrop helps explain why. Federal research security guidelines under National Security Presidential Memorandum 33, finalized by the White House in July 2024, form part of that regulatory backdrop.
A Fast-Growing Market for Privacy Tools
The push toward tools like PARAPET is also being driven by commercial growth in privacy-preserving computation. Privacy-enhancing technologies are being developed for healthcare, finance and telecommunications.
Whatever the market numbers ultimately show, PARAPET's stated purpose is narrower and more immediate: give researchers across the country — not just at SDSU, ASU and Utah — a shared, secure place to analyze sensitive data together. The project is funded through NSF grant #2537035 and is expected to run for five years, according to the EurekAlert! release, with the testbed intended to serve as a national-scale resource for privacy-enhancing technology research well beyond the three founding campuses.
