How HPC Systems Are Making Things Easier

Why It Matters

The CCC demonstrates how research institutions can achieve cloud agility and collaborative speed without exposing sensitive data to public clouds, accelerating discovery and controlling long‑term costs.

Key Takeaways

• $2M NIH S10 grant funds Moffitt’s Collaborative Computing Center.
• CCC offers 30 nodes, 1.3 PB high‑speed storage, dedicated internet.
• Private, multi‑tenant HPC mimics public cloud flexibility on‑premises.
• Secure environment streamlines external collaborator onboarding and data sharing.
• Plans include on‑site AI/ML inference for foundation models.

Pulse Analysis

The rise of private, cloud‑like HPC platforms reflects a broader industry shift toward balancing computational power with stringent data‑privacy requirements. Traditional on‑prem clusters often sit behind rigid institutional IT policies, creating bottlenecks for external collaborators and slowing data exchange. Moffitt’s CCC tackles these pain points by establishing a demilitarized zone (DMZ) that isolates research workloads, provides a dedicated high‑bandwidth connection, and leverages petabyte‑scale storage. This architecture not only reduces administrative overhead but also aligns with compliance mandates for protected health information, positioning the center as a model for secure, collaborative science.

From a cost and performance perspective, the CCC offers a compelling alternative to public‑cloud services. By provisioning a fixed pool of nodes and storage in‑house, Moffitt gains predictable long‑term expenditures while retaining the elasticity typically associated with cloud environments. Multi‑tenant capabilities enable simultaneous projects without sacrificing security, and the on‑prem setup eliminates data egress fees and latency associated with remote cloud resources. The system’s design also allows seamless integration with existing tools such as Globus and Open OnDemand, further streamlining researcher workflows.

Looking ahead, the center’s focus on on‑site AI and machine‑learning inference underscores the strategic value of keeping large language and vision models close to the data they process. Running foundation models within the secure perimeter minimizes PHI exposure and sidesteps the regulatory complexities of cloud‑based AI services. As more institutions adopt similar private‑cloud HPC models, we can expect accelerated adoption of advanced analytics, faster collaborative breakthroughs, and a reshaping of how biomedical research balances innovation with compliance.