Velocity • Durability

NMSU Research Cores Program data centre
VDURA
New Mexico State University
Resources / Success Story

VDURA powers AI and HPC research at New Mexico State University.

A 6PB VDURA Data Platform deployment anchors high-performance computing for NMSU's Research Cores Program, live in production August 2026.

6 PB usable capacity Carnegie R1 institution $141M+ annual research Existing InfiniBand fabric
6 PB
Usable capacity
R1
Carnegie R1 institution
2026
Live in production
Overview

One of America's newest R1 universities picked its data foundation.

New Mexico State University (NMSU), a Carnegie R1 research institution supporting more than $141 million in annual research expenditures, selected the VDURA Data Platform to anchor high-performance computing within its Research Cores Program, the university's network of shared research facilities, instrumentation and computational resources.

In production as of August 2026, the deployment gives NMSU researchers a high-performance, highly durable data foundation for workloads spanning aerospace, agriculture, cybersecurity, water science and biomedical research. The installation builds on the strategic partnership VDURA and NMSU announced in 2025 to co-develop post-quantum cryptographic technology for AI and HPC data infrastructure.

The Challenge

Research demand grew faster than the infrastructure under it.

As NMSU's research profile expanded following its R1 designation, its computing environment faced growing demands:

  • Diverse mixed I/O workloads across AI training, simulation and data-intensive science competing for shared resources
  • Growing datasets requiring both high-performance access and economical long-term capacity
  • The need for infrastructure that could serve a broad, multi-disciplinary research community through the Research Cores Program
The Deployment

Flash and HDD under one control plane, on the fabric NMSU already runs.

NMSU deployed the VDURA Data Platform in a Mixed-Fleet configuration combining NVMe flash and high-density HDD capacity under a single control plane.

Environment
High-performance computing within NMSU's Research Cores Program
Usable capacity
6 PB
Architecture
Three VeLO director nodes for metadata and orchestration, plus three Mixed-Fleet storage nodes, each combining 12 SSDs for low-latency active data with 78 high-density HDDs for bulk research data
Fabric
Data traffic runs over the university's InfiniBand fabric, shared with compute clients
Namespace
One unified global namespace, a single control plane spanning flash and HDD tiers
VDURA Data Platform deployment at NMSU
The VDURA Data Platform deployment at NMSU's Research Cores Program facility
Partnering with the NMSU Vision Team and the Research Cores Program allows us to deliver a world-class VDURA data storage solution to campus. This facility will ensure NMSU researchers continue to push boundaries in areas vital to New Mexico's growth and economic future.
Amy Wagler
Director of Research Computing and Data Science, New Mexico State University
Results and Benefits

What the platform delivers for NMSU researchers.

Performance. Flash-level responsiveness for active research data
Efficiency. Mixed-Fleet economics deliver bulk capacity at a fraction of all-flash cost
Simplicity. A single namespace across tiers eliminates separate performance and capacity silos
Scalability. Capacity and performance grow with research demand, no forklift upgrades
Durability. Production-hardened platform built for environments where failure is not an option
Conclusion

The VDURA Data Platform gives NMSU a future-ready data foundation for AI and HPC research, combining the velocity researchers demand with the durability an R1 institution requires. As one of America's newest R1 universities, NMSU now has storage infrastructure built to keep pace with its ambitions.

Why Choose VDURA

Performance where research needs it. Economics everywhere else.

Performance

Research-ready performance

Flash-level responsiveness for active datasets keeps AI training and simulation pipelines moving without storage bottlenecks.

Economics

Mixed-Fleet economics

NVMe flash and high-density HDD capacity managed as one namespace, delivering performance where it's needed and economical bulk capacity everywhere else.

Integration

Seamless fabric integration

The data platform runs over the university's existing InfiniBand fabric, shared with compute clients, with no separate storage network to build or manage.

Growth

Scale without re-architecture

Software-defined architecture and a subscription model let capacity and performance grow as research demand grows.

VDURA For Research

Where to go deeper.

Tell us what your researchers need to run.

We will show you what performance, capacity and economics look like on your environment.