NCShare

Creating a Science DMZ

Purpose

A Science DMZ is a dedicated portion of a site or campus network, located as close to the network perimeter as possible, whose equipment, configuration and security policy are tailored to serve high-performance science applications. This initiative will establish a Science DMZ for five to seven institutions so campus researchers can access external sites with a speed and style of connectivity that is typically only available at large research universities.

---

Goals

Building a shared, regionally based network operating on MCNC’s existing state-wide research and education network aims to:

• Lower costs for universities and require fewer campus support personnel
• Provide fast and unrestricted data movement to multiple institutions
• Increase accessibility of high-speed, data-driven research by democratizing access to advanced cyberinfrastructure
• Enhance research productivity and collaboration
• Reduce the time required for scientific discoveries at participating minority-serving and smaller institutions

---

---

Investigators

• Tracy Futhey

  Vice President for Technology and Chief Information Officer, Duke University, Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Joel Faison

  Chief Information Officer, North Carolina Central University, Co-Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Kevin Davis

  Chief Information Officer, Davidson College, Co-Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Tracy Doaks

  President and Chief Executive Officer, MCNC, Co-Principal Investigator-sDMZ

• Jon Gant

  Interim Associate Vice Chancellor for Faculty Development and Resources, North Carolina Central University, Co-Principal Investigator-sDMZ

Compute as a Service

Purpose

This collaboration among Duke University, North Carolina Central University (NCCU), and Davidson College builds a shared computing environment housed at North Carolina’s research and education network provider called MCNC. The shared computing environment supports a common set of software and science drivers, with priority use by Davidson, NCCU, and other North Carolina minority-serving and smaller institutions; OSG and Duke would have lower-priority use, if excess-capacity exists. As a result, faculty at participating institutions, who often juggle high course loads, have access to tools and services that ease the delivery of customized computing systems to meet their research needs and enable easy access to powerful tools for students.

---

Goals

The goal of this project is to provide a shared computing environment in a centrally managed, efficient and flexible fashion, that can support common, as well as individual, needs for faculty and researchers at the two pilot schools, and then extend that environment and support techniques to others within the region.

---

---

Investigators

• Charley Kneifel

  Senior Technical Director, Duke University, Principal Investigator-CaaS

• Tracy Futhey

  Vice President for Technology and Chief Information Officer, Duke University, Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Joel Faison

  Chief Information Officer, North Carolina Central University, Co-Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Kevin Davis

  Chief Information Officer, Davidson College, Co-Principal Investigator-sDMZ and Co-Principal Investigator-CaaS

• Mohammad Ahmed

  College of Health and Sciences and Associate Director of Triangle Universities Nuclear Laboratory (TUNL), North Carolina Central University, Co-Principal Investigator-CaaS

GPU as a Service

Purpose

NCShare AI-GaaS addresses two extremes on the demand spectrum for GPUs: First, GPUs are often unavailable or unaffordable to institutions with modest needs and budgets; and second, Artificial Intelligence research, including Large Language Model research, often requires GPUs at a scale even the largest research universities struggle to attain. By moving GPU sharing from the individual campus level to the state level the result is more efficient GPU utilization, greater aggregate computational capacity for solving research problems, improved economies of scale, and democratized access for smaller schools and under-resourced minority-serving institutions.

---

Goals

• Automated provisioning, containerization, GPU partitioning, advanced job scheduling, and federated login via either InCommon or bilateral federation with NCShare
• Horizontally scalable, shared environment of 32 H100s, with R1s and others able to buy-in with additional nodes
• Ensure sustainability /cost-effectiveness via a shared service from NC’s mature network operator (MCNC)

---

---

Investigators

• Tracy Futhey

  Vice President for Technology and Chief Information Officer, Duke University, Principal Investigator-sDMZ and Co-Principal Investigator-CaaS and Principal Investigator – AI-GaaS

• Charley Kneifel

  Senior Technical Director, Duke University, Principal Investigator-CaaS and Co-Principal Investigator – AI-GaaS

• Tracy Doaks

  President and Chief Executive Officer, MCNC, Co-Principal Investigator-sDMZ and Co-Principal Investigator – AI-GaaS

• J. Michael Barker

  Vice Chancellor for Information Technology and Chief Information Office, University of North Carolina at Chapel Hill, Co-Principal Investigator – AI-GaaS

• Kaushik Roy

  Professor and Chair in the Department of Computer Science at North Carolina Agricultural & Technical University, Co-Principal Investigator – AI-GaaS