Technology Powers University-Industry Research Partnerships

Georgia Tech has built a robust IT infrastructure to make the research possible, including high-speed networking and strong cybersecurity inside the facility. Outside it, researchers use the public cloud and the university’s high-performance computing (HPC) systems. 

“The goal is to engage industry in applied research, support their innovation efforts and create meaningful collaboration opportunities,” Ferguson says.

Georgia Tech is part of a growing trend in higher education in which universities are deepening their research partnerships with private industry. The federal funding landscape has shifted in recent years, forcing universities to adjust. 

“There are changes in what the federal government is interested in,” says Josh Tullis, Georgia Tech’s senior director of corporate engagement. “That has required some pivoting of research to align with national priorities.” 

Some research partnerships are between companies and universities. Others are tied to federal funding. The National Science Foundation, wanting research to produce impactful results, is funding more projects in which companies invest alongside the government and universities apply for the funding, says Anthony Boccanfuso, president and CEO of UIDP, a nonprofit that strengthens university-industry research partnerships. 

RELATED: Smart cities research projects help institutions optimize transportation.

Industry accounts for only a small share of university research funds, but the value goes beyond money, says Zachary Holman, Arizona State University’s vice dean for research and innovation. 

For universities, the partnerships keep research focused on real-world problems, offer researchers access to specialized equipment and give students hands-on experience that can lead to job offers, Holman says.

For companies, these relationships offer a way to explore long-term problems they lack the time or resources to pursue on their own, says Ratul Mahajan, a professor of computer science and engineering at the University of Washington.

The result is a “beautiful, virtuous cycle,” Tullis says. “The government is supporting the academic work. Industry is supporting the application of it, which then inspires the next generation of academic work.”

Georgia Tech Lab Shapes the Future of Manufacturing

At any given time, Georgia Tech in Atlanta has more than 1,000 industry-sponsored research projects underway with hundreds of companies, says Mark Nolan, the university’s associate vice president of corporate engagement.

DISCOVER: Modern tools enable collaboration for disparate teams.

AMPF, which opened in 2017 with an initial $3 million gift from Delta Air Lines Foundation, recently tripled in size to 60,000 square feet through a $40 million expansion backed by federal, state and university funding. 

Today, the shop floor features industrial equipment that includes metal 3D printers, robotic welding systems and autonomous, Wi-Fi-connected robots that move materials between machines.

Researchers design experiments, program systems and interpret resulting data, while robots can now handle much of the physical grunt work, Ferguson says. Georgia Tech integrates the operational technology to IT systems to create what he calls a “software-defined factory.” 

The facility’s machines generate huge volumes of data, from equipment telemetry to high-speed thermal and optical camera footage. The data streams into cloud storage for analysis, primarily on Amazon Web Services (AWS), with some on Microsoft Azure, he says.

From the cloud, researchers push experiments to lab equipment and analyze results using custom Python code. They also take advantage of PACE, the on-campus HPC center, and the new AI Makerspace supercomputing hub for compute-intensive work, such as training AI models or analyzing terabytes of data, Ferguson says.

Because some processes generate gigabits of data per second, Georgia Tech recently modernized AMPF’s network with 11 HPE Juniper switches and nine Juniper Mist wireless access points. Researchers, who work on personal or university-owned laptops, also use computer-aided design tools such as Autodesk Fusion to design parts before production. 

To support collaboration, Georgia Tech is equipping AMPF’s meeting rooms and a 150-person event space with Cisco Collaboration audiovisual equipment, including displays and cameras for Cisco Webex, Microsoft Teams, Zoom and Google Meet videoconferencing, Ferguson says. 

Security is also paramount. Georgia Tech relies on card- and digital-based physical access, Palo Alto Networks next-generation firewalls and role-based access controls tied to the university’s identity management system with single sign-on and multifactor authentication, he says. Devices must stay current with security updates. For sensitive projects, data lives in secure enclaves. 

“Research security is foundational to what we do,” Ferguson says. “Our partners trust us to protect their data and the knowledge that we create together.” 

LEARN MORE: University researchers rely on hybrid infrastructure.

Arizona State University Collaborates on Chip Research

Last October, ASU and semiconductor company Applied Materials opened the Materials-to-Fab Center, a shared $270 million research, development and prototyping facility housed within MacroTechnology Works, the university’s semiconductor research facility in Tempe, Ariz.

Applied Materials, a tenant at MacroTechnology Works, shares the MTF space with ASU researchers. The company supplied most of the advanced chipmaking equipment, while ASU provided the space and invested in facilities upgrades, Holman says.

While the space is available to other companies, government entities and academic institutions, ASU and Applied Materials are collaborating on more than a dozen research projects, he says. 

Collaboration takes different forms. In some cases, Applied Materials researchers work onsite with ASU researchers. In others, they hold regular calls to monitor progress and guide ASU researchers. They sometimes exchange samples for testing. 

Some research runs on laptops. Other research uses shared ASU core facilities, such as the university’s research computing center, which houses two supercomputers. Collaboration and file-sharing tools are not standardized.

“It’s kind of the Wild West,” Holman says. “The university makes tools available to us, but I will regularly be on Zoom meetings. We also have Teams. We have Dropbox subscriptions, but half the people use SharePoint and others use Box.” 

The university typically adapts to its industry partners’ software standards. “In many cases, it’s easier for the university to adapt to the norm of the company than the other way around,” he says. 

Security matters too. ASU tailors its protections to each project, such as limiting data access, and for the most sensitive work, restricting internet or USB access, Holman says.

University of Washington Tackles the Future of Cloud Infrastructure

In Seattle, the University of Washington launched the UW Center for the Future of Cloud Infrastructure with industry partners Alibaba, Amazon, Cisco, Google, Microsoft and VMware, who helped fund the effort.

The center is winding down its work. But over four years, faculty and students worked on next-generation networking, energy management and network verification, which validates network configurations and behavior. Some ideas have been deployed in partners’ production systems, says Mahajan, FOCI’s co-director. 

UP NEXT: Modern research requires a modern infrastructure.

All partners met with UW researchers to provide guidance and feedback, and some collaborated deeply and even co-authored papers with UW. 

Researchers used partner cloud platforms, such as AWS, Google Cloud and Azure, as well as CloudLab, an NSF-funded research cloud. For collaboration, they used Zoom, Google Meet and Google Docs, he says. 

Most of UW’s research is open: Code goes on GitHub, while papers are published in peer-reviewed journals. But the university would rather not manage industry data. Instead of bringing a partner’s data onto campus, UW sends its researchers and code to the company, where the data lives.

“If industry data is involved, we tend to not take ownership of the data. It’s a big risk to them and to us,” Mahajan says. “We can ship code, or students can go on an internship and test out ideas there.” 

Moving forward, Mahajan and his colleagues are pursuing a new research effort on self-defining systems, using AI agents to build and operate computing systems from design to daily operation. 

“We are in early conversations — what the shape should be, and which partners to approach,” he says.