Over the past year, UMass Amherst Information Technology has dramatically scaled its high-performance research capabilities through NET2, a high-capacity network built to support data-intensive, collaborative science. In early 2025, the network was capable of moving data at speeds of up to 400G per second. By the end of the year, that total had doubled to 800G.
This capability was showcased on a national stage at SC25, the world’s leading supercomputing conference, where UMass collaborated with partners to demonstrate how ultra-high-speed networks have the power to transform data-intensive science.
Led by Rafael Coelho Lopes de Sá, a principal investigator in UMass Amherst’s Department of Physics, the NET2 demonstration showed how massive datasets can be transferred between institutions in real time. As a collaborator on CERN’s ATLAS project, Lopes de Sá’s work involves moving staggering volumes of data generated by record-breaking experiments with the Large Hadron Collider, the world’s largest particle accelerator.
Researchers involved in ATLAS analyze collision data to investigate fundamental questions about the nature of matter, energy, and the forces that govern the universe – work that places extraordinary demands on research networks. This makes high-capacity, low-latency data movement critical.
“UMass is building the kind of networking infrastructure that modern research demands,” said Christopher Misra, Vice Chancellor for Information Technology and CIO. “Our investment in 800G networking is enabling groundbreaking work that is, without exaggeration, seeking to understand the secrets of the universe. There’s no telling what possibilities will continue to open up when researchers aren’t held back by the speed of data.”
UMass Amherst’s work on NET2 is closely connected to the Massachusetts Green High Performance Computing Center (MGHPCC), where NET2 is physically housed and operated. MGHPCC is a shared facility serving universities, government, and industry partners across the commonwealth, enabling a collaborative model that allows advanced research infrastructure to scale beyond any single institution.
That shared approach is what has made demonstrations like SC25 possible. Teams at MGHPCC provided significant investment to support the 800 Gbps network demonstration, with MGHPCC and UMass Amherst together contributing nearly $1 million in networking hardware and resources. This included Juniper networking equipment and shared optical infrastructure developed in partnership with institutions such as Caltech.
“We’re fortunate to work in the type of collaborative environment where this scale of research infrastructure can be built, maintained and demonstrated,” said Misra. “It takes shared expertise and care to make systems like this work at scale, and that work isn’t finished.”
As research continues to generate ever- larger datasets across disciplines and institutions, the demands placed on research networks will only grow. UMass Amherst’s work on NET2 and 800G networking shows what it takes to support science at a global scale, ensuring data movement never becomes the limiting factor in discovery.