FABRIC Testbed

What happens when a promising research idea works perfectly in theory—but encounters the complexity of the real world?

That is where testbeds come in.

A testbed is a controlled environment where researchers can safely experiment, measure results, identify unexpected problems, and refine new technologies before deploying them more broadly. Testbeds help answer questions that simulations alone cannot: What happens when a technology operates at scale? Where are the bottlenecks? Which assumptions break down? And how can the technology be improved?

FABRIC, a national-scale, programmable research infrastructure, gives researchers a place to explore those questions through cutting-edge experimentation at scale. Its capabilities support research across networking, cybersecurity, distributed computing, storage, virtual reality, 5G, machine learning, and science applications.

A Safe Place to Test the Unknown

For Kuang-Ching Wang, Co-PI on FABRIC (Binghamton University) , a testbed is simply a place to test something when you do not yet know the answer.

Wang explains “You can't test some questions in real-world systems because you may break things. Instead, you want to test it in a safe place where you can control the environment, do something, and measure it.”

That ability to experiment, measure results, identify what does not work, and try again is at the heart of why testbeds matter. When an experiment does not work as expected, the result is not a failure; it can reveal what needs to change.

Why Simulation Is Not Always Enough

Computer simulations are an important part of research, but they rely on assumptions. Real-world systems introduce complexity that can be difficult to reproduce in a simulation.

Paul Ruth, PI on FABRIC (RENIC) summed up the challenge with a familiar observation: “In theory, theory and practice are the same, but in practice they're not.”

Researchers can develop an idea and test it in a simulated environment where everything appears to work. But when they move into a real-world environment, unexpected bottlenecks and failures can emerge.

Testbeds provide a way to discover those problems before attempting broader deployment.

For example, researchers working with large scientific instruments may need to move enormous amounts of data across long distances while processing some of that data along the way. FABRIC allows researchers to measure actual performance, identify bottlenecks, and refine their approaches based on what they observe.

A Shared Environment for Discovery

Building a large-scale research environment is difficult and expensive. Without shared infrastructure, researchers may have to build and manage their own systems before they can even begin answering their research questions.

Shared testbeds such as FABRIC change that.

One major benefit is reproducibility. “You publish a paper and say, I did this experiment. These are the machines I used. Here are the results I got,” Ruth explained. “If you want to try it again, you can use the same testbed and verify my results.”

A shared infrastructure gives researchers a common environment for experimentation, allowing them to focus on their research while making it easier for others to reproduce and build upon their work.

That shared environment is also fostering a collaborative community where researchers can share their experiments, learn from one another, and build on each other’s work. Wang said one of the most exciting developments has been seeing researchers share their experiments on FABRIC and contribute to this growing community.

What Makes FABRIC Different?

The everyday internet is designed to provide reliable services to users—not to allow researchers to experiment with how the underlying network works. FABRIC gives researchers a level of visibility and control that is not typically available on commercial networks.

On a regular internet connection, users generally do not know exactly how their traffic is being routed. On FABRIC, researchers can control and observe the infrastructure their experiments use, allowing them to investigate questions about network performance, new architectures, distributed systems, and high-volume data transfers.

FABRIC also connects researchers with resources beyond the testbed itself.

“We also connect to all the supercomputing centers, so if you want to move data from your instrument to your supercomputer, and you want to do something interesting with it in the middle of the network, you can use the FABRIC” Ruth explained.

This connectivity allows researchers to experiment with increasingly realistic workflows that span instruments, networks, computing resources, and other research infrastructure.

From Theory to Experiment and Beyond

The role of a testbed is ultimately to help close the gap between an idea and its real-world application.

FABRIC supports that progression by giving researchers a place to test ideas, identify problems, refine their solutions, and experiment again.

“The idea is to get you from theory to experiment to production,” Ruth said.

That process has played an important role in the history of networking. Wang described how earlier shared networking experiments allowed researchers to deploy technologies across multiple universities, identify problems, improve their solutions, and share what they learned.

The same cycle—experiment, learn, refine, repeat—continues to drive research today.

Building the Foundation for What Comes Next

The questions researchers are asking today are becoming increasingly complex. Artificial intelligence, advanced wireless systems, cybersecurity, high-performance computing, and quantum technologies all require new ways to experiment with systems that are distributed, programmable, and connected across institutions and geographic locations.

FABRIC provides researchers with the infrastructure to explore those questions today while helping prepare for the technologies of tomorrow.

Wang sees particular opportunities ahead as quantum research continues to grow. “I think FABRIC is already connecting a lot of the locations where this quantum research is happening,” he said. “I do see quantum research, one way or the other, will be touching FABRIC, and FABRIC will enable a lot of that research.”

He also emphasized the importance of international collaboration and the opportunities emerging as researchers explore new areas involving AI and quantum technologies.

For most people, the work happening inside a research testbed may not be visible. But the technologies we rely on every day do not simply appear fully formed. They begin with questions. Testbeds give researchers a place to find the answers.

By providing a shared environment where ideas can be tested against the realities of scale, distance, performance, and complexity, FABRIC helps researchers move beyond what should work in theory and discover what actually works in practice.