Researchers poised to deliver key scientific findings in fast lane

Streaming data, the online process that brings movies, music and videos instantly into homes, is poised to accelerate the worldwide distribution of scientific discoveries in fusion research. Scientists at the U.S. Department of Energy’s (DOE) Princeton Plasma Physics Laboratory (PPPL) and Oak Ridge National Laboratory (ORNL) have demonstrated the streaming and rapid analysis of findings from a fusion reactor in South Korea to Princeton, New Jersey. Participating in the work were researchers from the Korean national data transmission network (KREONET), the DOE’s Energy Science Network (ESnet), and the Korea Institute of Fusion Energy (KFE) that operates the reactor.

Unveiling early stages

The streaming demonstration, performed in 2017, unveiled the early stages of a system whose development is forging ahead at PPPL and ORNL. Key goals of the format include the ability to rapidly deliver and analyze the massive flow of data expected from ITER, the international experiment under construction in France to reproduce the fusion energy that powers the sun and stars. The early demonstration was streamed from the Korea Superconducting Tokamak Advanced Research (KSTAR) fusion reactor in Daejon, South Korea, to PPPL in Princeton, New Jersey.

“We’ve continued to build on that framework,” said PPPL physicist Michael Churchill, lead author of a recent Fusion, Science & Technology paper that spells out the details. “The goal is to connect experimental data with larger computer facilities, to enable more compute and data intensive analyses during experiments. Faster, better analysis will lead to better and timely decision making, which will accelerate scientific discovery and ultimately the achievement of the goals of ITER.”

Streaming flows directly from sender to user without content being saved on disks at the user’s end – the capability that has made music CDs and video stores virtually obsolete. “The advantage for science is that we can analyze the data while they are moving from one computer to another,” said PPPL physicist C.S. Chang, a co-author of the paper who led the project together with computer scientist and co-author Scott Klasky from ORNL, with the support of DOE. “We don’t need to wait for the data to be downloaded on disks and then have to search for it.”

Fusion, the power that drives the sun and stars, combines light elements in the form of plasma – the hot, charged state of matter composed of free electrons and atomic nuclei, or ions, that makes up 99 percent of the visible universe – to generate massive amounts of energy. Scientists are seeking to reproduce and control fusion on Earth for a virtually inexhaustible supply of safe and clean power to generate electricity for all humanity.

PPPL and fusion laboratories around the world conduct experiments on doughnut-shaped tokamaks, the most advanced and widely used type of fusion facility, which heat the plasma to million-degree temperatures and confine it with magnetic fields to get the ions to fuse and release their energy. Theoretical physicists perform large-scale simulations of the findings to understand them and predict future results.

Into the fast lane

Scientists at PPPL and ORNL are now developing DELTA, a general framework for moving the delivery of magnetic fusion findings into the fast lane. Researchers Ralph Kube of PPPL and Jong Choi of ORNL have detailed the framework at international conferences, describing how the process links magnetic fusion experiments to high-performance supercomputers that can analyze the findings as they roll in.

“With a framework like this, you can stream data globally, do parallel analysis and modelling in near real-time, and you don’t have the bottlenecks of writing to disks as the data arrives,” Churchill said. Detailed analysis was performed on the Cori supercomputer at the National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory (LBNL), with NERSC researcher Laurie Stephey contributing.

A key benefit of streaming, said Churchill, is that, “There are a lot of international partners around the world sharing new results from advanced experiments and simulation codes and we could send the data anywhere for rapid analysis. We’re showing the research, we’re showing the utility, and we’re taking the approach that if you build it they will come.”

Going forward, “We’re hardening the framework with an eye toward wide usability,” he said. “The biggest thing for me is that the framework will really shine when we get other people involved in their own instant analysis.”

/Public Release. This material comes from the originating organization and may be of a point-in-time nature, edited for clarity, style and length. View in full here.