By Jason Kohn, Contributing Columnist
Like many of us, scientific researchers tend to be creatures of habit. This includes research teams working for the National Oceanic and Atmospheric Administration (NOAA), the U.S. government agency charged with measuring the behavior of oceans, atmosphere, and weather.
Many of these climate scientists work with massive amounts of data – for example, the National Weather Service collecting up-to-the-minute temperature, humidity, and barometric readings from thousands of sites across the United States to help forecast weather. Research teams then rely on some the largest, most powerful high-performance computing (HPC) systems in the world to run models, forecasts, and other research computations.
Given the reliance on HPC resources, NOAA climate researchers have traditionally worked onsite at major supercomputing facilities, such as Oak Ridge National Laboratory in Tennessee, where access to supercomputers are just steps away. As researchers crate ever more sophisticated models of ocean and atmospheric behavior, however, the HPC requirements have become truly staggering.
Now, NOAA is using a super-high-speed network called “n-wave” to connect research sites across the United States with the computing resources they need. The network has been operating for several years, and today transports enough data to fill a 10-Gbps network to full capacity, all day, every day. NOAA is now upgrading this network to allow even more data traffic, with the goal of ultimately supporting 100-Gbps data rates.
“Our scientists were really used to having a computer in their basement,” says Jerry Janssen, manager, n-wave Network, NOAA, in a video about the project. “When that computer moved a couple thousand miles away, we had to give them a lot of assurances that, one, the data would actually move at the speed they needed it to move, but also that they could rely on it to be there. The amount of data that will be generated under this model will exceed 80-100 Terabits per day.”
The n-wave project means much more than just a massive new data pipe. It represents a fundamental shift in the way that scientists can conduct their research, allowing them to perform hugely demanding supercomputer runs of their data from dozens of remote locations. As a result, it gives NOAA climate scientists much more flexibility in where and how they work.
“For the first time, NOAA scientists and engineers in completely separate parts of the country, all the way to places like Alaska and Hawaii and Puerto Rico, will have the bandwidth they need, without restriction,” says Janssen. “NOAA will now be able to do things it never thought it could do before.”
In addition to providing fast, stable access to HPC resources, n-wave is also allowing NOAA climate scientists to share resources much more easily with scientists in the U.S. Department of Energy and other government agencies. Ideally, this level of collaboration and access to supercomputing resources will help climate scientists continue to develop more effective climate models, improve weather forecasts, and allow us to better understand our climate.
Powering Vital Climate Research
The high-speed nationwide HPC connectivity capability provided by n-wave is now enabling a broad range of NOAA basic science and research activities. Examples include:
- Basic data dissemination, allowing research teams to collect up-to-the-minute data on ocean, atmosphere, and weather from across the country, and make that data available to other research teams and agencies nationwide.
- Ensemble forecasting, where researchers run multiple HPC simulations using different initial conditions and modeling techniques, in order to refine their atmospheric forecasts and minimize errors.
- Severe weather modeling, where scientists draw on HPC simulations, real-time atmospheric data, and archived storm data to better understand and predict the behavior of storms.
- Advancing understanding of the environment to be able to better predict short-term and long-term environmental changes, mitigate threats, and provide the most accurate data to inform policy decisions.
All of this work is important, and will help advance our understanding of Earth’s climate. And it is all a testament to the amazing networking technologies and infrastructure that scientists now have at their disposal, which puts the most powerful supercomputing resources in the world at their fingertips – even when they are thousands of miles away.