Living cyberinfrastructure (CI) is a vibrant collaborative venture that enables researchers to borrow the computational power and networking resources they need to complete their research — and publish it faster. This gives them a leg up on the competition, but here’s the icing on the cake: It doesn't cost them anything.
Supercomputers, data repositories, advanced instruments, visualization environments, and people — all linked by high-speed networks — make up CI. When centers and organizations provide CI for extended periods of time, living CI is the result. Together, these resources make scholarly innovation and discoveries that were once impossible, possible — at a surprising pace. In a single second, for example, a supercomputer can process over 100,000 times more data than a laptop in the same time. Some can process even more. In the race to discovery, speed is vital.
US national CI providers like the Extreme Science and Engineering Discovery Environment (XSEDE) and the Open Science Grid (OSG) are umbrella organizations that lead researchers to the CI resources needed to enable their research. Similar to a power grid, the CI computational grid serves the needs of many scientists. Instead of each researcher purchasing and maintaining all of the equipment necessary to conduct their research (which even then wouldn't add up to the muscle that living CI can provide), they can apply for allocation time at a living CI resource.
This is a different way of working for most scientists; applying for time on a machine does not guarantee their research proposals will be accepted. They might also need to adapt their code to run on the shared equipment. To use many living CI resources, scientists are required to cite the CI resource in their publications, and in some cases they must also share their data. In a sense, living CI claims a part of the scientific process.
"Living cyberinfrastructure has a longer lifespan and more community resources in support of basic research, which requires more of a give and take relationship," says Rich Knepper, manager of Campus Bridging and Research Infrastructure at Indiana University, US. "You can get great rewards from it, but it's a tradeoff. You have to change the way you think about how you're working with a supercomputer because you have to share it with other researchers, learn how to use it the correct way, and get over some hurdles to get your data and analysis there."
Living CI centers face challenges of their own. They cooperate and compete with other centers, are often inadequately funded, face scaling issues that come with exascale computing, and must attract larger scientific communities to stay viable.
XSEDE’s comprehensive three-year program report notes that living CI resources are often oversubscribed by three times. This means providers like XSEDE must choose which research benefits from their resources — and, ultimately, who is able to advance. "Having access to cyberinfrastructure means they can hit the fast-forward switch and speed up discovery," says Knepper. One catch is that the research must be data-driven to run on a supercomputer. For this reason, some qualitative researchers are beginning think of their work in new ways.
Living CI has helped scientists reveal the brain circuitry and mechanism behind anxiety, which could enable the development of treatments that could be used very early in life — such that those treated are more resilient and capable of adapting to stress when it occurs throughout life. Thanks to the advancement of scientific algorithms, supercomputers, and living CI, researchers have developed a cheaper, better way to store renewable energy and could have a home-prototype system — storing a days worth of sunshine — ready by 2017. Living CI has even led to tornado spawning predictions and helped with the development of seismic hazard maps in California.
Knepper recently gave a presentation on living CI at the International Conference on Infrastructures and Cooperation in E-science and E-humanities. A video of this presentation, along with feedback was part of the flipped conference format, in which a short video or narrated slides are prepared beforehand and published on the conference website for delegates to view before and during the conference.
Comments
Post new comment