As research opportunities continue to expand for the University of Guam EPSCoR- Guam Ecosystems Collaboratorium for Corals and Oceans (GECCO) program, so does the need to improve its cyberinfrastructure to keep up with the additional computational and data analysis requirements.
Part of the UOG EPSCoR-GECCO strategic plan is to implement high throughput computing (HTC) resources in Guam and to establish partnerships that would broaden access to off-campus HTC resources. According to the plan, “leveraging existing partnerships to enable remote access to HTC resources, implementation of local HTC hardware and effective user support will accelerate UOG’s capacity for data-intensive research, moving UOG closer to its goal of becoming a research-intensive university.”
To beef up research computational capacity, the program is looking at tapping into the Open Science Grid (OSG). According to the strategic plan, the OSG will provide project research access to its distributed computing network to facilitate parallel computing and support GECCO research.
Jeffrey Centino, research computing facilitator at EPSCoR –GECCO said the OSG
is a collaboration between institutions, universities, and other research organizations to forward the field of science through high throughput computing.
“So basically, you have these data centers located around the world and they are connected through high-speed internet, and they function together like a grid. So, say you need to run a job or an analysis, you can recruit these resources from around the world to complete your job in the fraction of the time compared to what is available to you in a single data center or your personal workstation.”
He said high throughput computing breaks the computational work into smaller tasks, which can run concurrently using these resources. “Right now, we are running an analysis server and it is very under powered so basically researchers are fighting for computational space and once we get those researchers onboarded to the Open Science Grid, they can have their jobs or their analysis run within a fraction of the time,” Centino added.
Centino said they are also expanding their computer clusters to support grid capacity, but the worldwide chip shortage caused by the COVID-19 pandemic restrictions delayed the process of acquiring the servers. “So, we are looking to launch early next year for the computer cluster, around the first quarter,” he said.
The OSG has over 100 participants that provide access to large computing resources. The most notable ones include the Large Hadron Collider Beauty Experiment, the Fermi Natural Accelerator Lab (Fermilab), and Dark Energy Survey.