The Center for Advanced Research Computing (CARC) at the University of Southern California (USC) supports a massive, multidisciplinary research ecosystem. As an interdisciplinary hub, CARC collaborates with faculty members, graduate students, and researchers across all campus divisions to empower computational, AI, and data-intensive research.
The Challenge: Data Explosion
The computational landscape shifted dramatically around 2010 with an explosion of genomics and life sciences data. “The volume of data was just overwhelming,” notes BD Kim, Associate Chief Research Information Officer at USC CARC. “The issue of how to deal with that large volume of data has become the main topic.”
As modern AI and deep learning applications expanded, the pressure on infrastructure intensified. Data fields grew beyond traditional scientific domains to encompass inputs from all areas of society. Furthermore, massive data scales inverted traditional workflow paradigms. As Kim explains, “Before, even a few years ago, you usually brought your data to computation. But now, because of the data volume and data locality, you now usually bring computation to the dataset.”
The Solution and Results: A Simple, Flexible, and Scalable Data Platform
To eliminate legacy storage bottlenecks and manage the pace of change, USC CARC deployed VAST. CARC consolidated its environment onto a 10 petabyte system, utilizing VAST’s highly flexible architecture to deliver seamless multiprotocol data access.
The impact on operations was immediate. “We switched to VAST a couple of years ago,” says Kim. “It’s very flexible, and administration has been very easy.” Beyond the technical advantages of simplified management, VAST’s co-pilot support model has acted as a true extension of the university’s infrastructure team. “The support from VAST has been pretty awesome, very responsive, and they always try to solve any problem together. We’ve been very happy with that.”
By partnering with VAST, USC CARC has future-proofed its infrastructure, allowing researchers to spend less time wrestling with data logistics and more time driving scientific breakthroughs.
