Cory Snavely Named Group Lead for Infrastructure Services

Cory Snavely, a senior computer systems engineer who joined NERSC in March 2015, has been named group lead for the Infrastructure Services Group (ISG). He has been acting lead of ISG since the group was created as part of a reorganization at NERSC in early 2016.

Snavely, who earned his bachelor’s of science degree in applied science at Miami University, was managing Library IT Core Services at the University of Michigan before making the trek west last year. In fact, his decision to move to the Bay Area – specifically the East Bay – was a conscious one, as was his choice to work in a research environment rather than a tech startup, corporation, or university.

“I specifically sought a position at NERSC because I’d enjoyed building large-scale digital library systems, felt very at home in academic environments and had grown curious about research and in particular how large-scale computational research was evolving,” he said. “So it was a very deliberate choice to jump into a different industry and bring my experience to bear in a new way.”

Snavely is drawn to the challenge of creating new infrastructure to support the evolving computational and data-intensive science workflow needs of NERSC’s 6,000 users.

“As we see continued growth in the amount of data produced at experimental facilities and increases in both network performance and computational capacity, we have an opportunity to support new, more complex DOE research projects and workflows,” he said. “ISG is building forward-looking services that are tightly integrated with the supercomputing environment to facilitate these projects.”

This new, scalable infrastructure will be container-based to support flexible computing and data services and enable, for example, real-time analysis of experimental or observational data from a remote facility. Containers allow applications to be packaged with their entire software stack, making them portable, scalable, and reliably reproducible – all critically important for researchers.

“We are embracing the container approach,” Snavely said, “to help us deploy services at NERSC for DOE Office of Science users, who will be able to quickly and easily couple them with large-scale computational analyses on NERSC’s HPC systems.”

ISG is working with NERSC’s Data and Analytics Services and Data Science Engagement groups over the next 6-12 months to demonstrate proof-of-concept deployments of databases, web-based data repositories, research workflow managers, automated software development tools, and other services running in the new environment, according to Snavely. In addition, ISG is identifying collaborators in different science areas to work with and provide feedback. For Snavely, it’s an exciting time to be in HPC and to be pioneering new approaches to facilitating data-intensive science.

“We need to conceive new models for how to deploy these services, but we also have to respond to ever-increasing demands – more data, faster access, quicker response times, etc.,” he said. “At the same time, there is greater interest in coordinating work across different facilities – the ‘superfacility’ concept – which creates other challenges for interoperability, authentication, and security. It’s an exciting and experimental time. We are building new capabilities and understanding the emerging use cases as we go.”