For research institutions, a solid IT foundation can prove to be the difference in delivering meaningful results for scientific endeavors — and thereby in securing valuable funding for further research.
To that end, University of California, Riverside has launched an ambitious cloud transformation to shift from a small on-premises data center to an advanced research platform powered by Google Cloud Platform and its various service offerings.
As part of a three-year partnership with Google Public Sector, which kicked off in January, UC Riverside aims to empower its researchers in computer science, materials and quantum engineering, genomics, and precision agriculture to fully exploit Google’s location-agnostic application modernization platform, as well as its scalable compute and high performance computing (HPC) capabilities, says Matthew Gunkel, CIO of IT solutions at UCR.
Gunkel enlisted Google Public Sector professional services specifically as part of a strategy to quickly evolve UC Riverside’s small data center into an advanced cloud hub with robust research computing capabilities that would enable researchers to better compete for grants and funding opportunities.
“We identified Google as being well aligned with us strategically,” says Gunkel. “They have an agile infrastructure. They have the ability to facilitate industry-leading service concepts in additional clouds through a service they run called Anthos.”
Google’s Anthos is a hybrid cloud container platform for managing Kubernetes workloads across on-prem and public cloud environments. Gunkel also cited Google’s Looker and Big Query BI data analysis tools and its Chronicle security operations suite as important for enabling the university to operate a wide variety of applications and research on the cloud.
A partnership and cloud training model
With roughly 180 staff members, UC Riverside IT is relatively small, with largely traditional on-premises IT skills. As such, migrating to the cloud alone was not part of Gunkel’s plan.
Google’s assistance in developing a more efficient cloud architecture and training UCR’s IT staff in cloud technologies has been an immeasurably valuable service, he says, adding that Google is in a support role and is not running the show. UCR’s cloud architecture, for example, has been designed to be location-agnostic so the university is not locked into any one vendor and can adopt a multicloud platform over the long term.
“The services engagement is consulting and training to assist us in moving initial cloud workloads and to assist in our architecture to align to GCP services,” Gunkel says. “This is a ‘teach us to fish’ model. It’s all our work.”
UC Riverside IT is well on its way to migrating its core data to the cloud, developing its research platform, and shifting a range of applications to support the needs of its user base, which ranges from quantum engineering researchers to administrators, faculty, and students.
To date, UCR has moved the “vast majority of our data stores to Google,” Gunkel says, noting that his staff is currently refining the architecture and ETL processes for management and organization of the data long term.
In addition, UC Riverside IT is aligning its data to be accessed from Looker, Google’s enterprise BI and analytics platform, though which UCR will be deploying its Oracle Finance application for scaled reporting. UC Riverside is also rewriting a number of legacy applications to be cloud-native while revamping others for the cloud — there will be no ‘lift and shift’ of any applications, Gunkel says.
To that end, Google helped UC Riverside re-architect and migrate certain legacy services, including an LDAP configuration on a Solaris Unix server, as part of a process of identifying increased efficiencies for the deployment and operation of those services, which has been “an educational experience for a lot of my staff,” Gunkel says, noting that the overall transformation has required “cultural change management.”
Empowering research in the cloud
But the university’s evolving research hub is the crown jewel of the cloud migration.
“We have been working with a number of researchers on a platform that we are calling ‘Ursa Major’ where we committed to a number of compute instances and storage and RAM and GPUs that would be available to our researchers over a three-year time period,” Gunkel says.
Jim Kennedy, CTO of UC Riverside, says Google is helping architect the research hub and is also helping the IT chiefs make connections with researchers beyond UCR to help train UCR’s research faculty on Ursa Major, which will expand and grow beyond the three-year agreement with Google.
“Google connects us to experts in various research fields, and have conversations with our faculty directly, such as our genomics researcher on campus. There are experts on Google’s side, too,” Kennedy says.
Google also helped the Gunkel and Kennedy extend the university’s subscription-based compute and storage services to researchers in a multitude of disciplines. In the past, if a materials engineering researcher wanted to run workloads on several thousand processors, they would often have to write proposals to gain access to external supercomputer clusters.
With HPC requiring vast computing power, Gunkel also notes the benefit for efficiency and sustainability of shifting those workloads to the cloud. “We’re in a fairly constrained region against mountains and our ability to bring power into the university is something we’re constantly battling,” Gunkel says. “One of the things our researchers were very concerned about was [building] a sustainable, more eco-friendly solution. It’s something UCR values heavily but it’s also a challenge for us locally.”
Still, the migration, still in its early days, is being designed to accommodate a wide range of computing constituencies. For instance, UCR is also using Salesforce and MuleSoft as well as Google’s API layer to provide the “connective tissue” that is required across the university’s many enterprise platforms.
“The best way to think of the university is really as a collection or community of small businesses,” Gunkel says. “A lot of what we try to provide on the service stack side are tools that empower all of them in their different endeavors.”
Cloud Computing, Education Industry, High-Performance Computing