On this edition of Inside Your Data Fabric, I had the pleasure of interviewing my long-time friend and colleague, Jonsi Stefansson, Chief Technology Officer (CTO) and Vice President of Cloud at NetApp. Jonsi is, to my mind, a cloud visionary, who has unfolded a new schematic for how disparate infrastructure resources—data centers, cloud technology, and edge technology—can work together. In this edition, we discuss “true cloud”, coding with your eyes glued to the keyboard, and the joy of innovation.
Jonsi’s “true cloud concept” is a level-headed vision of access. He wants to make a difference in the cloud by enhancing cloud orchestration practices and possibilities. “You should always be able to find the perfect home for your particular application or workload anywhere, at any given time. Any cloud, public or private,” he says. And so, together with NetApp engineers, Jonsi has done just that.
We Don’t Care What Hypervisor You’re Using
NetApp® Data Fabric—as the name suggests—is the nexus where data and application management, container orchestration, and monitoring meet. The Fabric enables you to operate from a single pane, regardless of hypervisor or hyperscaler. And creating the Data Fabric has been a long and arduous journey—but a necessary one, says Jonsi.
“I think we’ve had our heads down, coding and launching different products; it’s time for us to look a little up from the keyboard and start telling the world about the products that we have been launching over the last 18 months.”
As the CEO of Icelandic tech startup Greenqloud, Jonsi’s goal was to tie storage into applications using Kubernetes and to problem solve persistent data from the viewpoint of real production workloads. In other words, the goal was to control the infrastructure from the application.
It All Started With a Snapshot
“We want[ed] the application to be able to make a snapshot of the data,” Jonsi said. They were able to create application consistent snapshots and flex clones. The technology was the first of its kind. No other storage platform could do those things. As a result, Greenqloud joined forces with NetApp.
In some ways, there’s an ocean of distance between being a startup CEO and a Fortune 500 CTO. But, as Jonsi explains it, he’s a “control freak.” A high level of organization is essential to producing enterprise-grade cloud products, regardless of company size. In both settings, it’s really just about the product. Jonsi wants to focus on just the product in his list of priorities. And his product-first approach has been successful: in his tenure as CTO at NetApp, just 18 months, the company has released more cloud products than it has in its history.
A Rigorous QA Process is Key
He explains that NetApp’s QA process is sophisticated in a way that surpasses the capabilities of startups in general. Access to a larger company, with greater resources, more customer feedback, and extensive reach has led to better code. At NetApp, we’ve collectively arranged a QA process that might keep us glued to our keyboards for a while, but its efficiency yields startling results in a shorter time period.
That’s the case even when our cloud products are managed by smaller teams.
Managed by a seven-member SRE team, our “-as-a-Service” products have grown in popularity across 50 different public cloud regions. Jonsi joked about the creation of Azure NetApp Files and how “we drink our own champagne.”
Speaking of ‘–as-a-Service' Tech: CVS or CVO? Or CI?
A central point of focus in our conversation around cloud technology was the difference between CVS and CVO. CVS stands for, of course, Cloud Volumes Service—and CVO means Cloud Volumes ONTAP. Cloud Volumes Service is all about SLAs and SLOs that aren’t offered by the public clouds. CVS is built for automation.
Cloud Volumes ONTAP is for storage administrators who want full access to turn the knobs and push the buttons on EBS volumes. Both Cloud Volumes Service and Cloud Volumes ONTAP offer performance tiering—ONTAP is, of course, the technology underlying both CVS and CVO, enabling its dynamic functionalities in the cloud.
In building Cloud Volumes Service, Jonsi said, they wanted to make everything easier, at every stage of development and for every team. The end goal? You don’t have to rearchitect to make a massive performance gain.
Cloud Insights (CI) is a crucial product for running NetApp’s services and for cloud technology and infrastructure in general. It’s a product that we use internally, too, at NetApp, Jonsi explained, detailing how we use the historical performance view in Cloud Insights to find bottlenecks and peaks. We use that information to dial-up or down services. CI allows us to be proactive instead of reactive.
Eventually, like the cloud itself, Cloud Insights will be integrated into everything.
Find Out Why Jonsi Thinks “The New Core Is The Edge.”
Plus: Why NetApp takes a very agnostic approach to CI/CD and why it’s essential to building DevOps best practices. Watch now.