How Diamanti wants to bridge Kubernetes into the cloud
Commentary: Diamanti started off as a way to make on-premises Kubernetes deployments sing. Now it’s aiming to extend that to the cloud.
Despite Kubernetes being scorchingly hot, it’s nevertheless true that it remains too hard for many enterprises. Fortunately, Kubernetes is innovating at such a rapid clip that anything I write about Kubernetes today may no longer be true tomorrow. At the heart of this fast-paced innovation is a broad-based community that approaches Kubernetes development from diverse angles. Running in the cloud? You’re covered. On-premises? Also covered. Hybrid architecture? Here things get interesting.
Diamanti, which tackles the Kubernetes market with an appliance approach, has long positioned itself as the “easy button” for enterprises that struggle to weld together the software, hardware, and networking necessary to make Kubernetes hum on-premises. As Diamanti CEO Tom Barton said in an interview, it’s also trying to make it easier for companies that want to run Kubernetes in a hybrid cloud environment. Here’s how.
SEE: Kubernetes security guide (free PDF) (TechRepublic)
A generational sea change
As hot as Kubernetes is today, we’re nowhere near saturation, said Barton. A year or two back, enterprises were kicking the tires on vendors–today they’re in production. But that production is still nascent. Using the history of VMware and virtualization, Barton suggested it took roughly 15 years to go from 0% of enterprise workloads virtualized to 95% today. For containers, by contrast, he expects us to get close to 100% in less than half that time because all the major vendors have declared “Kubernetes is it.”
“It’s a generational sea change,” Barton declared. “It’s a big deal. It’s going to take over the world, because it’s just fundamentally more efficient. It’s game over.” Indeed, he continued, you’re doing it all wrong “if you’re a Fortune 500 company and you don’t develop using cloud native principles and containers and Kubernetes at this point.”
And yet many–too many–still aren’t. Why? Because it’s hard.
For Diamanti, the company tended to work with customers once it had developed some Kubernetes expertise but needed help optimizing software, hardware, and networking to boost performance. Finding customers who recognized the need for Diamanti’s appliance approach wasn’t too hard, Barton said, because the alternative was to “cobble together a bunch of crap and try to make it work together, and then try to maintain version control, ad nauseam, ad infinitum, into perpetuity, which is just really difficult.” This isn’t too different from the early days of Linux when companies tried to roll their own Linux distributions, only to discover it was far easier to rely on a Red Hat or SUSE for that service.
Another complicating factor in moving to Kubernetes is legacy: Any enterprise of size has to worry about its legacy infrastructure, which is all VM-based. A company may want to move forward with Kubernetes but first have to figure out how to retire VMware licenses, said Barton. Not only licenses, but the company also has to figure out how to re-skill employees for this new world of containers.
Now that enterprises are figuring out Kubernetes, whether with Diamanti or other vendors, they’re still stuck trying to piece together existing, on-premises infrastructure with a cloud future. For this, Barton said Diamanti has a great solution.
It’s a hybrid world
Though cloud may be the aspiration for enterprises, “They’ve got existing resources on-prem, or they’re in regulated industries, where a bunch of IT isn’t going to move,” Barton said. He cited a conversation with the CTO of Barclay’s, who said “Five years from now, 50% of our capacity will be on-prem, and 50% in the cloud.” Though Barton acknowledged such cloud estimates are “on the conservative side,” it’s still the case that demand for on-premises infrastructure continues to grow, even if the cloud infrastructure market grows faster.
Both sides are growing. Neither is going away. Hybrid, he suggested, is simply what IT will look like for a long time, and that’s particularly true of Kubernetes workloads. Said Barton:
We want to be the mechanism by which you think about brokering data and workloads back and forth to the public cloud. We ultimately don’t care what the underlying infrastructure is. That’s kind of the beauty of Kubernetes. If you write to the Kubernetes API, and somebody is Kubernetes certified, then it should work.
So for now, Diamanti is striking deals with server vendors anxious to make their hardware more attractive to would-be enterprise buyers with much better performance and the chance to gain an easier foothold in Kubernetes. But the longer-term vision is to marry the performance of on-premises, appliance-driven Kubernetes with the public cloud, serving as the hybrid control plane between the two. It’s an ambitious goal, but one more way to make a hot Kubernetes market even hotter.
Disclosure: I work for AWS, but the views herein are mine and don’t reflect those of my employer.