In another move beyond the bounds of strictly hyper-converged infrastructure (HCI), Nutanix has launched Cloud Native AOS.
In other words, this is the operating system (OS) that underlies all its HCI products but available for deployment separately and without the need for a hypervisor. In this way, Nutanix hopes to enable its customers to extend from their core HCI deployments and to roll out, for example, Kubernetes containerisation on bare metal servers, as well as Kubernetes runtimes in the cloud.
Here, the aim for Nutanix is to allow customers to run what they want, anywhere they want. This, said marketing vice-president Lee Caswell, is distinct from the approach of many other suppliers.
“We consider that data and applications will be more distributed over time,” said Caswell. “That’s actually a different model than any hyperscaler who’s trying to concentrate all data and applications into a single cloud. Our view is it’s going to be multicloud, distributed applications and data.”
Nutanix was a pioneer of so-called hyper-converged infrastructure, which saw compute and storage bundled together in nodes that could connect in grid-like fashion to form clusters, often with server and storage components scalable independently.
This was a particularly attractive proposition – something like server-and-storage-in-a-box – to customers that lacked deep skillsets, as they were relatively easily deployable and scalable.
The idea that Nutanix should allow containers to run on its OS without the use of a hypervisor somewhat goes against the original principle it pioneered. But it announced elsewhere at this week’s event that customers can now use Pure Storage FlashArray arrays as external storage to its HCI nodes.
All of which seems to be a play to extend its market reach, especially among customers that may be seeking alternatives to VMware since it was acquired by Broadcom and underwent significant licensing changes.
Caswell said: “The way I like to think about Cloud Native AOS is we’re pushing further into cloud native, or deeper into the cloud and then further into the edge. And so, the way to think about this is that when you get into cloud native, it’s not the case that every instance is going to be running AHV [Acropolis hypervisor].”
“If you think about where we are today, with 27,000-plus customers, then when you think about getting to the edge, this tends to look like more of a distributed datacentre today. And when you get to the cloud with [Nutanix’s hybrid cloud platform] NC2, we’re bringing our full stack, including AHV, into the cloud.”
Here, Caswell has in mind a containerised AOS running on, for example, Amazon EKS (Elastic Kubernetes Service) and providing data services such as snapshots, replication and disaster recovery, all without requiring use of the Acropolis hypervisor.
“It goes back to the architectural advantage of AOS being architected independently of a hypervisor,” he said. “And so now we can containerise it and run it directly on the Kubernetes runtime that’s in the public cloud providers and our data services.”
Caswell added: “And if you look at the edge and how it is going to evolve over the next five years, my expectation is we’re going to see edge that’s going to be smaller deployments for things like running on a wind farm, or an oil rig, or in MRI centres. And these could be that they’re container only, and so you’ve got the opportunity now, for the first time, that the hypervisor is optional.
“Sure, you could run our existing stack, but you need a three-node cluster right to start, or you could start taking cloud native AOS and get down to smaller form factors in a very cost-effective manner.”
Cloud Native AOS is currently in early access on Amazon EKS, and will be generally available this summer. Early access for on-premise containerised environments on bare metal servers is expected to be available by the end of this year.
