Nutanix Updates Kubernetes Platform
Nutanix today announced it has made available an update to its Karbon platform for deploying instances of Kubernetes on the company’s hyperconverged infrastructure (HCI) platform.
IT organizations can now upgrade an existing version of Kubernetes running on their cluster along with the guest operating system it is running on via a single click, without having to redeploy clusters or applications.
In addition, IT teams can now access to Karbon via the Nutanix Prism Central user interface, which makes it possible to employ Active Directory (AD) to add read-only only users to the Karbon platform.
Finally, IT teams can now deploy Karbon in “air-gapped’ environments where the Kubernetes cluster is physically isolated from any network. Most Kubernetes clusters are difficult to deploy in secure environments because they typically require downloading various containers from registries. That approach is not feasible in air-gapped environments, which can’t access registries over the internet, so Nutanix’s approach downloads a bundle of containers that have all the code needed to deploy and manage Kubernetes clusters in isolation.
Greg Muscarella, vice president of products at Nutanix, says the company has seen strong demand for Kubernetes clusters in on-premises IT environments. Nutanix makes Karbon available as part of the Nutanix AHV platform, which deploys a vanilla distribution of Kubernetes on top of a Nutanix-developed hypervisor. The company has not yet decided whether to make a bare-metal edition of Karbon available yet, he says, noting many organizations overestimate the impact that a virtual machine has on overall performance while underappreciating the need to isolate instances of Kubernetes.
Longer-term, lighter-weight virtual machines will be able to maintain those levels of isolation with even less impact on performance, he adds.
For now, most IT organizations are opting to deploy a larger number of smaller Kubernetes clusters on separate machines versus running many Kubernetes clusters on one large machine. The reason for that can be attributed to both the preference of application owners as well as the fact that it tends to be more operationally efficient to manage change within the context of a smaller cluster. The number of smaller clusters that will need to be managed is also likely to increase as the number of edge computing use cases for Kubernetes expands, he adds.
Muscarella notes there may never be one right model of deploying Kubernetes clusters that fits all customers. However, as many IT organizations embrace Kubernetes, they are not sticking with their existing IT infrastructure vendors by default.
At this point it’s clear Kubernetes is on the cusp of becoming a mainstream IT platform both in the cloud and in on-premises IT environments, says Muscarella. Eventually, adoption of Kubernetes clusters on-premises and in the public cloud should foster a new era of hybrid cloud computing. The challenge now is finding a way to not only spin up all those Kubernetes clusters but also manage them at scale.
