Oracle’s Container Engine for Kubernetes to offer managed virtual nodes

Oracle is adding virtual nodes to its managed Kubernetes service called Oracle Container Engine for Kubernetes (OKE) in an effort to let enterprises run their development operations without having to manage any infrastructure.

Nodes, which are one of the most fundamental building blocks of Kubernetes, are physical or virtual machines that make up clusters that in turn run Kubernetes and the containers managed by that particular instance of the orchestration system.

OKE’s new virtual nodes, which were first announced by the company in October last year, will eliminate the operational overhead of managing, scaling, upgrading, and troubleshooting worker nodes' infrastructure (servers), said Vijay Kumar, vice president, of product marketing, app development services and developer relations at Oracle.

OKE already offers a managed service that provides what the company calls managed Kubernetes nodes to enterprises. These nodes, according to Oracle, run on a shared operational model under which customers control the configuration of the nodes based on their requirements and Oracle is responsible for provisioning and updating software on the managed nodes.

The virtual nodes take away all responsibilities from enterprises and can scale automatically as required, Kumar said, adding that the new virtual nodes could result in more cost savings for many enterprises depending on the scale of their development operations.

OKE bets on lower pricing

Oracle claims that OKE is far cheaper than other Kubernetes services, including Amazon EKS with Amazon Fargate, Microsoft Azure’s AKS Virtual Nodes (using Azure Container Instances), and Google Cloud’s GKE Autopilot.

“Customers will be able to drive their businesses forward while realising cost savings of up to 50% compared to other cloud providers,” Kumar said. The pricing for OKE is the same across all regions in contrast to competing Kubernetes services from rival public cloud service providers where pricing varies from one region to another, Kumar added.

OKE’s pricing, according to Oracle, is calculated based on the sum of the CPU fee, memory fee, boot volume fee, node fee, and cluster fee.

While boot volume fee is only calculated for managed nodes, the node fee is only applicable for virtual nodes, Oracle said.

Oracle is using this aggressive pricing to gain a larger pie of the container infrastructure market, where it currently has very little presence, according to analysts.

“Oracle is a very small player in the container infrastructure market and is not among the top listed vendors in this market,” an analyst, who didn’t wish to be identified, said citing the aggressive pricing strategy.

Another reason why Oracle is doubling down on Kubernetes is its rising demand from enterprises due to the need for accelerating cloud-native application development in order to service customers, said Gary Chen and Lara Greden, research directors at IDC.

“Our most recent forecast for the Container Infrastructure Software market put its growth at 26.6% over five years. Note this is only for container orchestration services and does not count the infrastructure-as-a-service (IaaS) resources consumed by these services,” Chen said.

A separate survey conducted by IDC in the US showed that 45% of enterprises polled are increasing spending on Kubernetes compared to other areas such as database management, application platforms, events streaming, and robotic process automation.

Workload portability is another reason behind the rising popularity of Kubernetes, said Holger Muller, principal analyst at Constellation Research. “Workload portability can be difficult if you pick up outside of the container dependencies – which most use cases might require.”

Financially backed SLAs

In addition to service-level agreements (SLAs) for nodes, Oracle said it will provide “financially backed SLAs” for uptime and availability for the Kubernetes API server.

Additional new features of OKE include a Lifecycle Management service along with the ability to identify workloads.

The Lifecycle Management service allows enterprises to install and configure their chosen auxiliary operational software or related applications, the company said, adding that the service can manage the full lifecycle of this software, from initial set up and deployment through ongoing upgrades and patching.

Add-ons to the management service include essential software deployed on the cluster such as CoreDNS and kube proxy, and optional software operators such as Kubernetes dashboard, Oracle Database, and Oracle WebLogic among others.

The new features are also an indicator of a new strategy from Oracle for its Oracle Cloud Infrastructure (OCI) business, said Lee Sustar, principal analyst at Forrester. “OCI is positioning itself to be not only infrastructure for traditional Oracle workloads, but also as a provider of the cloud native service types that are most frequently used in enterprise-class IT. Managed Kubernetes services are foundational to that.”

In addition to the management service, OKE will now include the ability for organisations to identify workloads at the pod level. Pods, which are the smallest building block of Kubernetes, add up to make a node.

This ability at the pod level will strengthen OKE’s granular security, which is one of the top challenging aspects of cloud-native development, Sustar said.

OKE is now in general availability.