The Cloud Native Computing Foundation (CNCF) has released Kubernetes 1.13, an update that formally makes available a kubeadm capability to make it easier to administer Kubernetes.
Kubernetes 1.13 also makes available a much-anticipated Container Storage Interface (CSI), in addition to making open source CoreDNS now the default Domain Name System (DNS) for Kubernetes.
In terms of new capabilities, Kubernetes 1.13 also introduces support for third-party device monitoring tools as an alpha feature.
Capabilities that have graduated to stable with Kubernetes 1.13 include support for Kubelet Device Plugin Registration to make it easier to discover plugins and a topology Aware Volume Scheduling feature that makes it possible to determine the topology constraints, such as zone or node, of a volume running on a Kubernetes pod.
Graduating to a formal beta status with Kubernetes 1.13 is an APIServer DryRun that will take over “apply” and declarative object management; a Kubectl Diff capability that enables users to see the difference between a locally declared object configuration and the current state of a live object; and raw block device using persistent volume source capability that makes raw block devices available for consumption via a Persistent Volume Source.
The CNCF is on a quarterly cadence for delivering updates to Kubernetes. Aishwarya Sundar, a software engineer for Google who works on the Kubernetes project, says the CNCF expects to make support for Windows containers generally available with the next release of Kubernetes.
It typically takes most of the IT vendors that curate various distributions of Kubernetes another a couple of months to certify new capabilities once they become generally available. But in terms of usage of Kubernetes in production environments, CSI support is critical because it enables Kubernetes clusters to access persistent storage that resides outside the cluster.
CoreDNS, meanwhile, provides a faster implementation of DNS written in the Go programming language.
The rate at which Kubernetes is now being deployed in production environments inside and out of the cloud has risen dramatically in the last year. Going into 2019, that rate should only accelerate as more stateful applications that require access to external storage become more widely deployed on Kubernetes.
The biggest challenge facing IT organizations now is not so much learning how to provision Kubernetes as much as it is the ongoing management of those clusters. In addition to still being challenged with managing Kubernetes, IT organizations may soon find themselves confronted with a Kubernetes sprawl problem as each development team stands up their own clusters. IT organizations also may find themselves managing multiple distributions of Kubernetes running in a public cloud or in a local data center. Those local data centers could also be running Kubernetes on virtual machines or on a bare-metal server.
Whatever the path chosen, it’s pretty clear by now that many more IT administrators will be attaining Kubernetes certifications in the months ahead as the number of containerized applications that need to run on a Kubernetes cluster continue to increase.