kantver - Fotolia
Hyper-converged infrastructure has proven an effective strategy for running a wide range of virtualized applications. Hyper-converged systems are easy to deploy and manage and can help lower the total cost of ownership over traditional infrastructure.
Many organizations have been moving to modern, cloud-native applications, however, which rely on containers rather than VMs. Containers are more portable and scalable and less resource-intensive. Unfortunately, they don't always fit easily into the hyper-converged infrastructure model. But that's changing as more vendors add container support to their products.
Containers facilitate faster development and deployments and can easily scale across on-premises and public cloud platforms. Unlike a virtual machine, which runs its own operating system, containers share most of its system components with the underlying host OS, reducing much of the overhead that comes with a VM.
Enterprises that deploy containers often use Kubernetes to manage container execution and scheduling. Kubernetes is a popular container orchestration tool that many consider the de facto standard for container management.
However, Kubernetes is a complex product that requires a high level of expertise in order to build and maintain a production environment.
To complicate matters, companies adopting Kubernetes and containers are likely still supporting virtual server environments, adding yet another layer of complexity to their operations. For those that have deployed a hyper-converged infrastructure (HCI), managing these environments is no doubt easier. It doesn't help with their containers, though -- unless they've deployed hyper-converged systems that support Kubernetes, such as those offered by VMware, Cisco and Nutanix.
VMware vSphere with Kubernetes
VMware has integrated Kubernetes into the vSphere control plane. VMware vSphere with Kubernetes is part of VMware Cloud Foundation, an HCI software stack that provides unified management for compute, storage and network resources. The new integration makes it possible to deploy Kubernetes clusters, containers and VMs through the Kubernetes APIs without having to use the vSphere APIs.
VMware Cloud Foundation includes two Kubernetes-related families of services. The first is Tanzu Runtime Services, which delivers core Kubernetes development capabilities along with a distribution of the Tanzu Kubernetes Grid. The second family is Hybrid Infrastructure Services. This facilitates full Kubernetes and RESTful API access, making it possible to create and manipulate containers and VMs, as well as storage and networking resources.
Dell EMC has been quick to adopt VMware technology in its VxRail hyper-converged appliances. These appliances come with vSphere and other VMware Cloud Foundation components, and now they support Kubernetes. Dell Technologies worked in partnership with VMware to engineer multiple VMware Tanzu on VxRail systems, which range from a reference architecture to an automated turnkey product. In this way, customers have multiple options for implementing Kubernetes within a hyper-converged environment, while still being able to support their legacy applications.
Cisco HyperFlex Application Platform
In 2016, Cisco launched its line of HyperFlex HCI appliances. More recently, Cisco introduced the HyperFlex Application Platform (HXAP), which expands the HyperFlex architectural framework to provide native support for Kubernetes workloads. HXAP offers an integrated container-as-a-service platform that simplifies provisioning and maintenance across on-premises and public cloud environments. The service provides integrated lifecycle management and supports single-click installations and upgrades.
With HXAP, Cisco integrated Kubernetes components into the hyper-converged environment, providing customers with an easy-to-implement Kubernetes platform for their cloud-native applications. The platform maintains the components that make up the Kubernetes environment, including the OS, libraries, packages and patches. It also manages security updates, supports DevOps methodologies and includes additional tools for storage, container networking, load balancing and other operations.
Nutanix offers several software products for implementing a hyper-converged infrastructure. One of these is its AHV hypervisor, which provides license-free virtualization in the hyper-converged stack. A more recent addition to the software stack is Nutanix Karbon, an enterprise-grade Kubernetes management tool that's fully integrated into the AHV hypervisor. Karbon simplifies the process of deploying, configuring and managing Kubernetes clusters throughout their lifecycles, while providing open APIs for interfacing with the Kubernetes environment.
Nutanix recently released Karbon 2.0, which includes several important enhancements. For example, customers can now access Karbon's features through the Prism Central user interface. They can also initiate one-click upgrades that deploy newer Kubernetes versions without having to redeploy clusters or applications. In addition, Karbon 2.0 adds support for air-gapped environments, which are physically isolated from unsecured networks, including the internet.
Red Hat OpenShift Container Storage
Red Hat Hyperconverged Infrastructure is based on two platforms: Red Hat Virtualization and Red Hat OpenShift.
Red Hat Virtualization is a KVM-based virtualization platform, while OpenShift is a Kubernetes-based container orchestration platform. OpenShift is considered a downstream solution from the Kubernetes open source project. As such, it's undergone additional testing and includes productivity and security features not available to regular Kubernetes.
Red Hat has also partnered with NetApp to provide a data center product for running DevOps workloads on NetApp HCI appliances. NetApp HCI is a hybrid cloud infrastructure that incorporates both the VMware platform and OpenShift container platform. This combination enables NetApp HCI to deliver Kubernetes within the VMware virtualization environment, making it possible to take advantage of VMware's security, efficiency and high availability.
Red Hat also offers OpenShift Container Storage (OCS), software-defined storage that's integrated with and optimized for the OpenShift container platform. Red Hat OCS provides persistent data services for delivering file, block and object storage. It can run anywhere OpenShift runs, using a single set of Kubernetes operators across all cloud platforms.
Google Cloud Anthos
NetApp has also joined forces with Google to deliver hybrid cloud based on NetApp HCI. The appliances have been validated to work with Google Cloud Anthos, providing a verified hybrid cloud infrastructure for deploying the Google Kubernetes Engine (GKE). The GKE is a certified Kubernetes distribution that manages container-based operations both on premises and in Google Cloud, delivering a unified hybrid offering that facilitates container management across environments.
Additionally, HPE offers a validated design for Anthos on SimpliVity, its line of hyper-converged appliances. Similar to NetApp HCI, HPE SimpliVity has been validated to run the GKE, making it easier to develop, test and manage containerized applications.
Cisco's HyperFlex appliances have also embraced Anthos, with the goal of providing a common infrastructure for both container-based and server-based virtualized workloads.
Hyper-convergence and Kubernetes
The platforms mentioned here are only a sampling of the efforts currently underway to bring Kubernetes into the hyper-converged infrastructure fold. They represent a good cross-section of the different ways in which hyper-converged vendors now facilitate containerization, however.
Some vendors have also added support for the Container Network Interface, a specification that defines how to write plugins for configuring network interfaces in Linux containers.
Before long, most hyper-convergence products will support containers in some form. If they don't, they'll be left behind. Hyper-converged offerings that integrate Kubernetes make it possible to run traditional and modern applications within the same environment, while still delivering the benefits of hyper-convergence. Not only does this help to streamline IT operations, but it also makes room for new innovations.