Architecture overview

OKD is a cloud-based Kubernetes container platform. The foundation of OKD is based on Kubernetes and therefore shares the same technology. To learn more about OKD and Kubernetes, see product architecture.

Glossary of common terms for OKD architecture

This glossary defines common terms that are used in the architecture content.

access policies

A set of roles that dictate how users, applications, and entities within a cluster interacts with one another. An access policy increases cluster security.

admission plugins

Admission plugins enforce security policies, resource limitations, or configuration requirements.

authentication

To control access to an OKD cluster, a cluster administrator can configure user authentication and ensure only approved users access the cluster. To interact with an OKD cluster, you must authenticate to the OKD API. You can authenticate by providing an OAuth access token or an X.509 client certificate in your requests to the OKD API.

bootstrap

A temporary machine that runs minimal Kubernetes and deploys the OKD control plane.

certificate signing requests (CSRs)

A resource requests a denoted signer to sign a certificate. This request might get approved or denied.

Cluster Version Operator (CVO)

An Operator that checks with the OKD Update Service to see the valid updates and update paths based on current component versions and information in the graph.

compute nodes

Nodes that are responsible for executing workloads for cluster users. Compute nodes are also known as worker nodes.

configuration drift

A situation where the configuration on a node does not match what the machine config specifies.

containers

Lightweight and executable images that consist software and all its dependencies. Because containers virtualize the operating system, you can run containers anywhere, from a data center to a public or private cloud to your local host.

container orchestration engine

Software that automates the deployment, management, scaling, and networking of containers.

container workloads

Applications that are packaged and deployed in containers.

control groups (cgroups)

Partitions sets of processes into groups to manage and limit the resources processes consume.

control plane

A container orchestration layer that exposes the API and interfaces to define, deploy, and manage the life cycle of containers. Control planes are also known as control plane machines.

CRI-O

A Kubernetes native container runtime implementation that integrates with the operating system to deliver an efficient Kubernetes experience.

deployment

A Kubernetes resource object that maintains the life cycle of an application.

Dockerfile

A text file that contains the user commands to perform on a terminal to assemble the image.

hosted control planes

A OKD feature that enables hosting a control plane on the OKD cluster from its data plane and workers. This model performs following actions:

  • Optimize infrastructure costs required for the control planes.

  • Improve the cluster creation time.

  • Enable hosting the control plane using the Kubernetes native high level primitives. For example, deployments, stateful sets.

  • Allow a strong network segmentation between the control plane and workloads.

hybrid cloud deployments

Deployments that deliver a consistent platform across bare metal, virtual, private, and public cloud environments. This offers speed, agility, and portability.

Ignition

A utility that FCOS uses to manipulate disks during initial configuration. It completes common disk tasks, including partitioning disks, formatting partitions, writing files, and configuring users.

installer-provisioned infrastructure

The installation program deploys and configures the infrastructure that the cluster runs on.

kubelet

A primary node agent that runs on each node in the cluster to ensure that containers are running in a pod.

kubernetes manifest

Specifications of a Kubernetes API object in a JSON or YAML format. A configuration file can include deployments, config maps, secrets, daemon sets.

Machine Config Daemon (MCD)

A daemon that regularly checks the nodes for configuration drift.

Machine Config Operator (MCO)

An Operator that applies the new configuration to your cluster machines.

machine config pools (MCP)

A group of machines, such as control plane components or user workloads, that are based on the resources that they handle.

metadata

Additional information about cluster deployment artifacts.

microservices

An approach to writing software. Applications can be separated into the smallest components, independent from each other by using microservices.

mirror registry

A registry that holds the mirror of OKD images.

monolithic applications

Applications that are self-contained, built, and packaged as a single piece.

namespaces

A namespace isolates specific system resources that are visible to all processes. Inside a namespace, only processes that are members of that namespace can see those resources.

networking

Network information of OKD cluster.

node

A worker machine in the OKD cluster. A node is either a virtual machine (VM) or a physical machine.

OKD Update Service (OSUS)

For clusters with internet access, Fedora provides over-the-air updates by using an OKD update service as a hosted service located behind public APIs.

OpenShift CLI (oc)

A command line tool to run OKD commands on the terminal.

OpenShift Dedicated

A managed Fedora OKD offering on Amazon Web Services (AWS) and Google Cloud Platform (GCP). OpenShift Dedicated focuses on building and scaling applications.

OpenShift image registry

A registry provided by OKD to manage images.

Operator

The preferred method of packaging, deploying, and managing a Kubernetes application in an OKD cluster. An Operator takes human operational knowledge and encodes it into software that is packaged and shared with customers.

OperatorHub

A platform that contains various OKD Operators to install.

Operator Lifecycle Manager (OLM)

OLM helps you to install, update, and manage the lifecycle of Kubernetes native applications. OLM is an open source toolkit designed to manage Operators in an effective, automated, and scalable way.

OSTree

An upgrade system for Linux-based operating systems that performs atomic upgrades of complete file system trees. OSTree tracks meaningful changes to the file system tree using an addressable object store, and is designed to complement existing package management systems.

over-the-air (OTA) updates

The OKD Update Service (OSUS) provides over-the-air updates to OKD, including Fedora CoreOS (FCOS).

pod

One or more containers with shared resources, such as volume and IP addresses, running in your OKD cluster. A pod is the smallest compute unit defined, deployed, and managed.

private registry

OKD can use any server implementing the container image registry API as a source of the image which allows the developers to push and pull their private container images.

public registry

OKD can use any server implementing the container image registry API as a source of the image which allows the developers to push and pull their public container images.

Fedora OKD Cluster Manager

A managed service where you can install, modify, operate, and upgrade your OKD clusters.

Fedora Quay Container Registry

A Quay.io container registry that serves most of the container images and Operators to OKD clusters.

replication controllers

An asset that indicates how many pod replicas are required to run at a time.

role-based access control (RBAC)

A key security control to ensure that cluster users and workloads have only access to resources required to execute their roles.

route

Routes expose a service to allow for network access to pods from users and applications outside the OKD instance.

scaling

The increasing or decreasing of resource capacity.

service

A service exposes a running application on a set of pods.

Source-to-Image (S2I) image

An image created based on the programming language of the application source code in OKD to deploy applications.

storage

OKD supports many types of storage, both for on-premise and cloud providers. You can manage container storage for persistent and non-persistent data in an OKD cluster.

Telemetry

A component to collect information such as size, health, and status of OKD.

template

A template describes a set of objects that can be parameterized and processed to produce a list of objects for creation by OKD.

user-provisioned infrastructure

You can install OKD on the infrastructure that you provide. You can use the installation program to generate the assets required to provision the cluster infrastructure, create the cluster infrastructure, and then deploy the cluster to the infrastructure that you provided.

web console

A user interface (UI) to manage OKD.

worker node

Nodes that are responsible for executing workloads for cluster users. Worker nodes are also known as compute nodes.

Additional resources

About installation and updates

As a cluster administrator, you can use the OKD installation program to install and deploy a cluster by using one of the following methods:

  • Installer-provisioned infrastructure

  • User-provisioned infrastructure

About the control plane

The control plane manages the worker nodes and the pods in your cluster. You can configure nodes with the use of machine config pools (MCPs). MCPs are groups of machines, such as control plane components or user workloads, that are based on the resources that they handle. OKD assigns different roles to hosts. These roles define the function of a machine in a cluster. The cluster contains definitions for the standard control plane and worker role types.

You can use Operators to package, deploy, and manage services on the control plane. Operators are important components in OKD because they provide the following services:

  • Perform health checks

  • Provide ways to watch applications

  • Manage over-the-air updates

  • Ensure applications stay in the specified state

About containerized applications for developers

As a developer, you can use different tools, methods, and formats to develop your containerized application based on your unique requirements, for example:

  • Use various build-tool, base-image, and registry options to build a simple container application.

  • Use supporting components such as OperatorHub and templates to develop your application.

  • Package and deploy your application as an Operator.

You can also create a Kubernetes manifest and store it in a Git repository. Kubernetes works on basic units called pods. A pod is a single instance of a running process in your cluster. Pods can contain one or more containers. You can create a service by grouping a set of pods and their access policies. Services provide permanent internal IP addresses and host names for other applications to use as pods are created and destroyed. Kubernetes defines workloads based on the type of your application.

About Fedora CoreOS (FCOS) and Ignition

As a cluster administrator, you can perform the following Fedora CoreOS (FCOS) tasks:

The OKD installation program creates the Ignition configuration files that you need to deploy your cluster. Fedora CoreOS (FCOS) uses Ignition during the initial configuration to perform common disk tasks, such as partitioning, formatting, writing files, and configuring users. During the first boot, Ignition reads its configuration from the installation media or the location that you specify and applies the configuration to the machines.

You can learn how Ignition works, the process for a Fedora CoreOS (FCOS) machine in an OKD cluster, view Ignition configuration files, and change Ignition configuration after an installation.

About admission plugins

You can use admission plugins to regulate how OKD functions. After a resource request is authenticated and authorized, admission plugins intercept the resource request to the master API to validate resource requests and to ensure that scaling policies are adhered to. Admission plugins are used to enforce security policies, resource limitations, or configuration requirements.