In this introduction to Kubernetes, the first in a series, we’ll introduce you to the foundations of the container and container orchestration, the challenges that Kubernetes solves, and some of the associated high-level vocabulary.
What Is a Container and How Does It Work?
Containers are small, executable program components that integrate application source code with all of the OS libraries and dependencies needed to run the code in any environment.
A container, like a virtual machine, offers process and file system isolation when deployed, but with significant gains in server efficiency. This enables a considerably higher density of containers to be co-located on a single host.
Container technology has been a component of Unix-like operating systems since the turn of the century, but it wasn’t until Docker that containers were widely used.
Docker has achieved success through providing container runtime standards, for example, through the Open Container Initiative, as well as by building a full container management system around the underlying technology, easing the process of generating and deploying containers for end users. Docker, on the other hand, can only be used to run a container on a single host computer. That’s where Kubernetes comes into play.
Kubernetes: What Is It?
A container orchestration software known as Kubernetes (or “k8s” or “Kube”) automates and schedules the deployment and scaling of containerized apps.
Kubernetes allows many instances of a container to run on various computers at the same time, achieving fault tolerance and horizontal scale-out. Google developed Kubernetes after utilizing container orchestration internally to run their public services for over a decade. A long-time container user, Google had built its own proprietary methods for container deployment and scaling in data centers, which were available only to the company. Kubernetes is an open-source project that builds on such solutions, allowing a global community of software engineers to contribute to the platform’s development.
And Kubernetes adoption and deployment are increasing. Cloud-Native Computing Foundation’s bi-annual study of nearly 2000 IT professionals in North America and Europe revealed that 75% of respondents were currently utilizing containers in production, with the remaining 25% expecting to do so in the future, according to the findings. With 83% of companies adopting Kubernetes, up from 77%, and 58% using it in production, Kubernetes utilization has remained quite high.
Modern cloud infrastructure and applications rely on Kubernetes and the container ecosystem, which has evolved into a general-purpose computing platform and ecosystem that is comparable to, if not superior to, virtual machines (VMs). Because of this ecosystem’s high-productivity PaaS, development teams may focus entirely on coding and innovation instead of infrastructure-related duties and difficulties.
|
| Source aeliusventure |
Containerization and Kubernetes
For cloud deployments, the ability to manage apps independently of infrastructure is quite valuable. We can create a cloud cluster of computers that offers computing and storage resources for all of our apps, and then let Kubernetes assure optimal resource use. When demand changes, Kubernetes may be programmed to dynamically scale up and down the cluster.
Kubernetes provides several advantages for deployed applications, including service discovery, load balancing, rolling upgrades, and many others. Kubernetes serves as an application server, running all of the services, message queues, batch processes, database systems, cache services, and so on that comprise an enterprise application deployment.
Kubernetes’ versatility has fueled its adoption throughout the cloud, with all major cloud providers, such as Amazon EKS, Google Kubernetes Engine, and Azure Kubernetes Service, offering native Kubernetes services. Other container orchestration solutions, such as Red Hat OpenShift, are built on Kubernetes.
Terminology Used in Kubernetes
When it comes to Kubernetes, there is a lot to learn. This section will provide you an overview of Kubernetes vocabulary, which explains the service’s key moving elements.
Cluster: A Kubernetes environment is made up of a collection of compute nodes and storage resources. Each cluster contains at least one control plane, which is responsible for the general administration of the cluster, as well as a number of nodes on which containers will be scheduled to run. Each node needs to have a container runtime installed, which is often Docker but maybe another option such as rkt.
Pod: In the Kubernetes design, a pod is a group of containers that may be deployed and scaled collectively. This is accomplished through the use of pods, which are the smallest deployment unit in a Kubernetes cluster and allow many containers to share the same resources such as IP addresses, file systems, and so on.
Deployment: For stateless applications, deployment is used to govern Kubernetes pod generation, updates, and scaling inside the cluster. A stateless application does not need to keep track of its own client session information, allowing every instance of the program to be equally capable of handling client requests.
Stateful Sets: Maintaining the link between pods and data storage volumes is critical for some types of applications, such as database systems. In contrast to Kubernetes Deployments, Stateful Sets are used to provide each pod a distinct and persistent identity.
Stateless Apps: Stateless applications do not retain a private record of client session information, allowing any running instance of the same application to handle incoming requests. Because Kubernetes and containers, in general, are stateless, they’re easier to expand horizontally throughout the cluster.
Services: When several, interchangeable pod replicas are operational at the same time, clients require an easy way to locate any current pod they may submit queries to. Services overcome this problem by acting as a gateway to a collection of pods that may or may not be in the same Kubernetes cluster.
Storage Terminology
Volume: Storage that is assigned to a pod directly. In addition to Amazon EBS and Azure Disk Storage, Kubernetes also supports NFS and Google Persistent Disk. Volumes allow pod containers to communicate information and are destroyed when the parent pod is removed.
Persistent Volume: A volume that lives independently of a certain pod and has its own lifespan. Stateful applications such as database services may be supported by persistent volumes. This enables all components of an enterprise solution to be deployed and controlled by Kubernetes. Another significant benefit of persistent volumes is that they shield developers constructing pods from the low-level implementation details of the storage they use.
Persistent Volume Claim: A persistent volume claim that serves as a bridge between a pod and a persistent volume. Kubernetes searches for eligible persistent volumes that can fulfill the request using a persistent volume claim. The size of the necessary storage, the access mode, a selector definition used to match against labels on the persistent volume, and, optionally, a storage class name are used to conduct this search.
Storage Class: Storage classes offer another degree of abstraction to storage provisioning by allowing persistent volume claims to just define the type of storage they require. User-defined storage types such as slow, fast, and shared, for example, may all be valid. The provisioner to be utilized, the kind of volume to be created, and other provisioner-specific parameters are all encapsulated in the Storage class. Storage classes are commonly used with dynamic storage provisioning because they allow the cluster considerably more flexibility over how the storage is provisioned.
Dynamic Storage Provisioning: Static provisioning occurs when Kubernetes administrators are forced to manually configure persistent volumes ahead of time. Dynamic provisioning, on the other hand, is used to automatically assign persistent volumes depending on persistent volume claims received by the cluster. The storage class given in the claim determines the type of storage to allocate.
Provisioner: When utilizing dynamic storage provisioning, each storage class specifies the provisioner that should be utilized to build new persistent volumes. Kubernetes has internal provisioners for a variety of storage choices, such as Amazon EBS; however, external provisioners, such as NetApp Trident, may also be specified.
Conclusion
Kubernetes is the most commonly used container and microservices orchestration technology today, and it delivers the scale and flexibility necessary for delivering corporate applications and services. Using dynamic storage provisioning to manage storage in a Kubernetes cluster drastically lowers the amount of manual administration necessary for assigning cloud storage to pods and containers.
Source : kubernetes-introduction
