Kubernetes Demystified
Kubernetes is a complex open-source system used to manage Linux containers across platforms and systems. The above diagram, created using Inspiration 9, gives a bird’s eye view on how Kubernetes is organized. This diagram doesn’t show the optional connections which can be made between containers, pods and/or clusters.
Container Deployment
Kubernetes uses a container deployment system which allows it to keep each system separate from other systems. This allows maintaining secrecy where certain programming and information content is concerned simply by not activating communication with other containers or container systems.
Containers in Pods
Containers are deployed in pods. Because each pod contains connected containers that share IP addresses and share resources, they can be economically managed using kubelets, which are automatic programs that control functions. However, if it is more desirable, they can be manually managed, or managed using a combination of automation and manual operations.
Choice of Connections
Efficiency is enhanced by allowing users to choose how segments of their programming or their software connect.
These connections are handled in a variety of ways:
Kubelets
Kubelets manage the pods. They can auto restart failed containers.
Replication Controller
The replication controller also manages pods, but it uses a reconciliation loop and it scales horizontally. It can be used to create new pods or to destroy unused pods to maintain efficient use of space and to provide sufficient pods for required operations. It can restart or replace failed pods, thus preserving the integrity of the system.
Job Controller
The job controller works across the systems, handling and managing required operations.
Nodes
Nodes, which can also be called minions, can be either physical or virtual. They contain the necessary software to allow the job controller to do its work. The software might be Dockers, Kube-Proxy, Kubelets or any other OS type system that allows required operations to be performed.
Clusters
Clusters are a grouping of related pods that are controlled by a master node. The master node handles communication and cross functions between and across clusters. This enables users to efficiently carry out their work, locate their data, and generally manage operations.
Labels
Labels are used to create metadata that will allow computers and humans to locate information and functions and be able to use them efficiently. Without them, it would be difficult to deliver the desired service to all users.
This diagram and explanation might look difficult and complex. Not to worry – this is where Kubernetes expersts come in. They handle the behind-the-scenes complexity so that you and your clients have smoothly operating systems that manage your data and processes to get your needed jobs done.
