Kubernetes

Kubernetes (K8s) originated at Google (derived from its internal Borg scheduler) and was donated to the Cloud Native Computing Foundation (CNCF) in 2014. It provides a declarative API for defining the desired state of containerised applications — how many replicas, what resources they require, how they communicate — and continuously reconciles actual cluster state against the declared intent.

In datacentre and cloud contexts, Kubernetes is the standard for managing containerised microservices, data pipelines, and AI/ML workloads. The NVIDIA GPU Operator enables K8s to schedule GPU-accelerated training and inference jobs across clusters, with support for fractional GPU allocation via MIG and time-slicing. Major managed Kubernetes offerings include AWS EKS, Azure AKS, and Google GKE. On-premise K8s deployments run on bare-metal or VM infrastructure, and form the control plane of many enterprise AI factory deployments.