Google Kubernetes Engine

Google Kubernetes Engine
Name	Google Kubernetes Engine
Developer	Google
Genre	Container orchestration

Google Kubernetes Engine. It is a managed, production-ready environment for deploying containerized applications, built on the open-source Kubernetes system originally developed at Google. The service automates the management, scaling, and operations of Kubernetes clusters hosted on the Google Cloud Platform infrastructure. By handling complex administrative tasks, it allows developers and site reliability engineers to focus on application deployment and innovation.

Overview

This platform provides a fully managed experience, eliminating the need for users to install, operate, and maintain their own Kubernetes control plane. It integrates deeply with other Google Cloud services such as Cloud Load Balancing, Cloud Monitoring, and Cloud Storage. The environment supports a wide range of container images, including those stored in Google Container Registry or Artifact Registry, and is designed for high availability and security. Organizations from startups to large enterprises like Spotify and The New York Times utilize it to run their critical workloads.

Features

Key capabilities include automated cluster management, with features like automated upgrades, repair, and node pool scaling. It offers integrated logging and monitoring through Cloud Operations Suite and supports both horizontal pod autoscaling and cluster autoscaler functionalities. Advanced networking features are provided via Google Cloud VPC and support for network policies, while security is bolstered by Workload Identity federation and automatic security updates. The platform also includes built-in support for Istio to facilitate service mesh implementations and is compliant with standards like ISO 27001 and SOC 2.

Architecture

The fundamental unit is the cluster, which consists of a managed control plane and a set of worker nodes organized into node pools. The control plane, managed by Google's Site Reliability Engineering team, runs the Kubernetes API server, etcd, scheduler, and controller manager. Worker nodes are Compute Engine virtual machine instances that run the actual containerized applications using the container runtime. Cluster networking is handled by kube-proxy and the Container Network Interface, with persistent storage available through PersistentVolumes backed by Google Persistent Disk.

Use cases

Common applications include deploying and managing microservices-based architectures, enabling CI/CD pipelines in conjunction with tools like Cloud Build and Jenkins. It is extensively used for running stateful applications such as databases including MySQL and PostgreSQL, and for batch processing or machine learning workloads using frameworks like TensorFlow. Companies like Etsy and HSBC leverage it for hybrid and multi-cloud deployments, while others use it for hosting web applications and API backends at global scale.

Pricing

The cost structure is based on the resources consumed by the managed cluster, primarily the virtual CPUs, memory, and storage of the underlying Compute Engine instances that form the worker nodes. There is a standard management fee for the control plane, while users also pay for associated networking egress, Cloud Load Balancing forwarding rules, and persistent storage. Discounts are available through Committed Use Discounts and Sustained Use Discounts, and costs can be managed using tools like Recommender and budgets set in the Google Cloud Console.

History and development

The service was launched in 2015, following the open-sourcing of the Kubernetes project by Google in 2014, which was based on internal systems like Borg. It was one of the first managed Kubernetes offerings from a major public cloud provider, preceding similar services from Amazon Web Services and Microsoft Azure. Major updates have included the introduction of regional clusters for higher availability, the integration of the Anthos platform for multi-cloud management, and the general availability of Autopilot mode, which abstracts node management entirely. Its development continues to be closely aligned with the upstream Cloud Native Computing Foundation community.

Category:Google Cloud Platform Category:Cloud computing providers Category:Container orchestration