Kubernetes
Kubernetes (abbreviated K8s) is an open-source container orchestration platform originally designed by Google and donated to the Cloud Native Computing Foundation (CNCF) for maintenance. It automates the deployment, scaling, and management of containerized applications, supporting container scheduling and load balancing across multi-host clusters. Core capabilities include service discovery, rolling updates, and automatic recovery, making it a foundational infrastructure component for cloud-native application architectures.
Core Facts
Go compiles to a single binary file with no complex runtime dependencies, making it easy to deploy quickly across various environments.
容器技术（以Docker为代表）利用Linux内核的namespace机制实现进程、网络、文件系统隔离，借助cgroups限制CPU、内存等资源用量
Harness Engineering概念源自软件测试领域的Test Harness（测试框架），指围绕被测系统搭建的一套自动化测试基础设施
Timeline (last 90 days)
NVIDIA's official technical blog disclosed the engineering approach that gets a cluster to an AI-Ready state on day one of Vera Rubin hardware availability, the core of which is reorganizing resource abstractions around "compute domains".
Configuring and optimizing Kubernetes for new-generation GPU hardware is complex and error-prone, involving multiple layers such as drivers, the kernel, runtimes and networking.
"Day-one readiness" for an AI cluster requires all four layers—drivers, kernel, runtime and networking—to be working at the same time, and a missing link in any one of them will cause day-one readiness to fail.
NVIDIA's investment focus in the Kubernetes ecosystem is shifting from "supporting a certain GPU" to "enabling clusters to continuously accept new GPUs".
As hardware iteration continues to accelerate, manually maintaining compatibility matrices will become increasingly unsustainable, and converging adaptation logic into abstractions such as compute domains is key.
The compute domain approach adds another level of logical boundary on top of the existing GPU resource abstraction, aggregating compute resources of the same type or generation into units that can be described and scheduled as a whole.
AI cluster day-one readiness requires the Kubernetes side to rely on the NVIDIA Device Plugin to register GPUs as schedulable resources, and the scheduler and topology manager must correctly recognize NUMA and GPU affinity.
In Kubernetes, GPUs are exposed as extended resources (such as nvidia.com/gpu) through the Device Plugin mechanism, and the scheduler only performs count-based allocation with almost no awareness of model and generation differences.
The Linux Foundation is a nonprofit open source foundation that hosts numerous key open source projects such as the Linux kernel, Kubernetes, and Node.js.
Agent Chip's remote agent runs in Pod sandboxes within isolated namespaces; the agent image fixes the tools, command-line environment, and authentication and authorization information it can access, external connections must go through the proxy layer, and only egress actively granted by the team is opened.
40 more timeline events
All Facts (20)
Go compiles to a single binary file with no complex runtime dependencies, making it easy to deploy quickly across various environments.
80%Verified容器技术（以Docker为代表）利用Linux内核的namespace机制实现进程、网络、文件系统隔离，借助cgroups限制CPU、内存等资源用量
80%VerifiedHarness Engineering概念源自软件测试领域的Test Harness（测试框架），指围绕被测系统搭建的一套自动化测试基础设施
80%VerifiedOPA (Open Policy Agent) is a general-purpose policy engine commonly used to centrally manage permissions, access control, and compliance policies at the application layer.
75%VerifiedNVIDIA Device Plugin registers the nvidia.com/gpu resource type with the Kubernetes scheduler, enabling Pods to declare GPU requirements via resources.limits.
75%VerifiedNVIDIA GPU Operator is a Kubernetes Operator that automates the deployment of GPU drivers, container runtimes, device plugins, and monitoring components, managed uniformly through the ClusterPolicy CRD.
75%VerifiedKubernetes was open-sourced by Google in 2014 and has become the de facto standard for container orchestration.
75%VerifiedKubernetes adopts a control plane and data plane separation architecture, including etcd for storing desired state, API Server as the entry point for state changes, Controller Manager running control loops, and Scheduler for scheduling Pods.
70%VerifiedBoth Kubernetes and TensorFlow adopt the Apache 2.0 open-source license.
70%VerifiedYAML has become the de facto standard configuration language in the DevOps domain and is widely adopted by Kubernetes, Ansible, and GitHub Actions.
70%VerifiedKata Containers combines the hardware isolation of lightweight virtual machines with the ease of use of the OCI container standard.
65%VerifiedThe OCI standard in the containerization field allows Docker containers to run on any compatible runtime, and Kubernetes' CRI allows different container runtimes to be interchangeable.
65%VerifiedKubernetes is a container orchestration platform open-sourced by Google and has become the de facto standard for cloud-native deployment; its HPA can dynamically scale inference instances up or down based on request queue length.
65%VerifiedCEL (Common Expression Language) is a lightweight expression language developed by Google, commonly used in policy rules and data validation scenarios.
65%VerifiedThe three basic Kubernetes Service types are ClusterIP, NodePort, and LoadBalancer.
65%VerifiedKubernetes's NodePort mode forwards external traffic to Pods inside the cluster by opening a fixed port (in the range 30000–32767) on every node in the cluster.
65%VerifiedQuarkus is a new-generation cloud-native Java framework led by Red Hat, designed specifically for containerized and Kubernetes environments
65%VerifiedSAP AI Core is an AI infrastructure service within SAP Business Technology Platform (BTP), built on Kubernetes.
65%VerifiedK9S is a widely popular open-source terminal UI tool for managing and observing Kubernetes clusters, known for its Vim-style keyboard operations.
65%UnverifiedPrometheus最初由SoundCloud开发，2016年成为继Kubernetes之后第二个加入CNCF的毕业项目
90%