Kubernetes 1.0 and the Cloud Native Computing Foundation
Summary
On 21 July 2015 the container orchestration system Kubernetes reached version 1.0, after a year with more than 14,000 commits from over 400 contributors. At the same time Google announced it would found the Cloud Native Computing Foundation with the Linux Foundation, which was to manage the project in future. Kubernetes was thus considered ready for production and within a few years became the standard for running containers.
What version 1.0 brought
According to Google, 1.0 contained everything for production: DNS, load balancing, scaling, application-level health checks and service accounts. Namespaces divided clusters into areas, for example for test and production, and the system was tested with thousands of containers on hundreds of nodes.
Besides Google, contributors included Red Hat, CoreOS, IBM, Intel, Microsoft and VMware. Google handing the project to a neutral foundation was decisive for competing companies developing it together.
Ideas
- Orchestration turns many containers into a manageable service.
- A neutral foundation enables competitors to cooperate.
- Declarative configuration describes the target state instead of the steps.
- Version 1.0 means production-ready, not simple.
Insights
- Kubernetes transferred Google’s experience with large data centres into an open form; handing it to the CNCF turned it into an industry platform instead of a Google product.
- For small environments Kubernetes often remains too big; many homelabs and small companies are better off with Compose or system containers.
Facts
- Version 1.0 was released on 21 July 2015.
- By then, more than 14,000 commits from over 400 contributors.
- At the same time Google announced the Cloud Native Computing Foundation under the umbrella of the Linux Foundation.
References
Remarks
- Google writes here about its own project, and the tone is promotional accordingly.
Recommendations
- Before adopting Kubernetes, check whether the size of the environment justifies the effort.
- To get started, use a small distribution such as k3s instead of a full cluster.
Links to the original source and the Web Archive open in a new tab.