The future of container-enabled infrastructure
summary
CoreOS argues for separating apps from the OS using containers (Docker/rocket) and a minimal CoreOS Linux, then managing rollout and uptime with a scheduler and cluster store: etcd consensus key-value store drives leader election and config, Kubernetes handles rolling updates with replication controllers, and the OS gets atomic A/B OS upgrades so reboots happen safely. Claims include an automated vuln workflow (identify outdated images, trigger rebuilds), a measured finding that ~80% of public-registry containers are outdated, and production scale to thousands of nodes.
AI-written summary · about
chapters (6)
- 0:00Start
- 1:32Tolerate Machine Failures
- 1:53Simplify Application Upgrades
- 5:43Application Packaging
- 14:49Clustering
- 50:21Make Servers Consistent
related talks
53mManaging Containers at Scale with CoreOS and Kubernetes
24mCoreOS Demystified – The Specialized OS Powering OpenShift and Containers
13mCoreOS is Dead, Long Live CoreOS
07mKeynote: Pushing the Boundaries of Open Source for a Brand New Edge
25mContainers. Microservices. Cloud. Open Source. Fantasy or Reality for Enterprises?
07mKeynote: Containers Where YOU Want Them: All the Way to the Edge
10mContainers and Kubernetes are here - ok, but what’s next?
51mA Greybeard's Worst Nightmare