Understanding the Container Revolution
summary
Containers predate Docker: LXC (2008) uses Linux namespaces and cgroups to isolate processes, while Docker adds an image format, build DSL and registries that make packaging reproducible. Corporate turmoil prompts splitting out components like containerd; alternatives like rkt exist. Orchestration is the real shift — Kubernetes (Borg ideas from ~17 years at Google) schedules containers, relocates failed instances, mounts persistent volumes and autos-scales, turning containers into self-healing application-centric platforms that start to pay off once you run roughly three or more machines.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:53a brief history about containers
- 5:40security goods are not visible from outside
- 7:07containers in software
- 9:33Linux namespaces
- 11:46Docker as a "better virtual machine"?
- 16:31what was the problem again?
- 18:19a lot of work doesn't scale takes too long high failure rate deployment = downtime high mean time to recovery
- 18:40Normalization
- 19:51Continuous Delivery
- 20:52Infrastructure as Code
- 22:21This is all focussed on infrastructure, not the application!
- 22:35a container stack by itself is just a stack of containers
- 24:49Kubernetes will become for cloud hosting what Linux is for servers.
- 26:47Containers can make the process of delivering software more efficient
- 28:03will change the way you think about software architecture, operations, deployment, monitoring, debugging, ...
related talks
20mEmergence of Linux Application Containers and the Rise of Kubernetes
05mBare Metal to Containers to K8s: How We Got Here
49mContainers without Docker
30mContainers and Next Gen Infrastructure Ecosystem
35mAnd then there was k8s
05mIgnite: The History of Linux Containers
43mContainers Change Everything
35mIntroduction to containerd