Containers Change Everything
summary
Containers deliver seconds‑scale instantiation and much higher server density: Google achieves around 70% utilization versus industry ~10–15%. Using Docker + Kubernetes and a few hundred lines of Go for a custom scheduler, a prototype on standard hardware and AWS reassigns resources in real time (turns off low‑priority batch containers to free capacity for latency‑sensitive web containers), runs across major orchestrators, and produced measurable cost savings (example: saved $0.75 on an AWS test instance). Risks: weaker isolation, networking and disk I/O challenges, and stateful workloads.
AI-written summary · about
chapters (12)
- 0:00Introduction and background
- 1:19Traditional servers vs VMs
- 5:55Containers revolution
- 8:26Docker and packaging origins
- 10:32Resource utilization at Google
- 12:59Energy efficiency in data centers
- 17:12Orchestration and bin packing
- 19:20Clever scaling example
- 27:53Custom scheduler demo
- 33:21Weapons for resource optimization
- 35:25Containers winning the battle
- 36:26Q&A and closing remarks
related talks
01h 02mContainers in Production: It’s Like Orchestrating Cats
39mThe future of infrastructure will be containerized
38mThe Container Operator’s Manual
43mCloud Native - effective computing at scale
51mKubernetes: Changing the Way That we Think and Talk About Computing
56mContainers and the Cloud: Do You Need Another Virtual Environment?
29mUnderstanding the Container Revolution
05mBare Metal to Containers to K8s: How We Got Here