The Art of The Laptop Cluster
summary
Claim: running services on a laptop fails in different ways than production, so shipping a laptop container is not enough. Evidence: Docker (2013 promise) and Docker 1.12 features, local cluster managers like Mesos or minikube, visualizers such as Weave Scope and Raft demos, and topology tools (spiegel/simian style) are shown. Outcome: local clusters reproduce service discovery, DNS, network topology and failure scenarios reliably, but they do not produce realistic load or production-grade performance testing.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 2:10What do you guys nuts
- 2:22Shoutout
- 3:26Failure modes
- 5:56Performance differences
- 8:25Other contextual differences
- 9:48Installing software
- 10:34Security
- 10:58Mental Model
- 11:40The Problem
- 12:07False Hopes
- 13:59Dont Worry About It
- 15:25Zenos Paradox
- 16:22Getting Warmer
- 18:54Simian Biz
- 21:35Requirements
- 24:49Where are we
related talks
05mUsing production service definitions on developer laptops
32mKubernetes: Pimp My Laptop
46mDeploying and maintaining software for today's infrastructure
49mIn Silicon Valley, Engineers Deploy-and-Tell: Stories About Dockerizing an Application
59mBeyond Docker Buzzwords
57mMoby Dock in the Wind Tunnel
33mHow Container Clusters, like Kubernetes, ...
35mCluster as Code