Speed & Scale: How to Get There
summary
Netflix moves its customer-facing systems to the cloud in 2010 and replaces Oracle with Cassandra, running production clusters like a 288-node Cassandra cluster and testing at thousands of nodes, all backed by 40+ open source projects. Deployment is continuous with automatic deploys and feature flags, developers run code in production and are on call, resilience is enforced with tools like Chaos Monkey, and the result is much faster, incremental innovation.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 10:52Innovation
- 16:15Cassandra
- 18:15Cassandra scalability
- 21:21How to speed up
- 23:47What is SAS
- 24:59The Loop
- 27:38Incremental Deployment
- 30:46Continuous Delivery
- 31:06Dev Ops
- 33:41Less downtime
- 36:20DevOps is a reorg
- 38:47Developer adoption is a leading indicator
- 39:23The evolution of products
- 39:47Vertically integrated teams
- 41:10Crossing the chasm
- 42:04Out of the swamp
- 42:36Open source
- 44:56Open source benefits
- 45:25Hub
- 46:36Microservices
- 47:34Comparing monolithic vs microservices
- 49:56Conways law
- 52:29Client libraries
- 55:21Data stores
- 57:58Monitoring
related talks
29mHow Netflix Thinks About DevOps
38mIt's Simple…
20mSoftware Performance Optimization and Scaling
46mGo Faster -Impact on Technology, Process and Organization
47mWeb-Scale IT Engineering.
58mBuilding a Culture of Continuous Delivery
45mEvolution of Spinnaker Internals & Operations
30mA conversation with Netflix