Survival, "The Netflix way"

How Netflix Architects for Survival: @ InfoQ

The Netflix way:

• Everything is “built for three” (i.e. VMs, data centers, etc.)
• Fully automated build tools to test and make packages
• Fully automated machine image "bakery"
• Fully automated image deployment
• Independent teams responsible for both Dev and Ops
• Redundancy through multi-region deployment

Culture of "Freedom and Responsibility"

Automate all the things!

Shared state should be stored in a shared service

RESTfull services APIs (has its own CDN)

Highly aligned, loosely coupled (SOA)

All systems choices assume some part will fail at some point.

Reliability and $ (balance)

The Monkey Theory

• Simulate things that go wrong (automation, tool "ASGARD")
• Find things that are different

Execution

• Running a new and old version clusters (instead of instances) in parallel, until stable

"Baking VMs" means no configuration at startup time, it is fully ready to run when started.