Dystopia as a Service
summary
Argues that modern services run on ephemeral Linux AMIs with thousands of Tomcat JVMs (5–10k web servers), deploy via Asgard, store masters in AWS with SSD-backed Cassandra, and autoscale with average instance lifetimes ~1.5 days. Proves resilience by running a chaos gorilla that kills one third of infrastructure in production with only minor errors. Publishes Netflix OSS, ships tools like denominator to unify DNS, runs a cloud prize ($10k categories) and hits 1.0 releases in weeks.
AI-written summary · about
chapters (15)
- 0:00COLLABORATION
- 4:28Tipping the Balance
- 9:06Real Web Server Dependencies Flow
- 11:13Cloud Native Architecture
- 13:22Stateless Micro-Service Architecture
- 14:22Cassandra Instance Architecture
- 17:40Ephemeral Instances
- 18:59Managing Multi-Region Availability
- 22:29Inspiration
- 25:24What's Coming Next?
- 26:02Vendor Driven Portability Interest in using NetSS for Enterprise Private Clouds
- 30:49Infrastructure Pattern Failures
- 32:21Application Pattern Failures
- 34:21Linux Foundation
- 36:36Takeaway
related talks
01h 08mDystopia as a Service
50mFast Delivery
47mCloud Security at Scale and What it Means for Your Application
54mAdrian’s Greatest Hits, B-Sides & Re-issues
39mResilience & Chaos Engineering in the Cloud
31mKeeping Your Cloud Footprint in Check
51mApplication Stacks for the Modern World
42mPlanning To Fail