Cloud Native Architectures
summary
Shows that an AWS region outage in September 2015 lasts five hours while Netflix’s customer-facing outage is four minutes, credited to resilient design: stateless instances, dynamic load balancing and client-side routing, externalized config, health endpoints and automated failover. Recommends parallel deploys (canary/blue‑green), circuit breakers to stop retry storms, pervasive caching, data APIs or event-driven replication with a Kafka unified log so services materialize their own views and migrations work.
AI-written summary · about
chapters (19)
- 0:00Intro
- 0:32Architecture Matters
- 1:51State of the Cloud
- 2:21Cloud Native Definition
- 3:38Cornelias background
- 4:36Cornelias book
- 5:01Amazon Web Services
- 6:50Autonomy
- 9:45Agenda
- 10:36Cloud Native App
- 11:55Dynamic Load Balancing
- 13:00Client Side Load Balancing
- 14:04Application Instances
- 15:29Sticky Sessions
- 17:46Application Configuration
- 21:08Application Lifecycle
- 27:37Service Discovery
- 29:25Circuit Breakers
- 32:07Cloud Native Data
related talks
57mResilient Cloud Architecture Design Patterns
39mAvailability, Latency, and Cost: Withstanding Regional Outages
31mKeeping Your Cloud Footprint in Check
48mResiliency and Availability Design Patterns for the Cloud
22mCloud-Native Apps, recovery ready!
14mExpect the unexpected: Building resilience with AWS
25mChaos Engineering for Cloud native Apps
31mDevOps in the Cloud: Case Studies of Amazon.com teams and their resilient architectures