Distributed system design
summary
Builds a simple photo-uploading service as the running example and claims stateless services with images pushed to separate object storage and metadata in a database scale far better than monoliths. Uses load balancers, replicas, horizontal scaling and replication (mentions Raft/consensus) to avoid single points of failure, and suggests blocking bad traffic and autoscaling under overload. Emphasizes monitoring types (synthetic monitoring, black‑box monitoring), capacity planning, automation and blameless postmortems.
AI-written summary · about
chapters (12)
- 0:00Introduction to distributed systems
- 0:46Gathering requirements and constraints
- 3:59Designing stateless applications
- 6:27Horizontal vs vertical scaling
- 8:57Avoiding monolithic designs
- 12:22Handling storage and consistency
- 15:01Fault handling in distributed systems
- 18:50Scaling stateful components
- 20:34Mitigating overload failures
- 22:08Avoiding global configuration changes
- 23:15Service management and monitoring
- 27:29SRE principles and automation
related talks
41mIntroduction to distributed systems
01h 00mThe Secret Life of Distributed Systems
45mDo Distributed Systems they said. It'll be fun they said.
54mServices, Services, Everywhere!
11mExplaining distributed systems like I'm five
34mDistributed systems 102
13mExplaining Distributed Systems Like I'm 5
25mGrowing Distributed Systems