Distributed systems
Consistency, consensus, and failure — the advanced core.
The concepts behind large-scale services: partial failure, replication, consensus, messaging semantics, and designing for the real world.
Who this is for: Engineers ready to go deeper than single-node thinking.
outcomes
- Explain consistency models in plain language
- Design consumers that survive at-least-once delivery
- Reason about leader election and split brain risks
lessons
- 1 Partial failure is normal Networks lie; clocks drift; retries duplicate — design for it.
- 2 Replication and consistency Leader/follower, quorum, linearizability vs eventual.
- 3 Consensus and leadership Why Raft/Paxos exist; split brain; fencing tokens.
- 4 Messaging semantics At-most-once, at-least-once, exactly-once *effects*.
- 5 Design drills Unique ID generation, distributed locks, and saga workflows.