Raft: Distributed Consensus for Dummies
summary
A hands‑on Raft demo shows a three-node cluster with a small UI and a real implementation that correctly elects a leader and replicates a persistent log. Leadership changes use monotonically increasing terms and votes, and liveness is driven by heartbeats every second and election timeouts. Entries commit only after a majority commit, guaranteeing safety; the demo works but confirms Raft is not aimed at thousands of nodes and is best for configuration or critical replicated state.
AI-written summary · about
chapters (7)
- 0:00Raft consensus intro
- 2:03Three servers in cluster
- 3:49Leader election & heartbeat
- 6:58Leader failure handling
- 9:06State machine replication
- 11:55Safety & majority guarantees
- 12:38Raft trade-offs & use cases
related talks
42mPractical Raft Consensus
37mRaft - The Understandable Distributed Protocol
44mThe Road to Akka Cluster, and Beyond
44mStreams: past, present, and future
46mBuilding high available systems with Akka Distributed Data
30mDistributed system design
45m(...)Distributed Systems
01h 00mThe Secret Life of Distributed Systems