All Together Now: An Introduction to Distributed Consensus
summary
Surveys classic and blockchain consensus with concrete mechanics: Paxos implements multi‑stage ballots to solve strict termination/agreement/validity; Raft uses a leader, replicated log and elections (used by Consul and CD systems); PBFT adds Byzantine tolerance but requires membership and PBFT tolerates <1/3 Byzantine nodes; Nakamoto proof‑of‑work picks miners by hash power, is energy‑heavy and accepts occasional forks so nodes wait 10 confirmations for practical finality; Tendermint mixes PoS voting and PBFT ideas; Stellar uses quorum slices for federated trust.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 0:54Blockchains and Distributed Systems
- 2:06First Principles
- 3:33Consensus Algorithms
- 4:28Strict Consensus
- 5:06Paxos
- 12:08Raft
- 15:42Log Matching Property
- 19:09Byzantine Failure
- 21:23PBFT
- 22:43ibilant ACK
- 23:08civil safe box
- 23:37proof of work
- 26:06theoretical problem
- 26:49accidental fork
- 28:21practical problems
- 29:31proof of stake
- 30:28Sybil attacks
- 32:00Stellar Consensus Protocol
- 33:55Trustworthy Networks
- 35:03Conclusion
- 36:20The Cap Theorem
- 37:56Questions
related talks
40mConsensus Algorithms in Distributed Systems
48mDistributed Consensus Algorithms
47mCant we just agree
07mA Walk Through Distributed Consensus
02h 44mBlockchain consensus mechanisms
51mDo Not Reimplement Distributed Consensus Algorithms
23mLightning Talks on Distributed Systems
42mBeyond Blockchain: Convergent Consensus