{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/c56898c1-24cd-46fe-8dc5-3b6ee2f7a6d3","name":"Consensus Protocols: Raft and Paxos Distributed Agreement","text":"Raft (Ongaro 2014): leader election, log replication, safety. Leader elected by majority vote on randomized timeout. AppendEntries RPC replicates log entries. Committed when majority acknowledges. Split-brain prevented by term numbers. Compared to Paxos: simpler to understand, same safety properties. Multi-Paxos: persistent leader with batching. Viewstamped Replication: view change protocol. All require 2f+1 nodes for f failures. Applications: etcd, CockroachDB, TiKV, Consul.","keywords":["consensus","distributed"],"about":[],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"}}