consensus

Simplified Raft

Leader-based consensus: elect leader → replicate log entries → commit on majority ACK.

leaderN1
term1
commit index0
log entries0
👑
N1
leader (term 1)
log: []
👤
N2
follower (term 1)
log: []
👤
N3
follower (term 1)
log: []
Click followers to trigger election

// event log

[01]N1 elected leader (term 1)

How It Works

  • Leader elected by majority vote
  • All writes go through leader
  • Leader replicates entries to followers
  • Entry committed once majority ACK

Key Properties

  • Leader election via randomized timeouts
  • Log matching: same index+term = same entry
  • Safety: committed entries never lost
  • Used in: etcd, CockroachDB, TiKV