The previous chapter covered Paxos, the foundational consensus algorithm. Paxos is elegant but notoriously difficult to understand and implement. Google engineers spent years building Chubby on Paxos, and their famous "Paxos Made Live" paper documents the painful gap between the algorithm and a working system.

In 2014, Diego Ongaro and John Ousterhout at Stanford published Raft with a radical design goal: create a consensus algorithm that is easy to understand. Not just correct, not just efficient, but genuinely understandable by practitioners. Ongaro's PhD dissertation even includes a user study proving that students learned Raft faster and understood it better than Paxos.

The result has been transformative. Raft has become the consensus algorithm of choice for the distributed systems era. etcd (the brain of Kubernetes), Consul (HashiCorp's service mesh backbone), CockroachDB, TiDB, RethinkDB, and dozens of other systems use Raft. When engineers today need consensus, they reach for Raft.

Why should you learn Raft even if you never implement it? Because understanding Raft means understanding how your infrastructure works. When Kubernetes has a control plane failure, when your CockroachDB cluster elects a new leader, when Consul cannot reach quorum, you will know exactly what is happening and why.

This chapter provides a comprehensive treatment of Raft. By the end, you will understand not just what Raft does, but why each design decision was made.

Design Philosophy