Explaining the Raft Consensus Algorithm with "Mean Girls" (2019)

Raft is a consensus algorithm for safely and consistently replicating data in distributed systems. This article explains its core concepts through the clique relationships in Mean Girls.

• Replication and survivability: Data that stands alone like Cady is gone for good once it disappears, but with multiple replicas like the Plastics, information survives even if one node dies.
• Leader and followers: Regina corresponds to the Raft leader, while Gretchen and Karen correspond to follower replicas.
• Consensus and majority: A change cannot be committed without majority approval. With 3 members, 2 votes are needed.
• Quorum: A group too small to form a majority cannot finalize a commit. Meanwhile, a group with enough replicas can push a change through.
• Leader election: The leader periodically sends a heartbeat to signal its presence. If this fails to arrive, a new candidate is elected through voting.
• Handling client requests: Requests go to the leader, which propagates them to the replicas and commits once majority confirmation is received.

Along the way, the article connects CockroachDB with the role of the Raft leader, intuitively summarizing the core concepts of replication consistency through a teen comedy analogy.