Quorum (distributed systems)

In distributed systems, a quorum is the minimum number of nodes whose agreement is required to perform an operation safely. The concept is the mathematical backbone of consensus protocols like Paxos and the Raft algorithm: a value is only committed once a majority of nodes acknowledge it, ensuring that no two conflicting values can achieve majority simultaneously.

The simplest quorum is a majority of nodes, but more sophisticated schemes — weighted quorums, read/write quorums, and grid quorums — allow systems to trade latency, fault tolerance, and availability against each other. The CAP theorem constraint is visible here: a quorum system that tolerates more failures requires larger quorums, which increases the latency of every operation. The quorum is not merely a voting threshold; it is the system's safety boundary.

See Also

• Majority quorums — the simplest and most common quorum scheme in practice