Read and Write Quorums
You'll learn: How using separate quorum sizes for reads (R) and writes (W) gives you flexibility in balancing consistency, performance, and availability.
The Core Idea
A quorum is the minimum number of nodes that must agree on a read or write operation. Instead of requiring all replicas to participate (which would hurt availability), or just one replica (which risks inconsistency), quorum systems let you tune how many nodes must acknowledge reads (R) and writes (W).
The magic happens when you set R and W independently based on your application's needs.
The Quorum Rule
For strong consistency, you need: R + W > N
Where N is the total number of replicas. This overlap ensures read and write sets intersect—at least one node in your read quorum will have the latest write.
Tuning for Different Goals
Write-heavy workload? Use W=1, R=N. Writes are fast (only one node must acknowledge), but reads must check all nodes to find the latest value.
Read-heavy workload? Use W=N, R=1. Every write waits for all nodes, but reads are lightning-fast from any single node.
Balanced? Use W=2, R=2 (with N=3). You get reasonable performance for both operations while tolerating one node failure.
The Trade-off
Lower W means faster writes but slower reads (and vice versa). You can sacrifice strong consistency entirely by violating R + W > N—for example, W=1, R=1 gives speed but allows stale reads.
Key Takeaway: Read and write quorums let you independently tune the cost and guarantees of reads versus writes, giving you control over the consistency-performance trade-off for your specific workload.