Learn the vocabulary of requiring a minimum set of nodes to acknowledge a read or write.
0 / 5 completed
1 / 5
At standup, a dev mentions the minimum number of nodes in a distributed system that must acknowledge a read or write for it to be considered successful and consistent. What is this minimum number called?
A quorum is the minimum number of nodes in a distributed system that must acknowledge a read or write for it to be considered successful and consistent, without necessarily requiring every single node to respond. Requiring every node to acknowledge an operation makes the system fragile, since a single unavailable node would block every operation entirely. This quorum threshold is what balances a distributed system's consistency guarantee against its tolerance for a node being temporarily unavailable.
2 / 5
During a design review, the team wants to tune how many nodes must acknowledge a read versus how many must acknowledge a write, trading off consistency strength against latency and availability. Which capability supports this?
Read quorum and write quorum tuning adjusts how many nodes must acknowledge a read versus a write separately, trading off consistency strength against latency and availability depending on how each is configured. Using a single, fixed count for both ignores that a read-heavy or write-heavy workload might benefit from a different balance between the two. This separate tuning is what lets a distributed system's consistency-latency tradeoff be adjusted for a specific workload's needs.
3 / 5
In a code review, a dev notices the read and write quorum sizes are chosen so their combined total guarantees at least one overlapping node between the two sets, ensuring a read always sees the latest acknowledged write. What does this represent?
Quorum overlap guarantees the read quorum's set of nodes intersects with the write quorum's set of nodes for the most recent write, ensuring at least one node in a subsequent read has actually seen that latest acknowledged write. Choosing sizes with no guaranteed overlap risks a read completing entirely from nodes that never received the latest write. This overlap guarantee, often expressed as W plus R being greater than N, is the mathematical foundation of a quorum-based consistency guarantee.
4 / 5
An incident report shows a client read a stale value immediately after a supposedly successful write, because the system's read and write quorum sizes were configured so their combined total didn't guarantee any overlap between the two sets of nodes. What practice would prevent this?
Configuring the read and write quorum sizes so their combined total guarantees at least one overlapping node between the two sets ensures a subsequent read always includes a node that saw the most recent write. Configuring them with no such guarantee risks exactly the stale read this incident describes. This overlap requirement is the core condition that has to hold for a quorum-based system to actually deliver its intended consistency guarantee.
5 / 5
During a PR review, a teammate asks why the team requires a quorum of multiple nodes instead of just reading from or writing to a single node directly. What is the reasoning?
A quorum protects against a single node's failure or staleness by requiring an overlap between the nodes that acknowledged a write and the nodes consulted for a subsequent read. Relying on a single node directly offers no such protection, since that one node being stale or unavailable directly compromises the operation. The tradeoff is the added latency of waiting for multiple nodes to acknowledge an operation instead of just one.