Chain Replication
TechniqueDistributed SystemsSeniorStaffUpdated
Replicas arranged in a chain: writes enter at the head, flow through every link, commit at the tail. Reads served by tail give linearizable results with simple logic.
When to Use Chain Replication
Storage systems wanting linearizable reads without quorum coordination, willing to pay write latency = sum of hops. Simpler failure handling than Paxos.
In Practice
Microsoft Azure Storage (the layered DFS), CRAQ (Chain Replication with Apportioned Queries), Hibari, and FAWN. Used in Azure's stream layer at exabyte scale.
How Chain Replication Comes Up in Interviews
Scenario
Azure-style blob storage wants linearizable reads from a replica without running Paxos or quorum reads on every GET.
Chain replication serves all reads from the tail, which by construction holds only fully committed writes. No quorum math, no leader round-trip per read.
CRAQ extends this to let any node serve reads when there are no in-flight writes, dramatically improving read throughput.