Distributed Systems also: ISR

In-Sync Replicas

Replicas caught up to the leader within a time bound.

In plain terms

Kafka. Falling behind for >30s evicts you; catching back up readmits. Producer durability scales with ISR size.

Origin

A core Kafka concept since the original 2011 LinkedIn paper. ISR membership is dynamic — falling behind the leader by more than replica.lag.time.max.ms drops you out; catching back up re-admits you.

Where it shows up in production

Apache Kafka min.insync.replicas controls durability — produce can require N ISR members before acking.
RocketMQ Similar ISR-style replication adopted from Kafka's model.

On Semicolony

Guide kafka

Sources & further reading

Docs Kreps et al. — Kafka design (LinkedIn)

Found this useful?