Software Engineering Glossary

Linearizability

Also known as: Strong Consistency Atomic Consistency External Consistency

Definition

Linearizability is the strongest single object consistency guarantee. Every operation appears to take effect instantly at some single point between when it was called and when it returned. Once a write completes, every later read sees that value or a newer one, and everyone agrees on the same order of events, so the system behaves as if there is just one copy of the data.

Popular reads

View All
Payment System Design: Ledger, Idempotency, and Settlement
Payment System Design: Ledger, Idempotency, and Settlement Jul 18, 2026
X Algorithm Explained: How the Open Source Recommendation System Works
X Algorithm Explained: How the Open Source Recommendation System Works Jan 22, 2026
Cursor Skills: How to Create and Use Agent Skills
Cursor Skills: How to Create and Use Agent Skills Jun 23, 2026
Complete Guide to Graph Data Structure: BFS, DFS, Adjacency List vs Matrix
Complete Guide to Graph Data Structure: BFS, DFS, Adjacency List vs Matrix Jan 20, 2026
The Complete HTMX Guide: From Zero to Production
The Complete HTMX Guide: From Zero to Production Dec 22, 2025
Transactional Outbox Pattern: Never Lose an Event Again
Transactional Outbox Pattern: Never Lose an Event Again Apr 07, 2026

Key Takeaways

  • Linearizability makes a distributed system behave like a single machine. There is one order of operations that everyone agrees on.
  • A read is guaranteed to see the most recent completed write. There is no window where a client reads stale data.
  • It is not free. Reads and writes usually go through a leader and a quorum, which costs latency and caps throughput.
  • It is stronger than eventual consistency, which only promises replicas converge later. Use linearizability for metadata and coordination, eventual consistency for scale.

How It Works

  1. Writes go through a consensus protocol so a majority quorum agrees on the order before a write is committed.
  2. Reads are served in a way that reflects all committed writes, for example by routing through the current leader.
  3. To stay correct, the leader confirms it is still the leader, using a lease or a quorum check, before answering a read so a stale leader cannot serve old data.
  4. The result is that the whole cluster looks like one register that applies operations one at a time in real time order.

Where It Is Used

  • etcd serves linearizable reads by default so Kubernetes always sees the latest cluster state.
  • A consistent core like ZooKeeper provides linearizable operations that larger data clusters rely on for coordination.
  • Google Spanner offers external consistency, a form of linearizability across the globe, using TrueTime and Paxos.

Related glossary terms