Eventual Consistency
Also known as:
Optimistic Replication
Definition
Eventual consistency is a promise that, if no new writes happen, all replicas will converge on the same value given enough time. Reads in between can return stale or out of order data. The system promises that updates will reach every replica eventually through gossip, anti entropy, or read repair.
Popular reads
View All
Cursor Skills: How to Create and Use Agent Skills
Jun 23, 2026
Payment System Design: Ledger, Idempotency, and Settlement
Jul 18, 2026
X Algorithm Explained: How the Open Source Recommendation System Works
Jan 22, 2026
Complete Guide to Graph Data Structure: BFS, DFS, Adjacency List vs Matrix
Jan 20, 2026
The Complete HTMX Guide: From Zero to Production
Dec 22, 2025
Transactional Outbox Pattern: Never Lose an Event Again
Apr 07, 2026
Key Takeaways
- Eventual consistency trades immediate global agreement for high availability and low latency, especially across regions.
- Stale reads, lost updates, and out of order observations are normal. The application has to handle them.
- You need a way to resolve conflicts. Common options are last writer wins, vector clocks, CRDTs, or app specific merge logic.
- Despite the name, well tuned systems usually converge in milliseconds. The term describes the worst case, not normal latency.
How It Works
- A write goes to one replica and is acknowledged right away.
- Background work like gossip, anti entropy, or read repair pushes the update to other replicas.
- Conflicting writes are detected with timestamps, vector clocks, or version vectors.
- Replicas merge concurrent writes with a fixed rule so every node ends up with the same final value.
Where It Is Used
- DynamoDB defaults to eventual consistency. Strongly consistent reads cost extra.
- Cassandra has tunable consistency, with eventual being the cheapest and most available setting.
- DNS, S3 in its early years, and Git itself are all eventually consistent in real life.