What is Write Amplification?

Write amplification is a storage effect where the amount of physical data written to disk is larger than the data requested, commonly seen in SSDs and LSM-tree databases.

When to Use

This concept is important in workloads with frequent writes—databases like RocksDB or LevelDB, caching systems, or SSD-heavy infrastructure. Engineers and system designers must consider it when optimizing for durability and performance.

Example

Saving a 1 MB file may cause the SSD to internally write 2 MB due to garbage collection and block erases. That’s a 2× write amplification.

Want to strengthen your system design prep?

Explore Grokking System Design Fundamentals, Grokking the System Design Interview, or practice in Mock Interviews with ex-FAANG engineers.

Why Is It Important

Write amplification reduces SSD lifespan, increases latency, and affects throughput. Managing it is crucial to ensure reliable, long-lived storage systems.

Interview Tips

Define it concisely, explain why it happens (flash memory requires block erases), and mention techniques like over-provisioning or optimized compaction. Framing both the problem and mitigation strategies shows depth.

Trade-offs

Minimizing write amplification often means reserving storage space or performing more background tasks. This balances durability with performance costs.

Pitfalls

A common mistake is assuming write amplification can be eliminated. In reality, it can only be reduced through careful design and workload tuning.