Latency Numbers Every Developer Should Know
Practical intuition for network and storage costs.
Performance1 minInspired by Distributed systems fundamentals
Why this matters
Most architecture mistakes come from bad latency intuition.
Typical hierarchy
- CPU cache: nanoseconds
- RAM: tens to hundreds of nanoseconds
- SSD: microseconds to milliseconds
- Network: milliseconds
Design implication
Batch round trips and avoid synchronous chatty boundaries. Why Redis Is Fast is a concrete example of designing around those costs.