#system-design
-
Latency Numbers and Throughput Ceilings: The Constants Behind Capacity Estimation
Part 2 of the napkin-math series — the latency numbers every programmer should know, from L1 cache to cross-region round trips, plus the per-node throughput ceilings of Postgres, Redis, Kafka, and WebSocket servers, and how to derive ceilings you never memorized.
-
Back-of-the-Envelope Estimation for System Design: Load, Units, and the Two Ceilings
A practical guide to capacity estimation — turning "200,000 devices sending every 4 seconds" into QPS, bandwidth, and storage verdicts, with interactive calculators for unit conversion, bits vs bytes, peak vs average traffic, and CPU-bound vs network-bound systems.