Capacity Estimation/

Thumb Rules

Lesson overview

Thumb Rules

Key thumb rules for back-of-envelope estimations.

1.Always round numbers to powers of 10 for fast estimation. example : 9.2 million → 10⁷

2.Know Common Powers of 2 (Memory & Storage) example : Unit Power of 2 Approx 1 KB 2¹⁰ ~10³(thousand) 1 MB 2²⁰ ~10⁶(million) 1 GB 2³⁰ ~10⁹(billion) 1 TB 2⁴⁰ ~10¹²(trillion) 1 PB 2⁵⁰ ~10¹⁵(Quadrillion)

3.Always Take the Maximum (Worst Case) When in doubt, design for the upper bound, not the average. examples: Max profile size, not average

4.Include Replication & Overhead Stored data ≠ usable data Add: Replication factor (2x–3x) Backups Example: Raw data = 10 TB Replication (3x) → 30 TB Index + metadata (~20%) → 36 TB total

5.Separate Hot vs Cold Data Not all data needs fast storage. Hot data → SSD / cache Cold data → HDD / cheap object storage

6.Think in Growth Over Time Rule: Always multiply by retention period. Example: 1 TB/day Retention = 1 year → 365 TB/year Add buffer: 20–30% extra for growth

Loading Thumb Rules