# Rate Limiting Algorithms, Deep Dive — System Design

Source: https://www.geekswithgeeks.com/en/system-design/sd-rate-limiting-deep

> Token bucket vs leaky bucket vs sliding window, compared.

## Leaky bucket

A **leaky bucket** models requests as water poured into a bucket with a fixed hole at the bottom: it processes (leaks) at a **constant rate** regardless of burst size, queuing or dropping the rest. It smooths traffic; token bucket instead allows bursts up to the cap.

## Sliding window

A fixed window (e.g. 'max 100 req per minute, reset at :00') lets a client burst 100 at 11:59:59 and another 100 at 12:00:01 — 200 in 2 seconds. A **sliding window log or counter** weighs recent sub-windows to smooth this edge effect.

**Quiz:** Which algorithm enforces a strictly constant output rate, smoothing out bursts entirely?

- [ ] Token bucket
- [x] Leaky bucket
- [ ] Fixed window counter

*Answer:* Leaky bucket. Leaky bucket processes at a fixed rate no matter how bursty the input is, unlike token bucket which allows saved-up bursts through.