17. System Design Case Study 1: URL Shortener (TinyURL / Bitly) at 100M Daily Requests • High-Level System Design (HLD): Distributed Systems at Scale • Rishabh Ki Kaksha

*   **Scale:** 100 Million new URLs created / day; 10:1 Read-to-Write ratio (1 Billion reads/day).
*   **Write QPS:** $\frac{100 \times 10^6}{86400} \approx 1,160 \text{ reqs/sec}$.
*   **Read QPS:** $\approx 11,600 \text{ reqs/sec}$ (Peak: 25,000 QPS).
*   **Storage (5 Years):** $100\text{M} \times 365 \times 5 \times 500\text{ bytes} \approx 91 \text{ TB}$.

---

### 2. Base62 Encoding Architecture 🔤

Characters: `[a-z, A-Z, 0-9]` (Total 62 characters).
$$62^7 \approx 3.5 \text{ Trillion Unique URLs with a 7-character string!}$$

```
[ Unique 64-bit ID from Snowflake (e.g. 10023412) ] ➔ [ Base62 Encoder ] ➔ "aZ9kX1w" (7 Chars)
```

---

### 3. End-to-End High-Level Architecture Diagram

```
[ Client ] ➔ [ Global Anycast CDN / Cloudflare ]
                   ⬇
           [ Nginx Load Balancers ]
                   ⬇
           [ URL Shortener Service (Go/Node.js) ]
              ├── 1. Read: Check Redis Cluster (80-20 Pareto Rule Cache) ➔ Hit? Return 301/302!
              └── 2. Miss: Query Distributed NoSQL (MongoDB / Cassandra)
                   └── Write: Snowflake ID Gen ➔ Base62 ➔ Save in DB & Redis
```

---

### 4. HTTP 301 vs 302 Redirect Tradeoff ⚖️

*   **301 Permanent Redirect:** Browser client side par redirect cache kar leta hai (Origin server load zero hota hai, lekin analytics click tracking impossible ho jaati hai!).
*   **302 Temporary Redirect (Industry Standard):** Har click server se guzarta hai (Perfect real-time analytics & telemetry tracking).

---

### 5. Interview Ke 3 Sawal Is Topic Se

1. **Custom Short Aliases (e.g. `tiny.url/rishabh`) collision kaise handle karte hain?** -> Unique constraint in DB; agar alias taken ho toh user ko 409 Conflict return karo.
2. **Expired URLs ko cleanup karne ka best mechanism kya hai?** -> Lazy deletion on access + TTL index background cleanup script off-peak hours me.
3. **80-20 Pareto rule se Redis Cache RAM size kitni compute hogi?** -> Daily 1 Billion reads ka 20% hot data = $200\text{M} \times 500\text{ bytes} \approx 100 \text{ GB RAM}$ in Redis Cluster.

---
**Next Topic:** 18. System Design Case Study 2: Distributed Rate Limiter & Web Application Firewall

Loading lesson content...

Python 3.12

Full IDE

Live Execution Terminal

Click "Execute Live" to run code.
                

Lesson Mastery AI Quiz

AI Doubt Solver

CERTIFICATE OF COMPLETION

Rishabh Gehlot

PHP 8+ & Laravel 11 Master Track