In system design, the term “back-of-the-envelope” refers to a quick, rough estimation or calculation done informally, usually without detailed tools or exact data, to see if an idea or approach is feasible.
It’s called that way because people often literally scribble on a napkin or small piece of paper — you don’t need a full spreadsheet, simulator, or diagram to get the gist.
🔹 Key Points
-
Purpose:
- Quickly validate assumptions
- Check if a system can handle expected load
- Estimate rough costs or resources needed
-
Not Precise:
- These are approximations, not exact numbers
- They help identify bottlenecks or feasibility before diving deep
-
Common Use Cases in System Design Interviews:
- Estimating traffic/load: How many requests per second will the system handle?
- Estimating storage: How much disk space is needed for user data?
- Estimating bandwidth: How much network capacity is required?
- Estimating compute resources: How many servers or CPUs do you need?
🔹 Example
Suppose you are designing a URL shortening service:
- Expected: 1 billion URLs in 10 years
- Average URL length: 100 bytes
- Extra metadata: 50 bytes
Rough calculation (back-of-the-envelope):
Total storage = 1B * (100 + 50) bytes
= 150B bytes
≈ 150 GB
✅ Conclusion: Storage requirement is manageable. ❌ No need to calculate exact overheads or indexing yet — this is just a rough check.
🔹 Why It’s Useful in Interviews
- Shows you can think about scalability and resource requirements quickly
- Helps guide the design: e.g., whether you need sharding, caching, or distributed databases
- Demonstrates practical engineering intuition
🔹 Analogy
It’s like estimating if a car can reach a destination without calculating exact fuel consumption — you just check if the trip is roughly possible.