Estimate

Estimating scale helps you choose the right technologies, plan for growth, and communicate your understanding of real-world systems.

✅ What to Estimate?

1. 👥 Users (DAU / MAU)
2. 🔄 Requests / Query per second (RPS / QPS)
3. 📦 Storage
4. 🧮 Bandwidth

👥 1. Estimate Users

Estimate Daily Active Users (DAU) and Monthly Active Users (MAU) based on the app type.

Heuristics:

• Social media: DAU ≈ 10–30% of MAU
• Admin/dashboard apps: DAU ≈ MAU
• Ticketing/event apps: Spiky traffic

Example:
If MAU is 10M, assume 1–3M DAU.

Say in interview:

“Let’s assume 5M monthly active users, and around 1M daily active users.”

🔄 2. Requests Per Second (RPS)

🎯 Formula

RPS ≈ (DAU × Requests per User per Day) ÷ 10^6

Use 10^6 (instead of 86400 seconds in a day) for simplicity in interviews.

🧠 Example

• DAU = 1,000,000
• Avg 5 requests/day/user

RPS = (1M × 5) ÷ 1M ≈ 5 RPS

Advanced: Split by Type

• 70% reads → 3.5 RPS
• 30% writes → 1.5 RPS

“We’ll need to optimize for reads and cache aggressively.”

📦 3. Estimate Storage

🎯 Formula

Storage = Avg object size × Number of objects

Example

• 1 post = 500 bytes
• 10M posts/month = 10^6 posts/month

Storage = 500 bytes × 10^6 = ~5 GB/month

Use 10^6 (instead of 1024 take 10^3 in unit calculation) for simplicity in interviews.

• 1-year retention = 60 GB

“We’ll need ~5 GB/month, or ~60 GB/year for post storage.”

🧮 4. Estimate Bandwidth

🎯 Formula

Bandwidth = RPS × Avg response size

Example

• 60 RPS
• Avg response = 50 KB

Bandwidth = 60 × 50 KB = 3 MB/s

• Daily: 3 MB/s × 86400 = ~250 GB/day

“Our outgoing traffic will be ~250 GB/day, mainly due to read-heavy load.”

📋 Summary Table

💡 Interview Power Move

Say something like:

“Given 1M DAU and ~5 RPS, we expect read-heavy traffic, so we’ll cache results using Redis/CDN. Storage needs are modest (~5 GB/month), and bandwidth can be optimized via compression.”

Tip: Clear, quantitative estimates show you understand real-world system constraints and can communicate trade-offs effectively.