How to Calculate Throughput & Database Size

Alright, so picture this: You walk into that interview feeling pretty damn good. You've practiced system design questions until you could solve the URL shortener problem in your sleep. You're ready.

Then you hit the Non-functional Requirements section. The interviewer leans back and casually asks, "So, what kind of storage are we talking about for your database? And what's the throughput looking like?"

And just like that, everything falls apart. Now you're stuck doing mental gymnastics with math you haven't touched since college, trying to multiply huge numbers in your head or scribbling frantically on a whiteboard—when in real life you'd just pull out a calculator like a normal human being.

Good news though: I put together a cheat sheet that'll save your ass in these situations.

Here's the thing—in interviews, they only care about order of magnitude. Not exact numbers. So we can estimate.

The one trick you absolutely need to remember:

2^10 ≈ 10^3

This is your golden ticket. Memorize it.

Let's work through this. Say each URL is 8 characters long, and we're only using lowercase letters. That's 26 letters in the alphabet, so we're looking at 26^8 possible combinations.

Step 1: Estimate to the nearest power of 2 (2, 4, 8, 16, 32, 64, 128)

26^8 ≈ (32)^8

Step 2: Rewrite as 2^x

(32)^8 = (2^5)^8

Step 3: Multiply the powers

(2^5)^8 = 2^40

Step 4: Break down into a multiple of 10

2^40 = (2^10)^4

(We can see that 40 = 10 × 4)

Step 5: Swap using our cheat code (2^10 ≈ 10^3)

(2^10)^4 ≈ (10^3)^4

Step 6: Simplify

(10^3)^4 = 10^12

Each character in UTF-8 is 1 byte (8 bits, but we never use bits anyways).

Quick storage reference to memorize:

• 1 KB = 10³ bytes (thousand)
• 1 MB = 10⁶ bytes (million)
• 1 GB = 10⁹ bytes (billion)
• 1 TB = 10¹² bytes (trillion)

Result: We need about 1 Terabyte of storage.

Step 1: Start with our total bytes

10^12 bytes total

Step 2: Divide by bytes per URL (round 8 → 10 for easier math)

10^12 ÷ 8 bytes per URL ≈ 10^12 ÷ 10

Step 3: Simplify

10^12 ÷ 10 = 10^11 URLs

That's a hell of a lot of URLs.

Say we've got 10 million users per day, and each user makes about 6 requests on average. That's 60 million requests per day.

There are:

24 hours × 60 min × 60 sec = 86,400 seconds/day
≈ 8.6 × 10^4
≈ 10^5 seconds per day

(We're rounding for estimation)

60 × 10^6 requests ÷ 10^5 seconds = 600 requests/second

Now if we've got a 5:1 read-to-write ratio:

• 500 req/sec for reads
• 100 req/sec for writes

2^10 ≈ 10^3

• 2, 4, 8, 16, 32, 64, 128
• Always round to the nearest power of 2

• Seconds per day: ~10^5 (actually 86,400)
• Seconds per month: ~2.6 × 10^6
• Seconds per year: ~3.2 × 10^7

And there you have it. That's the framework. Memorize that 2^10 ≈ 10^3 trick and you'll be able to estimate your way through any storage or throughput question they throw at you.

The key takeaways:

1. Always estimate to the nearest power of 2
2. Break down exponents into multiples of 10
3. Use the 2^10 ≈ 10^3 conversion
4. Round for easier mental math (8 → 10)
5. Focus on order of magnitude, not exact numbers