पाठ 1 / 30

Why Engineers Estimate

यह पाठ अभी हिंदी में अनुवादित नहीं है, इसलिए अंग्रेज़ी में दिखाया जा रहा है।

Explain what back-of-the-envelope estimation is for and why an answer within 2-10x is useful.

लेखक: · 15 साल से वेब डेवलपर

Estimation is a decision tool

A back-of-the-envelope estimate is a quick calculation, done with round numbers, that tells you roughly how much traffic, storage, compute or money a system will need. The goal is not precision. The goal is to land within the right order of magnitude, because most design decisions change only when a number crosses a power of ten. Ten requests per second fits on one small server with a single database; ten thousand per second needs load balancing, caching and probably replicas; ten million needs sharding and a very different team. If your estimate says 2,000 and the truth is 5,000, you still pick the same architecture. If it says 2,000 and the truth is 2,000,000, you built the wrong thing. Estimation also exposes which resource will run out first, so you spend effort on the real bottleneck instead of the one you guessed.

The estimation loop

Every estimate in this course follows the same four steps: clarify requirements, state assumptions, calculate, then sanity-check and revisit the assumptions if the answer looks wrong.

Four connected stages: clarify requirements, state assumptions, calculate, sanity-check.
Figure 1.1: clarify requirements -> state assumptions -> calculate -> sanity-check

Planning a wedding caterer order

Nobody counts exact guests before booking a caterer. You say roughly 400 people, about 2 rotis each plus buffer, so order around 1,000. Being off by 50 rotis is fine; ordering 100 or 10,000 is a disaster. Capacity estimation is the same habit applied to servers and storage.

Where you will use this

You will meet estimation in system design interviews, in design reviews before a launch, when choosing a database or instance size, and when a manager asks what a feature will cost per month. The same small set of techniques covers all of them.

त्वरित जाँच: Why is an estimate that is 2x off usually acceptable?

  • Because cloud providers refund overestimates automatically
  • Because architecture choices usually change only when a number moves by an order of magnitude
  • Because estimates are never checked after launch
  • Because servers always run at 50% utilisation
Answer

Because architecture choices usually change only when a number moves by an order of magnitude — Designs differ mainly between 10, 10,000 and 10 million requests per second. A 2x error rarely changes the design, while a 100x error does.