Why Throughput, latency, and availability in HLD? - Purpose & Use Cases

The Big Idea

What if your favorite app suddenly slowed down or stopped working during a big event? Understanding these concepts can prevent that!

The Scenario

Imagine running a busy coffee shop where every order is taken and served by just one person without any tools or system.

Customers line up, wait a long time, and sometimes orders get mixed up or missed.

The Problem

Handling many customers manually is slow and stressful.

Orders pile up, mistakes happen, and some customers leave unhappy because they waited too long or never got their coffee.

The Solution

By understanding throughput, latency, and availability, we can design a system that serves many customers quickly, keeps wait times low, and stays open without interruptions.

This means using multiple baristas, efficient order tracking, and backup plans for busy times or machine failures.

Before vs After

✗ Before

serve_one_customer_at_a_time()
wait_for_order()
make_coffee()
serve_customer()

✓ After

parallel_serve_customers()
optimize_order_flow()
monitor_system_health()

What It Enables

It enables building systems that handle many requests fast, respond quickly, and stay reliable even under pressure.

Real Life Example

Think of a popular online store during a sale: it must process thousands of orders per second (throughput), show pages quickly (low latency), and never crash (high availability).

Key Takeaways

Throughput measures how much work a system can do in a time.

Latency is the delay before a system responds.

Availability means the system is ready and working when needed.