Theerej C

Posted on May 1

Scalability in simple terms - 1

Theerej C · 2026-05-01T12:57:12Z

🚀 Intro So as usual, I was reading something about system design and came across the concept of scalability as an essential concept. So I thought of going a bit deep into it regarding the fact that a A system without scaling is a system without safety So in this blog, I will explain scalability in simple terms which can be understood by anyone. 📈 What is Scalability? So let's look into the talk of the hour — scalability. Basically, scalability is a concept where a system should be able to scale into a bigger one or a smaller one. 🏨 Example Hotel Scenario If there is a hotel which is being operated at a scale of serving 150 persons per hour , during festival season the crowd will require at least that much capacity. But during non-festival days, it requires only 75 servings per hour . What does the hotel do? Splits the hall into two → serves 75 people during non-rush Opens both sections during rush If still not enough → constructs another building nearby Likewise, in software and IT terminology, scalability refers to the concept of scaling a service or hardware capacity bigger or smaller. 🏗️ Horizontal vs Vertical (Wide vs Tall) So consider the same hotel example — the hotel has two options to expand: 🧱 Vertical Scaling (Tall) Expand the same building (more floors / bigger infrastructure) 🌐 Horizontal Scaling (Wide) Open another branch (separate building) 💻 In Servers Increase server power → Vertical Scaling Add more servers → Horizontal Scaling ⚠️ Limitation In vertical scaling, we cannot upgrade forever (we can’t add infinite RAM or CPU to a single machine) 👉 So practically, horizontal scaling is what is used in the current world. ⚖️ Load Balancing (Multi-Serve) Now consider the hotel scales horizontally (good choice). They open another branch in the next street. 🏨 Example Problem: If all online orders go to Hotel 1 , it becomes a choking point Solution: Split orders: Orders 1, 3, 5 → Hotel 1 Orders 2, 4, 6 → Hotel 2 👉 This is exactly the concept of a load balancer 💻 Technical Meaning A load balancer is a service that splits incoming client requests across multiple servers properly. 🔄 Round Robin One of the most commonly used algorithms Server 1 → Server 2 → Server 3 → Server 1 → ... 👉 Requests are distributed in order (circular fashion) 🍪 Sticky Sessions & Cookies (Problem Arises) 🏨 Example Two hotels (scaled horizontally) Two customers: A and B Scenario: Customer A orders → routed to Hotel 1 Next order → routed to Hotel 2 → delay happens 💡 Solution The system remembers where the customer was first served Future requests go to the same place 💻 Technical Explanation First request → load balancer assigns a server Server sends its ID → stored as a cookie on client side Future requests → include cookie → routed to same server 🔁 Alternative Load balancer can map: Client IP → Specific Server 💾 RAID (Redundant Array of Independent Disks) 📖 Example Author Scenario An author writes a book and worries about losing it So every day: Makes a xerox copy Stores it in a bank locker 👉 If something happens → he can recover it 💻 In Computers Data is stored in disks (HDD/SSD) To avoid data loss: Same data is replicated into another disk ❓ Why in Scalability? As systems scale: Multiple failure points exist Data loss risk increases Multiple servers need consistent data 📌 Definition RAID is a concept of using multiple disks to store duplicate copies of data There are many types of RAID, but that is a topic for another day. 🔚 Closing There will be a continuation of this in upcoming blogs, as this is just a small portion of system design. 📚 References Harvard Lecture: https://www.youtube.com/watch?v=-W9F__D3oY4&t=503s

#systemdesign #distributedsystems #beginners #backend

🚀 Intro

So as usual, I was reading something about system design and came across the concept of scalability as an essential concept. So I thought of going a bit deep into it regarding the fact that a

A system without scaling is a system without safety
So in this blog, I will explain scalability in simple terms which can be understood by anyone.

📈 What is Scalability?

So let's look into the talk of the hour — scalability. Basically, scalability is a concept where a system should be able to scale into a bigger one or a smaller one.

🏨 Example

Hotel Scenario

If there is a hotel which is being operated at a scale of serving 150 persons per hour, during festival season the crowd will require at least that much capacity.
But during non-festival days, it requires only 75 servings per hour.

What does the hotel do?

Splits the hall into two → serves 75 people during non-rush
Opens both sections during rush
If still not enough → constructs another building nearby

Likewise, in software and IT terminology, scalability refers to the concept of scaling a service or hardware capacity bigger or smaller.

🏗️ Horizontal vs Vertical (Wide vs Tall)

So consider the same hotel example — the hotel has two options to expand:

🧱 Vertical Scaling (Tall)

Expand the same building (more floors / bigger infrastructure)

🌐 Horizontal Scaling (Wide)

Open another branch (separate building)

💻 In Servers

Increase server power → Vertical Scaling
Add more servers → Horizontal Scaling

⚠️ Limitation

In vertical scaling, we cannot upgrade forever (we can’t add infinite RAM or CPU to a single machine)

👉 So practically, horizontal scaling is what is used in the current world.

⚖️ Load Balancing (Multi-Serve)

Now consider the hotel scales horizontally (good choice).
They open another branch in the next street.

🏨 Example

Problem:

If all online orders go to Hotel 1, it becomes a choking point

Solution:

Split orders:
- Orders 1, 3, 5 → Hotel 1
- Orders 2, 4, 6 → Hotel 2

👉 This is exactly the concept of a load balancer

💻 Technical Meaning

A load balancer is a service that splits incoming client requests across multiple servers properly.

🔄 Round Robin

One of the most commonly used algorithms Server 1 → Server 2 → Server 3 → Server 1 → ...

👉 Requests are distributed in order (circular fashion)

🍪 Sticky Sessions & Cookies (Problem Arises)

🏨 Example

Two hotels (scaled horizontally)
Two customers: A and B

Scenario:

Customer A orders → routed to Hotel 1
Next order → routed to Hotel 2 → delay happens

💡 Solution

The system remembers where the customer was first served
Future requests go to the same place

💻 Technical Explanation

First request → load balancer assigns a server
Server sends its ID → stored as a cookie on client side
Future requests → include cookie → routed to same server

🔁 Alternative

Load balancer can map: Client IP → Specific Server

💾 RAID (Redundant Array of Independent Disks)

📖 Example

Author Scenario

An author writes a book and worries about losing it
So every day:
- Makes a xerox copy
- Stores it in a bank locker

👉 If something happens → he can recover it

💻 In Computers

Data is stored in disks (HDD/SSD)
To avoid data loss:
- Same data is replicated into another disk

❓ Why in Scalability?

As systems scale:
- Multiple failure points exist
- Data loss risk increases
- Multiple servers need consistent data

📌 Definition

RAID is a concept of using multiple disks to store duplicate copies of data

There are many types of RAID, but that is a topic for another day.

🔚 Closing

There will be a continuation of this in upcoming blogs, as this is just a small portion of system design.

📚 References

Harvard Lecture: https://www.youtube.com/watch?v=-W9F__D3oY4&t=503s

Top comments (1)

Naveen Kumar • May 1

In scaling, can we say it like this...? vertical scaling is paying more to the cooks and giving them better tools and bigger equipment so they can work faster and horizontal scaling is adding more cooks..