System Design Series — Part 25

Imagine you're building an e-commerce application.

The homepage needs to display:

• Product name
• Price
• Product image
• Average rating
• Seller name

Your frontend calls a REST API:

GET /products/101

The server responds with:

• Product description
• SKU
• Weight
• Dimensions
• Stock history
• Created date
• Updated date
• Supplier information
• Warehouse details
• ...and 20 more fields.

But your page only needs five.

The rest of the data travels across the network for no reason.

Now imagine millions of users making the same request every day.

That's a lot of unnecessary data.

This is one of the problems GraphQL was created to solve.

Let's understand it in the simplest way possible.

The Real Problem

Imagine you're building an Instagram profile page.

The UI needs:

• Username
• Profile picture
• Followers count
• Latest 10 posts

With a traditional REST API, you might call:

GET /users/101

GET /users/101/posts

GET /users/101/followers

Multiple requests.

More network traffic.

Higher latency.

Slower page loads.

Can we get everything in one request?

Yes.

That's exactly what GraphQL does.

A Simple Real-World Analogy

Imagine you're ordering food.

REST API

The waiter brings you a fixed meal.

You only wanted rice and paneer.

But you also receive soup, salad, dessert, and drinks.

You paid for—and received—more than you needed.

GraphQL

The waiter asks:

"What exactly would you like?"

You order:

✔ Rice

✔ Paneer

✔ Coke

Nothing more.

Nothing less.

That's GraphQL.

You request exactly the data you need.

What is GraphQL?

GraphQL is a query language for APIs developed by Facebook.

Instead of exposing many endpoints,

GraphQL exposes a single endpoint.

The client decides:

• Which fields it wants
• How much data it needs
• Which related objects should be returned

The server simply delivers exactly that.

How GraphQL Works

Step 1

Client sends a query.

↓

Step 2

GraphQL Server receives it.

↓

Step 3

Resolvers fetch data.

↓

Step 4

Database is queried.

↓

Step 5

Only requested fields are returned.

No unnecessary data.

REST vs GraphQL

REST

Different endpoints.

GET /users

GET /posts

GET /comments

Multiple requests.

Fixed responses.

GraphQL

Single endpoint.

POST /graphql

Client decides:

• Which objects
• Which fields
• Which relationships

One request.

Custom response.

Real-World Example

Suppose you're building a product page.

Frontend only needs:

• Product Name
• Price
• Rating

GraphQL Query:

Product

• Name
• Price
• Rating

Server returns only those fields.

No extra data.

This reduces network usage and improves performance.

Another Example: Netflix

Imagine Netflix loads the home screen.

It needs:

• User profile
• Continue Watching
• Trending Movies
• Recommendations

With REST:

Multiple API calls.

With GraphQL:

One query can fetch everything.

Fewer requests.

Faster loading.

Why Do Companies Use GraphQL?

GraphQL solves two common REST problems.

Over-Fetching

Receiving more data than required.

Example:

Need:

Name

Receive:

Name + 30 unnecessary fields.

Under-Fetching

One API doesn't provide enough data.

Frontend must call multiple APIs.

This increases latency.

GraphQL solves both.

GraphQL Schema

Every GraphQL API defines a schema.

The schema describes:

• Available objects
• Relationships
• Fields
• Operations

Think of it as the contract between frontend and backend.

Queries and Mutations

GraphQL mainly uses two operations.

Query

Retrieve data.

Equivalent to:

GET

Mutation

Create, update, or delete data.

Equivalent to:

POST

PUT

DELETE

Production Architecture

A typical GraphQL architecture looks like this:

Users

↓

CDN

↓

Load Balancer

↓

GraphQL Server

↓

Business Services

↓

Redis Cache

↓

Database

The GraphQL server acts as an orchestration layer between clients and backend services.

Real-World Companies Using GraphQL

Many large companies use GraphQL.

Examples include:

• Facebook
• Instagram
• GitHub
• Shopify
• Pinterest
• Airbnb

It's especially popular for applications with complex user interfaces.

Advantages of GraphQL

✔ Single API endpoint

✔ Client requests exactly the data it needs

✔ Reduces over-fetching

✔ Reduces under-fetching

✔ Faster frontend development

✔ Great for mobile applications

✔ Excellent developer experience

Challenges of GraphQL

More Complex Backend

Resolvers and schema management require additional effort.

Caching Is Harder

REST endpoints are easier to cache.

GraphQL queries are more dynamic.

Query Complexity

Clients can accidentally request huge datasets.

Production systems often enforce:

• Query depth limits
• Rate limiting
• Complexity analysis

Security

Field-level authorization becomes important.

Users should only access data they are allowed to see.

GraphQL vs REST

Choose REST when:

✔ APIs are simple

✔ Resources are well defined

✔ HTTP caching is important

✔ Public APIs need simplicity

Choose GraphQL when:

✔ Frontends need flexible data

✔ Mobile apps must minimize bandwidth

✔ Complex dashboards require multiple resources

✔ Many client applications consume the same backend

There is no universally "better" choice.

Choose the right tool for the problem.

Common Developer Mistakes

Returning Everything

GraphQL allows flexibility.

Don't ignore performance.

Ignoring Query Limits

Large nested queries can overload servers.

Poor Schema Design

A poorly designed schema becomes difficult to maintain.

Forgetting Authorization

Every field should be secured appropriately.

Production-Level Insight

One misconception is:

"GraphQL replaces REST."

It doesn't.

Many companies successfully use both.

For example:

REST APIs for public integrations.

GraphQL for web and mobile applications.

Good architects choose the technology that best fits the business requirements—not the latest trend.

Interview Tip

A common System Design interview question is:

"GraphQL vs REST: Which one would you choose?"

A strong answer should discuss:

• Over-fetching
• Under-fetching
• Single endpoint
• Client-driven queries
• Caching
• Complexity
• Performance trade-offs

Interviewers are looking for architectural reasoning, not technology preference.

Key Takeaways

✔ GraphQL is a query language for APIs

✔ Clients request exactly the data they need

✔ It reduces over-fetching and under-fetching

✔ A single endpoint can serve multiple use cases

✔ GraphQL is ideal for complex frontends and mobile apps

✔ REST and GraphQL can coexist in the same system

✔ Choose the technology based on your application's requirements

One of the biggest lessons in System Design is this:

The best API isn't the one with the most features.

It's the one that delivers exactly what the client needs—nothing more, nothing less.

That's why GraphQL has become an important tool for building modern, scalable applications.

This is Part 25 of the System Design Simplified series.

Next Article: Part 26 — API Versioning Explained Simply

If this article helped you understand GraphQL better, consider sharing it with fellow developers.

#SystemDesign #GraphQL #SoftwareArchitecture #BackendDevelopment #APIDesign #RESTAPI #DistributedSystems #Microservices #Scalability #CloudComputing #SoftwareEngineering #SystemDesignInterview #BackendEngineer #Programming #TechArchitecture