Chapter 10 — System Design Interview Questions (40+ Questions)

System Design interviews evaluate your ability to design large-scale, reliable, maintainable software systems.

For Senior Full Stack / Team Lead roles, interviewers expect you to discuss:

• scalability
• availability
• performance
• security
• data storage
• architecture trade-offs

1. How do you approach a system design interview?

Interview Answer

I start by understanding requirements, defining the system boundaries, estimating scale, and then designing the architecture.

Step 1 — Clarify Requirements

Ask:

Functional requirements

What should the system do?

Example:

For an online shopping system:

• users browse products
• add items to cart
• place orders
• make payments

Non-functional requirements

How should the system behave?

Examples:

• response time
• availability
• scalability
• security

Step 2 — Estimate Scale

Understand:

• number of users
• requests per second
• data size
• traffic patterns

Example:

Users:

10 million

Requests:

50,000 per second

Step 3 — Design High-Level Architecture

Example:

Users

 |

Load Balancer

 |

Web/API Servers

 |

Business Services

 |

Database

Step 4 — Identify Bottlenecks

Ask:

• Where can it fail?
• What becomes slow under load?
• How do we scale?

Senior Answer

I approach system design by first clarifying requirements, estimating scale, designing a high-level architecture, then refining areas such as data storage, scalability, reliability, and security.

2. What are the main characteristics of a scalable system?

Interview Answer

A scalable system can handle increasing workload by adding resources without significant performance degradation.

Main characteristics:

1. Horizontal Scaling

Adding more servers.

Example:

Before:

Application Server

After:

       Load Balancer

       /      |      \

    Server Server Server

2. Stateless Services

Servers should not store user session data locally.

Bad:

User Session

     |

Server A Memory

Problem:

If Server A fails:

Session Lost

Better:

Server

 |

Redis / Database

3. Caching

Reduce database load.

Example:

User Request

 |

Cache

 |

Database (only if needed)

4. Database Scaling

Options:

• indexing
• replication
• partitioning
• sharding

5. Asynchronous Processing

Move slow operations to background workers.

Example:

Instead of:

Upload File

 |

Process Immediately

 |

Return Response

Use:

Upload File

 |

Queue

 |

Background Worker

Senior Answer

A scalable system uses stateless services, horizontal scaling, caching, asynchronous processing, and scalable data storage strategies.

3. Explain the difference between vertical and horizontal scaling.

Interview Answer

Scaling can be achieved by increasing the capacity of one machine or adding more machines.

Vertical Scaling (Scale Up)

Increase server power.

Example:

Before:

4 CPU

16 GB RAM

After:

32 CPU

128 GB RAM

Advantages:

• simple
• fewer application changes

Disadvantages:

• hardware limits
• expensive
• single point of failure

Horizontal Scaling (Scale Out)

Add more servers.

Example:

Before:

One Server

After:

       Load Balancer

       /       |       \

    Server   Server   Server

Advantages:

• better availability
• unlimited growth potential
• fault tolerant

Disadvantages:

• more complexity
• requires distributed architecture

Senior Answer

Vertical scaling increases the capacity of an existing server, while horizontal scaling adds more servers. Modern cloud applications usually prefer horizontal scaling because it improves availability and elasticity.

4. What is load balancing?

Interview Answer

A load balancer distributes incoming traffic across multiple servers to improve availability and performance.

Architecture:

            Users

               |

        Load Balancer

          /      |      \

      Server  Server  Server

Responsibilities

Traffic Distribution

Example:

Request 1 → Server A

Request 2 → Server B

Request 3 → Server C

Health Checks

The load balancer removes unhealthy servers.

Example:

Server A ❌

Removed from pool

SSL Termination

The load balancer can handle HTTPS encryption.

Client

 HTTPS

 |

Load Balancer

 HTTP

 |

Application

Load Balancing Algorithms

Round Robin

Requests distributed sequentially.

A → B → C → A

Least Connections

Send traffic to server with fewer active requests.

IP Hash

Same client goes to same server.

Examples

Cloud:

• AWS Elastic Load Balancer
• Azure Load Balancer
• Azure Application Gateway

Senior Answer

A load balancer improves system availability by distributing traffic across healthy servers and providing features such as health monitoring, SSL termination, and traffic routing.

5. What is caching and why is it important?

Interview Answer

Caching stores frequently accessed data in faster storage to reduce latency and database load.

Without Cache:

User Request

 |

Application

 |

Database

 |

Response

With Cache:

User Request

 |

Cache

 |

Response

Types of Cache

1. Client-Side Cache

Stored in browser.

Examples:

• images
• JavaScript files
• CSS

2. Application Cache

Stored in application memory.

Example:

.NET MemoryCache

3. Distributed Cache

Shared between servers.

Examples:

• Redis
• Memcached

Architecture:

Server A

   |

Server B

   |

 Redis Cache

Cache Strategies

Cache Aside

Application checks cache first.

Flow:

Request

 |

Cache

 |

Miss

 |

Database

 |

Update Cache

Write Through

Data written to cache and database together.

Write Behind

Write to cache first, database later.

Senior Answer

Caching improves performance by reducing database access and response times. In distributed systems I usually prefer a distributed cache such as Redis.

6. What is Redis and when would you use it?

Interview Answer

Redis is an in-memory data store commonly used as a cache, session store, message broker, and fast database.

Architecture:

Application

     |

 Redis

     |

Database

Common Uses

1. Application Cache

Example:

Product catalog:

Product Data

 |

Redis

 |

Database

2. Session Storage

Example:

User Login

 |

Redis Session

Useful when multiple servers exist.

3. Distributed Locking

Example:

Prevent two processes updating the same resource.

4. Real-Time Data

Examples:

• leaderboards
• counters
• notifications

.NET Example

IDistributedCache cache;

Senior Answer

Redis is useful for low-latency data access scenarios such as caching, session management, distributed locks, and real-time counters.

7. What is database replication?

Interview Answer

Database replication creates copies of data on multiple database servers.

Architecture:

             Application

                   |

             Primary DB

              /      \

       Replica DB   Replica DB

Benefits

Read Scaling

Heavy reads go to replicas.

Example:

Write

 |

Primary DB

Read

 |

Replica DB

High Availability

If primary fails:

Replica

 |

Become Primary

Types

Synchronous Replication

Data copied immediately.

Advantages:

• no data loss

Disadvantages:

• slower

Asynchronous Replication

Copy happens later.

Advantages:

• faster

Disadvantage:

• possible data loss

Examples

AWS:

• RDS Read Replicas

Azure:

• SQL Database Replication

Senior Answer

Database replication improves availability and read scalability by maintaining synchronized copies of data across multiple database instances.

8. What is database partitioning?

Interview Answer

Database partitioning divides a large table into smaller logical pieces while keeping it as one database object.

Example:

Large Orders table:

Orders

100 million rows

Partition by year:

Orders_2024

Orders_2025

Orders_2026

Benefits

• faster queries
• easier maintenance
• better data management

Types

Horizontal Partitioning

Split rows.

Example:

Customer ID

1-1M

1M-2M

Vertical Partitioning

Split columns.

Example:

Customer table:

Before:

Customer

ID

Name

Address

Photo

After:

Customer

ID

Name

CustomerDetails

Address

Photo

Senior Answer

Partitioning improves database performance and manageability by dividing large datasets into smaller, more efficient structures.

9. What is database sharding?

Interview Answer

Sharding is a database scaling technique where data is distributed across multiple independent databases.

Example:

Before:

One Database

1 billion users

After:

Shard 1

Users 1-10M

Shard 2

Users 10M-20M

Shard 3

Users 20M-30M

Sharding Key

Determines where data goes.

Example:

UserId % 3

Result:

User 1 → Shard 1

User 2 → Shard 2

User 3 → Shard 3

Advantages

• unlimited scaling
• distributed load

Challenges

• joins become difficult
• data balancing
• transactions across shards

Senior Answer

Sharding horizontally distributes data across multiple database instances. It provides massive scalability but introduces complexity around transactions and data distribution.

10. What is a CDN?

Interview Answer

A Content Delivery Network (CDN) is a distributed network of servers that delivers static content from locations closer to users.

Without CDN:

User Canada

 |

Server USA

 |

Image

With CDN:

User Canada

 |

CDN Canada Edge Server

 |

Image

CDN Stores

• images
• videos
• JavaScript files
• CSS
• static files

Benefits

• lower latency
• reduced server load
• better global performance

Examples:

• CloudFront (AWS)
• Azure Front Door
• Cloudflare

Senior Answer

A CDN improves performance by caching static content closer to users and reducing latency between clients and origin servers.

11. What is an API Gateway?

Interview Answer

An API Gateway is a single entry point between clients and backend services.

It handles:

• routing
• authentication
• authorization
• rate limiting
• logging
• request transformation

Without API Gateway:

id="api1"

Client

  |

  +---- Service A

  |

  +---- Service B

  |

  +---- Service C

Problems:

• client knows all services
• duplicated security logic
• difficult maintenance

With API Gateway:

id="api2"

             Client

                |

          API Gateway

        /       |       \

    Service A Service B Service C

Responsibilities

Routing

Example:

/api/orders

       |

Order Service

/api/users

       |

User Service

Authentication

Example:

Request

 |

Validate JWT Token

 |

Forward Request

Rate Limiting

Example:

User

 |

100 requests/minute allowed

 |

Block excess requests

Response Aggregation

Example:

Mobile app needs:

• user information
• orders
• recommendations

Instead of:

Mobile

 |

 |

Service A

 |

Service B

 |

Service C

Gateway combines responses:

Mobile

 |

API Gateway

 |

Combined Response

Examples

AWS:

• API Gateway

Azure:

• API Management

Enterprise:

• Kong
• Ocelot (.NET)

Senior Answer

An API Gateway provides a centralized entry point for client applications and handles cross-cutting concerns such as authentication, routing, rate limiting, and monitoring.

12. What is microservice architecture?

Interview Answer

Microservice architecture divides an application into small, independently deployable services that communicate through APIs or messaging systems.

Monolith:

id="mono1"

Application

+----------------+

| User Module    |

| Order Module   |

| Payment Module |

| Reporting      |

+----------------+

Microservices:

id="micro1"

        API Gateway

             |

+-------------+-------------+

User       Order       Payment

Service    Service     Service

+-------------+-------------+

             |

        Databases

Characteristics

Independent Deployment

Each service can be deployed separately.

Example:

Order Service v2

without deploying

Payment Service

Independent Scaling

Example:

Order Service

100 servers

User Service

10 servers

Separate Data Ownership

Example:

Order Service

Order Database

User Service

User Database

Benefits

• scalability
• independent teams
• fault isolation
• technology flexibility

Challenges

• distributed complexity
• network failures
• monitoring
• data consistency

Senior Answer

Microservices allow teams to build independently deployable services with clear boundaries. They improve scalability and team autonomy but require strong practices around communication, monitoring, and data management.

13. How do microservices communicate?

Interview Answer

Microservices communicate using synchronous or asynchronous communication.

1. Synchronous Communication

Usually HTTP/REST or gRPC.

Example:

id="sync1"

Order Service

      |

HTTP Request

      |

Payment Service

Example:

Create order:

Order Service

 |

Payment API

 |

Payment Result

Advantages:

• simple
• immediate response

Disadvantages:

• tight coupling
• failures propagate

2. Asynchronous Communication

Using message brokers.

Example:

id="async1"

Order Service

      |

Message Queue

      |

Notification Service

Examples:

• RabbitMQ
• Kafka
• AWS SQS
• Azure Service Bus

Advantages:

• loose coupling
• better resilience
• handles traffic spikes

Senior Answer

I choose synchronous communication when immediate responses are required and asynchronous messaging when I need loose coupling, reliability, and independent processing.

14. What is a message queue?

Interview Answer

A message queue is a system that allows applications to communicate asynchronously by sending and receiving messages.

Architecture:

id="queue1"

Producer

   |

Message Queue

   |

Consumer

Example:

Order processing:

Customer places order

        |

Order Service

        |

Queue

        |

Inventory Service

        |

Email Service

Benefits

Decoupling

Services do not directly depend on each other.

Reliability

Messages remain available if consumer is temporarily down.

Load Handling

Queue absorbs traffic spikes.

Example:

10,000 requests

        |

Queue

        |

Workers process gradually

Examples

AWS:

• SQS

Azure:

• Service Bus

Open Source:

• RabbitMQ

Senior Answer

Message queues improve system reliability by allowing asynchronous processing and reducing direct dependencies between services.

15. Explain RabbitMQ.

Interview Answer

RabbitMQ is a message broker that implements Advanced Message Queuing Protocol (AMQP).

It routes messages between producers and consumers.

Architecture:

id="rabbit1"

Producer

   |

Exchange

   |

Queue

   |

Consumer

Components

Producer

Creates messages.

Example:

Order Created Event

Exchange

Receives messages and routes them.

Types:

• Direct
• Topic
• Fanout
• Headers

Queue

Stores messages.

Consumer

Processes messages.

Example:

Order system:

Order Service

creates:

OrderCreated Event

        |

RabbitMQ

        |

Email Service

Features

• acknowledgments
• retries
• dead-letter queues
• routing rules

Senior Answer

RabbitMQ is suitable for reliable asynchronous communication where message routing, delivery guarantees, and task processing are important.

16. Kafka vs RabbitMQ. What is the difference?

Interview Answer

Both are messaging systems, but they are designed for different scenarios.

Kafka:

id="kafka1"

Producer

 |

Topic

 |

Consumers

Messages remain:

Event 1

Event 2

Event 3

RabbitMQ:

id="rabbit2"

Producer

 |

Queue

 |

Consumer

After processing:

Message Removed

Kafka Use Cases

• financial transactions
• analytics pipelines
• event sourcing
• IoT data

RabbitMQ Use Cases

• background jobs
• notifications
• order processing

Senior Answer

Kafka is better for high-volume event streaming and data pipelines, while RabbitMQ is often better for traditional asynchronous task processing and message routing.

17. What is event-driven architecture?

Interview Answer

Event-driven architecture is a design pattern where services communicate through events instead of direct calls.

Traditional:

id="event1"

Order Service

      |

Payment Service

      |

Notification Service

Services are coupled.

Event-driven:

id="event2"

Order Service

      |

OrderCreated Event

      |

+-------------+-------------+

Payment     Notification   Inventory

Example Event:

{

 "eventType":"OrderCreated",

 "orderId":123,

 "timestamp":"2026-07-25"

}

Benefits

• loose coupling
• scalability
• easier integration

Challenges

• debugging complexity
• eventual consistency
• event versioning

Senior Answer

Event-driven architecture allows services to react to business events independently, improving scalability and reducing coupling between components.

18. What is eventual consistency?

Interview Answer

Eventual consistency means that distributed systems may temporarily have different data states, but eventually all replicas become consistent.

Example:

Bank transfer:

Account A

-$100

        |

Event

        |

Account B

+$100

There may be a short delay before both systems reflect the final state.

Strong consistency:

Write

 |

All systems updated immediately

Eventual consistency:

Write

 |

Update asynchronously

 |

Eventually synchronized

Common in:

• microservices
• distributed databases
• cloud systems

Example

Order system:

Order Service

Order Created

        |

Message Queue

        |

Inventory Updated

Inventory may update a few seconds later.

Senior Answer

Eventual consistency is a trade-off used in distributed systems to achieve scalability and availability while allowing temporary data differences between services.

19. What is rate limiting?

Interview Answer

Rate limiting controls how many requests a client can make within a specific time period.

Purpose:

• prevent abuse
• protect services
• control costs
• avoid overload

Example:

User:

100 API calls/minute

Allowed:

100

Blocked:

101+

Architecture:

id="rate1"

Client

 |

API Gateway

 |

Rate Limiter

 |

Backend

Algorithms

Fixed Window

Example:

100 requests per minute

Sliding Window

More accurate:

Last 60 seconds

Token Bucket

Tokens are generated over time.

Example:

Bucket:

100 tokens

Each request:

Consumes 1 token

Implementation

Possible locations:

• API Gateway
• Redis
• Load Balancer

Senior Answer

Rate limiting protects APIs from excessive usage and is usually implemented at the gateway layer using techniques such as token bucket or sliding window algorithms.

20. Design a notification system.

Interview Answer

A notification system should support sending:

• email
• SMS
• push notifications
• in-app messages

while handling high volume reliably.

High-Level Design

id="notif1"

Application

     |

Notification Service

     |

Message Queue

     |

+-------------+-------------+

Email       SMS        Push

Service     Service    Service

Flow

Example:

User receives order confirmation.

Order Created

      |

Notification Event

      |

Queue

      |

Email Worker

      |

Send Email

Components

Notification API

Receives requests.

Queue

Handles spikes.

Workers

Process messages.

Templates

Store message formats.

Example:

Order {id} shipped

User Preferences

Store:

• email enabled
• SMS enabled
• language

Scaling

Use:

• multiple workers
• queues
• retries
• dead-letter queues

Senior Answer

I would design a notification system using asynchronous processing with queues, independent notification workers, retry mechanisms, and user preference management to support scalability and reliability.

21. Design a URL Shortener (like Bitly)

Interview Answer

A URL shortener converts long URLs into short unique URLs and redirects users back to the original URL.

Examples:

Long URL:

https://www.example.com/products/category/item?id=123456

Short URL:

https://short.ly/aB91x

Requirements

Functional

• Create short URL
• Redirect short URL
• Track clicks
• Expiration support

Non-functional

• Low latency
• High availability
• Scalable storage

High-Level Architecture

            User

              |

        Load Balancer

              |

        URL Service

              |

       +--------------+

       |              |

   Database       Cache

Data Model

Table:

UrlMapping

Id

ShortCode

OriginalUrl

CreatedDate

ExpirationDate

ClickCount

Example:

Creating Short URL

Algorithm:

1. Generate unique ID
2. Convert ID to Base62
3. Store mapping

Example:

Database ID:

123456

Base62:

aB91x

Redirect Flow

User opens:

short.ly/aB91x

Flow:

Browser

 |

URL Service

 |

Cache

 |

Database

 |

Original URL

Scaling

Use:

• Redis cache for popular URLs
• database indexing on ShortCode
• CDN for global access

Senior Answer

I would design a URL shortener using a stateless URL service, a distributed database for mappings, Redis caching for frequent redirects, and a Base62 encoding strategy to generate compact unique identifiers.

22. Design a File Storage System (like Google Drive)

Interview Answer

A file storage system allows users to upload, store, retrieve, and share files.

Requirements

Functional:

• Upload files
• Download files
• Delete files
• Share files
• Version files

Non-functional:

• Large storage capacity
• High availability
• Secure access

Architecture

            Client

               |

          File Service

               |

       +---------------+

       |               |

 Metadata DB       Object Storage

Components

Metadata Database

Stores:

FileId

Name

Owner

Size

Location

CreatedDate

Object Storage

Stores actual files.

Examples:

• AWS S3
• Azure Blob Storage

Architecture:

Upload File

    |

File Service

    |

Object Storage

    |

Save Metadata

Large File Upload

Use:

Chunk Upload

Split file:

Large File

 |

+----+----+----+

Part 1 Part 2 Part 3

Upload separately.

Security

Use:

• authentication
• authorization
• encryption
• signed URLs

Senior Answer

I would separate metadata management from file storage. Metadata belongs in a database, while large binary objects should be stored in scalable object storage such as S3 or Azure Blob Storage.

23. Design a Chat Application (like WhatsApp)

Interview Answer

A chat system requires real-time communication, message persistence, and delivery tracking.

Requirements

Functional:

• Send messages
• Receive messages
• Online status
• Message history

Architecture

             Client

                |

          WebSocket Server

                |

          Message Service

                |

          Message Database

Real-Time Communication

Use:

• WebSockets
• SignalR
• Socket.IO

Flow:

User A

 |

WebSocket

 |

Chat Server

 |

WebSocket

 |

User B

Message Storage

Example:

Messages

Id

SenderId

ReceiverId

Text

Timestamp

Status

Handling Offline Users

Use queue:

Message

 |

Queue

 |

Store

 |

Deliver when online

Scaling

Use:

• multiple WebSocket servers
• Redis pub/sub
• message queues

Senior Answer

I would use WebSockets for real-time communication, a message service for processing, persistent storage for history, and queues for handling offline users and reliability.

24. Design a Stock Trading Platform

Interview Answer

A trading platform requires real-time market data, low latency, security, and reliable transaction processing.

Architecture

             Users

                |

          Trading API

                |

       +----------------+

       |                |

 Order Service    Market Data Service

       |                |

 Database        WebSocket Server

Components

Market Data Service

Receives:

• stock prices
• volume
• trades

Example:

AAPL

Price:

$215.50

Volume:

1,000,000

Order Service

Handles:

• buy orders
• sell orders
• validation
• execution

Database

Stores:

Orders

Id

UserId

Symbol

Quantity

Price

Status

Real-Time Updates

Use:

Market Feed

 |

WebSocket

 |

Browser

Important Design Points

Low Latency

Use:

• caching
• asynchronous processing
• optimized databases

Reliability

Use:

• message queues
• transaction logging
• audit trails

Security

Use:

• MFA
• encryption
• authorization

Senior Answer

For a trading platform I would separate order processing from market data streaming. I would use WebSockets for real-time prices, reliable order services, event-driven processing, and strong auditing because financial systems require accuracy and traceability.

25. Design a Search System (like Google Search)

Interview Answer

A search system indexes data and allows users to quickly find relevant information.

Architecture

            Documents

                 |

             Crawler

                 |

             Indexer

                 |

           Search Index

                 |

              Query API

                 |

              User

Components

Crawler

Collects documents.

Indexer

Creates searchable structures.

Example:

Word:

database

Documents:

1,5,10,20

Search Engine

Processes queries.

Examples:

• Elasticsearch
• Apache Solr

Search Optimization

Use:

• inverted indexes
• ranking algorithms
• caching

Example

User searches:

"ASP.NET Core authentication"

System:

Query

 |

Search Index

 |

Rank Results

 |

Return Results

Senior Answer

A scalable search system uses crawlers, indexing pipelines, and search engines such as Elasticsearch. The key design challenge is building efficient indexes and ranking relevant results quickly.

26. Design an Authentication System

Interview Answer

An authentication system verifies user identity and manages access.

Architecture

           User

             |

       Authentication API

             |

        User Database

             |

        Token Service

Login Flow

User enters credentials

        |

Validate password

        |

Generate JWT token

        |

Return token

JWT Structure

Example:

{

 "sub":"123",

 "role":"Admin",

 "exp":123456789

}

Components

User Store

Stores:

• username
• password hash
• roles

Token Service

Creates:

• access tokens
• refresh tokens

Authorization

Checks:

User Role

      |

Permission

      |

Allow / Deny

Security

Use:

• password hashing
• MFA
• HTTPS
• token expiration

Senior Answer

I design authentication systems using secure password storage, JWT or OAuth2 tokens, refresh token management, and role-based authorization.

27. Explain OAuth 2.0.

Interview Answer

OAuth 2.0 is an authorization framework that allows applications to access resources on behalf of users without sharing passwords.

Example:

"Login with Google"

Flow:

Application

      |

Google Authorization Server

      |

User Login

      |

Access Token

      |

Application

Components

Resource Owner

User.

Client

Application requesting access.

Authorization Server

Issues tokens.

Example:

• Microsoft Identity Platform
• Google Identity

Resource Server

API protecting data.

Common Flow

Authorization Code Flow:

User Login

 |

Authorization Code

 |

Exchange Code

 |

Access Token

 |

Call API

Senior Answer

OAuth2 separates authentication and authorization by allowing applications to obtain limited access tokens instead of handling user credentials directly.

28. What is JWT and how does it work?

Interview Answer

JWT (JSON Web Token) is a compact token format used for securely transmitting claims between parties.

Structure:

Header.Payload.Signature

Example:

xxxxx.yyyyy.zzzzz

Flow

User Login

 |

Server validates credentials

 |

Creates JWT

 |

Client stores token

 |

Send token with requests

Request:

Authorization:

Bearer eyJhbGci...

Advantages

• stateless authentication
• scalable
• works across services

Disadvantages

• difficult to revoke
• token size
• requires expiration strategy

Senior Answer

JWT enables stateless authentication by carrying user claims inside signed tokens. In production systems I combine JWT with expiration, refresh tokens, and secure storage.

29. What are distributed transactions?

Interview Answer

Distributed transactions occur when a single business operation involves multiple independent services or databases.

Example:

Order creation:

Order Service

+

Payment Service

+

Inventory Service

Problem:

Payment succeeds:

Payment completed

Inventory fails:

Stock update failed

System becomes inconsistent.

Traditional Solution

Two-Phase Commit (2PC)

Prepare Phase

 |

Commit Phase

Problems:

• slow
• difficult across services

Modern Solution

Use:

• Saga Pattern
• Event-driven architecture

Senior Answer

In microservice systems I avoid traditional distributed transactions when possible and use patterns such as Saga with compensating actions.

30. Explain the Saga Pattern.

Interview Answer

Saga is a pattern for managing distributed transactions using a sequence of local transactions with compensation actions.

Example:

Order process:

Create Order

      |

Reserve Inventory

      |

Process Payment

      |

Send Confirmation

If payment fails:

Compensating actions:

Cancel Order

      |

Release Inventory

Architecture:

Order Service

     |

Event

     |

Inventory Service

     |

Event

     |

Payment Service

Types

Choreography

Services react to events.

Service A

 |

Event

 |

Service B

 |

Event

 |

Service C

Orchestration

Central coordinator controls flow.

Saga Coordinator

 |

Service A

 |

Service B

 |

Service C

Senior Answer

Saga provides a way to maintain consistency across microservices by using local transactions and compensating actions instead of distributed database transactions.

31. What is CQRS (Command Query Responsibility Segregation)?

Interview Answer

CQRS is a design pattern that separates operations that change data (commands) from operations that read data (queries).

Instead of using one model for everything, we use separate models optimized for different purposes.

Traditional CRUD:

Application

      |

  Single Database Model

      |

 Read + Write

CQRS:

              Application

                  |

        +---------+---------+

        |                   |

    Command Side       Query Side

        |                   |

   Write Database     Read Database

Command Side

Handles:

• Create
• Update
• Delete

Example:

CreateOrderCommand

{

    CustomerId = 10,

    ProductId = 100

}

Query Side

Handles:

• Reading data
• Reporting
• Searching

Example:

GetCustomerOrdersQuery

Why use CQRS?

Different Optimization

Write model:

• transactions
• validation
• consistency

Read model:

• fast queries
• denormalized data
• reporting

Example:

Trading platform:

Command:

Buy Stock

 |

Order Service

 |

Trading Database

Query:

Show Portfolio

 |

Read Database

 |

Fast Response

Benefits

• scalable reads
• cleaner business logic
• independent optimization

Challenges

• more complexity
• data synchronization
• eventual consistency

Senior Answer

CQRS separates write operations from read operations, allowing each side to be optimized independently. I use it when systems have very different read and write workloads, such as financial or reporting applications.

32. What is Event Sourcing?

Interview Answer

Event sourcing stores the sequence of events that changed the state instead of storing only the current state.

Traditional database:

Account

Id:

100

Balance:

500

Only the final state is stored.

Event sourcing:

Account Events

Created

Deposit +1000

Withdraw -500

Deposit +200

Current balance is calculated from events.

Architecture

Command

   |

Application

   |

Event Store

   |

Events

   |

Build Current State

Example:

Bank Account:

Events:

AccountCreated

DepositMade(1000)

PaymentMade(300)

PaymentMade(200)

Current state:

Balance = 500

Benefits

Complete History

You know:

• what happened
• when
• who did it

Auditability

Useful for:

• banking
• trading
• insurance

Challenges

• complex queries
• event versioning
• storage growth

Senior Answer

Event sourcing stores business events as the source of truth. It is useful in systems where audit history and traceability are critical, such as financial systems.

33. Design an E-commerce Platform (Amazon-like)

Interview Answer

An e-commerce system requires product management, shopping cart, orders, payments, and inventory.

High-Level Architecture

                 Users

                   |

             API Gateway

                   |

+-----------+-----------+-----------+

Product    Cart      Order     Payment

Service   Service   Service    Service

                   |

             Databases

Main Services

Product Service

Handles:

• catalog
• categories
• pricing

Database:

Products

Id

Name

Price

Category

Cart Service

Stores:

• user cart
• selected items

Order Service

Handles:

• order creation
• status tracking

Example:

Created

 |

Paid

 |

Shipped

 |

Delivered

Payment Service

Integrates with:

• payment providers
• fraud detection

Inventory Service

Tracks:

• available quantity
• reservations

Scaling

Use:

• Redis for product cache
• queues for order processing
• search engine for products

Senior Answer

I would design an e-commerce platform using separate services for catalog, cart, order, payment, and inventory. I would use asynchronous messaging between services to handle high traffic and maintain reliability.

34. Design an Uber-like Ride Sharing System

Interview Answer

A ride-sharing system requires real-time location tracking, matching drivers with passengers, and trip management.

Architecture

Passenger App

       |

Ride Service

       |

Matching Engine

       |

Driver Service

       |

Location Service

Components

Location Service

Receives GPS updates:

Driver:

Latitude

Longitude

Speed

Matching Engine

Finds nearby drivers.

Example:

Passenger Request

        |

Search radius 5 km

        |

Available Driver

Real-Time Communication

Use:

• WebSockets
• SignalR

For:

• driver location
• trip status

Database

Stores:

Trips:

TripId

DriverId

PassengerId

Start

End

Status

Challenges

• millions of locations
• real-time updates
• geographic search

Technologies

Possible:

• Redis Geo
• Elasticsearch
• Kafka
• WebSockets

Senior Answer

A ride-sharing system requires a real-time location service, matching engine, and event-driven architecture. Geographic indexing and WebSocket communication are key for scalability.

35. Design a Payment System

Interview Answer

Payment systems require high reliability, security, and transaction consistency.

Architecture

Client

 |

Payment API

 |

Payment Service

 |

+--------------+

|              |

Bank Gateway   Transaction DB

Components

Payment Service

Handles:

• payment requests
• validation
• processing

Transaction Database

Stores:

Transaction

Id

UserId

Amount

Status

Date

Payment States

Example:

Created

 |

Processing

 |

Completed

 |

Failed

Important Features

Idempotency

Prevents duplicate payments.

Example:

User clicks Pay twice.

System should create:

One Transaction

not:

Two Charges

Audit Logs

Every transaction recorded.

Security

Use:

• encryption
• tokenization
• PCI compliance

Senior Answer

Payment systems require strong consistency, idempotent operations, audit logging, and security controls. I would design them with transactional storage and reliable event processing.

36. Design a Reporting System

Interview Answer

Reporting systems usually separate transactional workloads from analytical workloads.

Problem

Production database:

Orders

Users

Payments

Heavy reports slow down applications.

Solution

Create reporting pipeline:

Production DB

      |

ETL Pipeline

      |

Data Warehouse

      |

Reporting System

Components

Data Extraction

Collect data from:

• SQL databases
• APIs
• logs

Data Warehouse

Optimized for analytics.

Examples:

• Snowflake
• Redshift
• Azure Synapse

Reporting Layer

Examples:

• Power BI
• QuickSight
• SSRS

Example

Daily sales report:

Orders Database

 |

Nightly ETL

 |

Data Warehouse

 |

Dashboard

Senior Answer

I separate reporting workloads from transactional systems using ETL pipelines and analytical databases to prevent performance impact on operational applications.

37. What is a multi-tenant SaaS architecture?

Interview Answer

Multi-tenancy allows multiple customers to use the same application infrastructure while keeping their data isolated.

Example:

SaaS application:

Application

     |

+-----------+-----------+

Company A             Company B

Data                  Data

Common Models

1. Shared Database, Shared Tables

Example:

Customers Table

TenantId

UserId

Name

Data:

TenantId=1

TenantId=2

2. Shared Database, Separate Schemas

Database

 |

 +-- CompanyA Schema

 |

 +-- CompanyB Schema

3. Separate Database

Tenant A DB

Tenant B DB

Tenant C DB

Considerations

• security isolation
• scalability
• cost
• backup strategy

Senior Answer

Multi-tenant systems require strong tenant isolation, security controls, and a database strategy that balances scalability, maintenance, and cost.

38. Explain cloud architecture best practices.

Interview Answer

Cloud architecture should focus on scalability, reliability, security, and cost optimization.

Key Principles

1. Design for Failure

Assume components fail.

Example:

Server Failure

       |

Automatic Recovery

2. Use Managed Services

Examples:

AWS:

• RDS
• S3
• Lambda

Azure:

• SQL Database
• Blob Storage
• Functions

3. Automate Deployment

Use:

• CI/CD
• Infrastructure as Code

Examples:

• Terraform
• CloudFormation
• Bicep

4. Monitor Everything

Collect:

• metrics
• logs
• traces

Senior Answer

Good cloud architecture uses managed services, automation, observability, security controls, and designs for failure rather than assuming infrastructure will always be available.

39. What is disaster recovery (DR)?

Interview Answer

Disaster recovery defines how a system recovers after major failures.

Important metrics:

RTO (Recovery Time Objective)

How quickly must the system recover?

Example:

RTO = 1 hour

RPO (Recovery Point Objective)

How much data loss is acceptable?

Example:

RPO = 5 minutes

Meaning:

Maximum 5 minutes of data loss.

DR Strategies

Backup and Restore

Cheapest.

Pilot Light

Minimal infrastructure running.

Warm Standby

Partially running environment.

Active-Active

Multiple regions serving traffic.

Region A

    |

Users

    |

Region B

Senior Answer

Disaster recovery planning defines recovery objectives, backup strategies, and failover mechanisms to minimize downtime and data loss.

40. What is observability?

Interview Answer

Observability is the ability to understand the internal state of a system by analyzing telemetry data.

The three main pillars are:

1. Logs

Detailed events.

Example:

User login failed

UserId=123

Time=10:05

2. Metrics

Numerical measurements.

Examples:

• CPU usage
• response time
• requests per second

3. Traces

Follow a request across services.

Example:

API Gateway

 |

Order Service

 |

Payment Service

 |

Database

Tools

Examples:

• Application Insights
• CloudWatch
• Grafana
• Prometheus
• ELK Stack

Senior Answer

Observability allows teams to diagnose production problems using logs, metrics, and distributed tracing. For microservice systems it is essential for troubleshooting and performance monitoring.

Final section of System Design.

41. Design a Notification Platform at Scale

Interview Answer

A large notification platform must support millions of notifications through multiple channels:

• Email
• SMS
• Push notifications
• In-app messages

The main challenges are scalability, reliability, retries, and delivery tracking.

High-Level Architecture

                 Applications

                      |

              Notification API

                      |

              Message Queue

                      |

          Notification Workers

        /          |           \

     Email        SMS        Push

    Service     Service     Service

                      |

              Delivery Tracking DB

Components

Notification API

Receives requests:

Example:

{

 "userId":123,

 "type":"ORDER_CREATED",

 "channel":"EMAIL"

}

Message Queue

Examples:

• RabbitMQ
• Kafka
• AWS SQS
• Azure Service Bus

Purpose:

• absorb traffic spikes
• decouple services

Example:

1 million notifications

        |

Queue

        |

Workers process gradually

Worker Services

Responsible for:

• sending messages
• retries
• failures

Example:

Send Email

 |

Failed

 |

Retry after 5 minutes

Delivery Database

Stores:

Notification

Id

UserId

Channel

Status

CreatedDate

SentDate

Scaling Strategies

Use:

• multiple workers
• partitioned queues
• batch processing
• rate limiting

Senior Answer

I would design a notification platform using asynchronous processing with queues, independent workers for each channel, retry mechanisms, and delivery tracking to achieve scalability and reliability.

42. Design a Stock Market Data Pipeline

Interview Answer

A stock market data system must process high-volume real-time market events with very low latency.

Architecture

Market Data Providers

        |

   Data Ingestion

        |

 Message Streaming Platform

        |

+---------------+---------------+

Real-Time API              Storage

        |

 WebSocket Servers

        |

 Trading Clients

Components

Data Ingestion Service

Receives:

• price updates
• trades
• volume
• order book changes

Example:

{

 "symbol":"AAPL",

 "price":215.50,

 "volume":1000

}

Streaming Layer

Examples:

• Kafka
• AWS Kinesis
• Azure Event Hubs

Purpose:

• high throughput
• ordered events
• replay capability

Real-Time Service

Uses:

• WebSockets
• SignalR

Example:

Market Update

       |

WebSocket

       |

Browser Dashboard

Storage

Different storage for different needs:

Time-series database

For:

• historical prices
• charts

Examples:

• InfluxDB
• TimescaleDB

Relational Database

For:

• accounts
• orders
• transactions

Senior Answer

A market data platform should separate ingestion, streaming, real-time delivery, and historical storage. Event streaming platforms such as Kafka help process large volumes of market events reliably.

43. Design a Document Management System

Interview Answer

A document management system stores, searches, secures, and manages business documents.

Examples:

• contracts
• invoices
• reports

Architecture

              User

               |

        Document Service

               |

      +----------------+

      |                |

 Metadata DB     Object Storage

                    |

              File Content

Components

Metadata Database

Stores:

Document

Id

Name

Owner

Version

CreatedDate

Location

Object Storage

Stores actual files.

Examples:

• AWS S3
• Azure Blob Storage

Features

Version Control

Example:

Contract.doc

Version 1

Version 2

Version 3

Access Control

Example:

Manager

Can Edit

Employee

Read Only

Search

Use:

• Elasticsearch
• Azure Cognitive Search

Security

Use:

• encryption
• access policies
• audit logging

Senior Answer

I would separate document metadata from binary storage, using databases for metadata and object storage for files. Security, versioning, and audit history are critical parts of the design.

44. Design a Centralized Logging System

Interview Answer

A centralized logging system collects logs from many applications and makes them searchable.

Architecture

Applications

     |

Log Collectors

     |

Message Queue

     |

Log Processing

     |

Search Storage

     |

Dashboard

Components

Log Collectors

Collect logs from:

• applications
• servers
• containers

Examples:

• Fluent Bit
• Logstash

Processing

Responsibilities:

• parsing
• filtering
• enrichment

Example:

Add:

Application Name

Environment

Timestamp

Storage

Examples:

• Elasticsearch
• OpenSearch
• Splunk

Visualization

Examples:

• Kibana
• Grafana

Example Log

{

 "level":"ERROR",

 "service":"OrderAPI",

 "message":"Payment failed"

}

Senior Answer

A centralized logging platform collects logs through agents, processes them, stores them in searchable systems, and provides dashboards for troubleshooting and operational visibility.

45. Design a Monitoring Platform

Interview Answer

A monitoring platform collects system health information and alerts teams about problems.

Architecture

Services

   |

Metrics Collectors

   |

Time-Series Database

   |

Alert Engine

   |

Dashboard

Metrics Examples

Application:

• response time
• errors
• requests/sec

Infrastructure:

• CPU
• memory
• disk

Database:

• connections
• query performance

Alert Example

CPU > 90%

for 10 minutes

        |

Send Alert

        |

Team Notification

Tools

Examples:

• Prometheus
• Grafana
• Azure Monitor
• AWS CloudWatch

Senior Answer

Monitoring systems collect metrics, analyze system health, and provide alerts. A good monitoring strategy combines infrastructure metrics, application metrics, logs, and distributed tracing.

46. What is API versioning and why is it important?

Interview Answer

API versioning allows APIs to evolve without breaking existing clients.

Example:

Version 1:

GET /api/v1/users

Version 2:

GET /api/v2/users

Common Approaches

URL Versioning

Example:

/api/v1/orders

Header Versioning

Example:

Accept-Version:2

Query Parameter

Example:

/api/orders?version=2

Why Needed?

Because clients may:

• update slowly
• depend on old contracts
• require migration time

Best Practices

• maintain backward compatibility
• document changes
• deprecate old versions gradually

Senior Answer

API versioning allows systems to evolve while maintaining compatibility with existing consumers. I prefer clear versioning strategies and gradual migration paths.

47. Explain security architecture in a modern application.

Interview Answer

Security should be designed at every layer:

• frontend
• API
• database
• infrastructure

Typical Architecture

User

 |

Identity Provider

 |

API Gateway

 |

Application Services

 |

Database

Authentication

Answers:

"Who are you?"

Technologies:

• OAuth2
• OpenID Connect
• JWT

Authorization

Answers:

"What can you do?"

Examples:

Admin

Create/Delete Users

Employee

Read Data

Data Security

Use:

• encryption at rest
• encryption in transit
• secrets management

Infrastructure Security

Use:

• firewalls
• private networks
• security groups

Senior Answer

Modern security architecture uses defense in depth, combining authentication, authorization, encryption, secure networking, and continuous monitoring.

48. How do you handle data consistency in distributed systems?

Interview Answer

Distributed systems require balancing consistency, availability, and performance.

Strategies

Strong Consistency

All nodes see the same data immediately.

Example:

Bank balance.

Eventual Consistency

Systems become consistent over time.

Example:

Social media likes.

Techniques

Transactions

For single database operations.

Distributed Locks

Prevent conflicting updates.

Example:

Only one process updates inventory

Event-driven Updates

Example:

Order Created

 |

Inventory Updated

 |

Notification Sent

Saga Pattern

Uses compensation.

Senior Answer

I choose consistency strategies based on business requirements. Financial transactions may require strong consistency, while large-scale systems often use eventual consistency with event-driven patterns.

49. Explain architecture trade-offs.

Interview Answer

Every architecture decision involves balancing competing goals.

Common trade-offs:

Performance vs Cost

Example:

More servers:

Higher performance

Higher cost

Consistency vs Availability

Example:

Distributed databases:

Strong consistency

slower

Eventual consistency

faster

Simplicity vs Scalability

Monolith:

Simple

Less scalable

Microservices:

Complex

Highly scalable

Speed of Development vs Flexibility

Example:

Framework with conventions:

• faster initially
• less customization

Senior Answer

Good architecture decisions are based on business requirements. I evaluate scalability, reliability, cost, complexity, and maintainability before choosing a solution.

50. Describe a senior architect-level system design scenario.

Interview Answer

A senior engineer should demonstrate:

• requirement analysis
• architecture decisions
• technical trade-offs
• operational awareness

Example:

"Design a global banking platform."

Approach

Step 1: Requirements

Functional:

• accounts
• transfers
• payments
• reports

Non-functional:

• security
• availability
• compliance

Step 2: Architecture

             Users

               |

          API Gateway

               |

       Microservices

               |

       Databases + Events

               |

       Monitoring

Step 3: Important Decisions

Use:

• microservices for domains
• queues for async processing
• databases with replication
• encryption
• audit logs

Step 4: Failure Handling

Plan:

• backups
• disaster recovery
• retries
• monitoring

Senior Answer

At a senior level, I focus not only on designing components but also on operational concerns such as reliability, security, scalability, deployment, monitoring, and future growth.

Chapter 10 — System Design Completed ✅

Total:

50 System Design Interview Questions

Covered:

✅ Scalability
✅ Load Balancing
✅ Caching
✅ Databases
✅ Microservices Architecture
✅ API Gateway
✅ Messaging
✅ Kafka/RabbitMQ
✅ Event Driven Design
✅ CQRS
✅ Event Sourcing
✅ Cloud Architecture
✅ Security
✅ Disaster Recovery
✅ Observability
✅ Real-world system designs