🗄️ AWS Databases for Backend Engineers

RDS (PostgreSQL/MySQL) • DynamoDB • Aurora

🚀 Introduction

AWS offers multiple managed databases for different backend use cases — from transactional systems to high-throughput NoSQL workloads.

This guide covers the three most critical database services backend engineers must know:

Service	Type	Best For
RDS (PostgreSQL/MySQL)	Relational	Traditional applications, OLTP workloads
Aurora	High-performance relational	Scalable, fault-tolerant modern apps
DynamoDB	NoSQL Key-Value / Document DB	Low-latency, large-scale apps, IoT, serverless

This documentation includes real-world examples, architectures, and interview-grade insights.

🟦 1️⃣ Amazon RDS (PostgreSQL / MySQL)

Managed Relational Database | Automatic Backups | Multi-AZ

✅ What is RDS?

RDS is AWS’s managed relational database service supporting engines like:

PostgreSQL
MySQL
MariaDB
Oracle
SQL Server

AWS handles:
✅ Backups
✅ Patching
✅ Failover
✅ Monitoring
✅ Storage scaling

✅ Real-World Use Case – Backend API with PostgreSQL


API → Private Subnet → RDS PostgreSQL

Text Diagram


             Internet
                │
                ▼
        ┌────────────────┐
        │  ALB (Public)  │
        └────────────────┘
                │
                ▼
       [ App Server / ECS / Lambda ]
                │ JDBC/PG Driver
                ▼
        ┌────────────────┐
        │ RDS PostgreSQL │
        └────────────────┘

Common backend workloads:

User accounts
Transactions
Logging
Sessions
Product catalog

✅ Key Features

Feature	Description
Multi-AZ	Standby replica for failover
Read Replicas	Scale read traffic
Automated Backups	Point-in-time recovery
Storage autoscaling	Avoid storage exhaustion
Enhanced Monitoring	OS-level insights

✅ Best Use Cases

Monolithic applications
Traditional transactional workloads
Apps requiring strong ACID consistency
Reporting and analytics using read replicas

🧠 Interview Tip

RDS = fully managed relational DB with multi-AZ failover and read replicas.

🟩 2️⃣ Amazon Aurora (MySQL/PostgreSQL Compatible)

Distributed Storage | High Performance | Serverless Options

✅ What is Aurora?

Aurora is AWS’s modern relational database, compatible with:

PostgreSQL
MySQL

Aurora provides:
✅ 5× performance of MySQL
✅ 3× performance of PostgreSQL
✅ 6-way replicated storage
✅ Automatic failover
✅ Autoscaling (Aurora Serverless v2)

✅ Real-World Use Case – Scalable E-Commerce Application


ALB → App → Aurora Cluster (Writer + Readers)

Text Diagram


               ┌─────────────────────┐
               │   Aurora Cluster    │
               └───────┬────┬────────┘
                       │    │
          ┌────────────┘    └────────────┐
          ▼                                ▼
   [ Writer Node ]                 [ Read Replica ]
       (writes)                       (reads)

Aurora storage layer:

6 copies across 3 AZs
Auto-healing
Auto-scaling

✅ Key Features

Feature	Description
Serverless v2	Auto-scaling compute
Multi-master	Multi-writer support (Aurora MySQL)
Fast failover	< 30 seconds
Global Database	Low-latency cross-region reads
Backtrack	"Undo" DB changes by minutes/hours

✅ Best Use Cases

High-traffic backend systems
SaaS platforms
Enterprise OLTP systems
Multi-region apps
Microservices needing relational consistency at scale

🧠 Interview Tip

Aurora = high-performance, distributed relational database with auto-scaling.

🟨 3️⃣ Amazon DynamoDB

NoSQL Key-Value & Document Store | Serverless | Low-latency

✅ What is DynamoDB?

DynamoDB is AWS's serverless NoSQL database offering:

⚡ Ultra-low latency (single-digit ms)
Infinite horizontal scaling
Built-in HA across multiple AZs
Automatic partitioning
Pay-per-use (on-demand mode)
Fine-grained access control (IAM)

💡 Real-World Use Case – Metadata / IoT / High TPS Workloads


IoT Device → API → DynamoDB

Text Diagram


[  API Gateway  ]
        │
        ▼
[ Lambda Function ]
        │
        ▼
[ DynamoDB Table ]

Use cases:

IoT telemetry
User sessions
Shopping carts
Feature flag storage
Configuration/state
Gaming leaderboards

✅ Key Features

Feature	Description
On-demand mode	Auto-scaling read/write capacity
Global Tables	Multi-region, active-active replication
Streams	Event handling for CDC
TTL	Auto-expiration of data
Single-digit ms latency	High throughput

✅ DynamoDB vs RDS vs Aurora (Conceptual)

Feature	DynamoDB	Aurora	RDS
Type	NoSQL	Relational	Relational
Scaling	Automatic	Automatic	Manual
Consistency	Eventually/Strong	Strong	Strong
SQL support	No	Yes	Yes
Transactions	Yes	Yes	Yes
Best for	High TPS	Enterprise scale	Traditional apps

🧠 Interview Tip

DynamoDB = NoSQL, auto-scaling, low-latency, high throughput.

🟥 4️⃣ High-Level Comparison: RDS vs Aurora vs DynamoDB

✅ Decision Table (Backend Engineer View)

Feature	RDS	Aurora	DynamoDB
Database Model	Relational	Relational	NoSQL
Scaling	Vertical + Read replicas	Horizontal + Serverless	Fully horizontal
Performance	Good	Excellent	Extreme
Pricing	Instance-based	Higher (cluster-based)	Usage-based
Availability	Multi-AZ	6 copies across 3 AZs	Multi-AZ (default)
Schema	Strict	Strict	Flexible
Use Case	Traditional apps	High-scale apps	High-throughput apps

📘 5️⃣ Architecture Diagrams (ASCII)

✅ Multi-AZ RDS Architecture


[ App Servers ]
       │
       ▼
[ RDS Primary ]  → synchronous → [ Standby Replica ]
       │
       ▼
   Read Replica (async)

✅ Aurora Cluster Architecture


             ┌─────────────────────────────┐
             │  Distributed Storage Layer   │
             │ (6 copies across 3 AZs)      │
             └───────────┬────────────┬────┘
                         │            │
            ┌────────────┘            └────────────┐
            ▼                                       ▼
    [ Writer Node ]                         [ Reader Nodes ]

✅ DynamoDB Architecture


Client
  │
  ▼
DynamoDB API
  │
  ▼
Auto-Partitioned Storage Layer
  │
  ▼
Multi-AZ Replication

🧠 6️⃣ Interview Questions & Answers

✅ RDS

Q: What is Multi-AZ?
A: Synchronous replication to a standby for automatic failover.

Q: How do read replicas work?
A: Asynchronous replication for read scaling only.

✅ Aurora

Q: Why is Aurora faster than RDS?
A: Distributed storage engine + high-performance cache + parallel writes.

Q: What is Aurora Serverless?
A: Auto-scaling compute capacity based on load.

✅ DynamoDB

Q: How does DynamoDB scale?
A: Automatic partitioning based on key access patterns.

Q: What is DynamoDB Streams?
A: Real-time change data capture (CDC).

✅ 7️⃣ Best Practices Summary

✅ RDS

Use Multi-AZ for production
Use parameter groups for tuning
Avoid storing files/blobs
Use read replicas for read-heavy workloads

✅ Aurora

Prefer Serverless for unpredictable workloads
Use performance insights
Use global database for cross-region reads

✅ DynamoDB

Use partition keys designed for distribution
Use TTL for expiring data
Enable auto-scaling
Use Streams for event-driven patterns

✅ Final Takeaways

Service	Choose When
RDS	You need traditional SQL with stable workloads
Aurora	You need relational with high performance & auto-scaling
DynamoDB	You need massive scale, ultra-low latency, flexible schema

Backend engineers typically use all three in different architectures — the key is choosing the right database for the job.