Relational Databases (RDS & Aurora)

AWS Certified Solutions Architect – Associate (SAA-C03) Study Guide

1. Overview

Amazon Relational Database Service (RDS) is a managed service that makes it easy to set up, operate, and scale a relational database in the cloud. Instead of managing the underlying OS and hardware, AWS handles the “undifferentiated heavy lifting” like patching, backups, and hardware provisioning.

The Analogy: Think of EC2 with a DB like buying a fixer-upper house. You are responsible for the plumbing, electrical, and maintenance. RDS is like renting a luxury apartment: you live there and use the facilities, but the landlord handles the repairs, security, and upgrades.

Core Concepts & Well-Architected Framework

When designing database solutions on AWS, we align with these pillars:

  • Reliability: Use Multi-AZ deployments for synchronous replication and automatic failover.
  • Performance Efficiency: Use Read Replicas to offload read traffic from the primary instance.
  • Security: Always encrypt at rest (KMS) and in transit (SSL/TLS). Use Security Groups to restrict access to the DB port (e.g., 3306 for MySQL).
  • Cost Optimization: Use Aurora Serverless for unpredictable workloads to avoid paying for idle compute.

Service Comparison

Feature RDS (Standard) Amazon Aurora
Engines MySQL, Postgres, MariaDB, Oracle, SQL Server MySQL & PostgreSQL Compatible
Replication Up to 5 Read Replicas Up to 15 Read Replicas
Storage Manual scaling (EBS volumes) Auto-healing, scales in 10GB increments to 128TB
Failover Minutes (DNS swap) Seconds (Instantaneous)
Durability Stored in 1 AZ (standard) or 2 (Multi-AZ) 6 copies across 3 AZs by default

Scenario-Based Decision Matrix

If the requirement is…Then use…

  • …High Availability (HA) for Disaster Recovery → Multi-AZ
  • …Scaling read-heavy traffic for a global app → Read Replicas
  • …A legacy SQL Server or Oracle database → RDS (Standard)
  • …Maximum performance and 15+ replicas → Amazon Aurora
  • …A dev/test environment with intermittent use → Aurora Serverless

Exam Tips (SAA-C03 Golden Nuggets)

  • Multi-AZ vs. Read Replicas: Multi-AZ is for Availability (Synchronous). Read Replicas are for Performance/Scaling (Asynchronous).
  • Storage Scaling: RDS can scale storage automatically, but you cannot decrease storage size once allocated.
  • Aurora Global Database: Use this for low-latency cross-region reads and disaster recovery (RTO < 1 min).
  • IAM Authentication: You can use IAM to manage database access instead of traditional passwords for RDS MySQL and PostgreSQL.

Infographic: RDS Architecture Flow

Primary DB (Writes/Reads) Sync (Multi-AZ) Standby (Passive) Async Replication (Read Replicas) Read Replica (Read Only)

Key Services

RDS Proxy: Pools connections to handle thousands of concurrent serverless connections (Lambda).

Aurora Serverless v2: Scales instantly from 0.5 to 128 ACUs.

Common Pitfalls

Downtime: Changing instance class or applying patches without Multi-AZ causes downtime.

Replication Lag: Read Replicas are asynchronous; “Read-after-write” consistency is not guaranteed.

Quick Patterns

Pattern A: Web App → RDS Multi-AZ (High Availability).

Pattern B: Reporting App → Read Replica (Offload heavy queries).

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top