Skip to content

DB Aurora

Purpose: Automatically scale capacity

AWS Aurora

  • is: a fully managed RDBMS

  • does: provides performance & scalability

  • compatible with: MySQL & PostgreSQL

  • designed for: enterprise workloads

architecture: distributed, shared-storage. (separates compute from storage)

Aurora vs RDS

RDS ecosystem: automates provisioning / scaling

  • Aurora

  • RDS = managed database on a single primary instance.

  • Aurora = distributed storage + multiple instances.

RDS deployment

  • tools: AWS Management Console, CLI, or API.

  • Aurora clusters: Aurora MySQL / Aurora PostgreSQL

Features and variants

  • Aurora Serverless v2: automatically scales compute capacity

  • Aurora Global Database: (global apps) replicates data across multiple regions

  • Aurora DSQL (Distributed SQL): adds globally distributed query execution for near-limitless scaling within and across regions

All variants include

  • continuous backup to S3

  • integration with AWS Key Management Service for encryption at rest

1. What Aurora is (conceptual)

Read first: Amazon Aurora overview
Goal: understand what problem it solves.

Focus on:

  • Managed relational database (MySQL / PostgreSQL compatible).

  • Storage and compute are decoupled.

  • One writer, many readers.

  • Built-in replication, backups, and failover.

2. How Aurora differs from RDS

Read next: Aurora vs Amazon RDS. Understand, in order:

  1. RDS: managed database on a single primary instance.

  2. Aurora: distributed storage + multiple instances.

  3. Result: faster failover, better read scaling, higher availability.

Mental model: Aurora is “cloud-native Postgres/MySQL,” not just hosted Postgres/MySQL.

3. Core Aurora concepts

  • Cluster

  • Writer endpoint

  • Reader endpoint

  • Instances

  • Storage layer (6 copies across 3 AZs)

4. Hands-on

Follow one tutorial:

  • Create an Aurora PostgreSQL cluster.

  • Connect via psql.

  • Add a read replica.

  • Observe reader vs writer endpoints.

5. What to learn after

Only proceed if you’ll actually use Aurora:

  • Serverless v2

  • Auto-scaling readers

  • Failover behavior

  • Cost model (this is where surprises happen)