AWS RDS: Read Replicas vs. Multi-AZ

Understanding the distinction between RDS Read Replicas and Multi-AZ deployments is crucial for AWS DevOps. This summary highlights their differences, use cases, and how they impact application scalability and disaster recovery.

RDS Read Replicas

• Purpose: Scale read operations.
• Configuration:
  • Up to 15 Read Replicas.
  • Can be in the same Availability Zone (AZ), across AZs, or cross-region.
• Replication: Asynchronous, leading to eventual consistency.
• Use Cases:
  • Offloading read traffic from the primary database.
  • Supporting reporting and analytics without impacting the production database.
• Promotion: Can be promoted to standalone databases.
• Networking Costs:
  • Within the same region but different AZs: No cost.
  • Cross-region: Incurs replication fees.

RDS Multi-AZ

• Purpose: Ensure high availability and disaster recovery.
• Configuration:
  • Synchronous replication to a standby instance in a different AZ.
• Replication: Synchronous, ensuring data is consistent across the primary and standby instances.
• Failover: Automatic, with a single DNS name for seamless application connectivity.
• Use Cases: Protecting against AZ failures, instance failures, and network disruptions.
• Scalability: Not used for scaling; standby instance is for failover only.

Key Differences and Considerations

• Read Replicas:
  • Aimed at scaling read capacity.
  • Supports asynchronous replication.
  • Can be promoted to primary databases.
• Multi-AZ:
  • Designed for high availability and disaster recovery.
  • Utilizes synchronous replication.
  • Standby instance can only serve read or write requests once it becomes the primary after failover.

Exam Tips