Question

Difficulty: MediumResilient Database Configurations and High Availability

A logistics company is migrating its core shipment tracking system to AWS. The application uses a PostgreSQL database and must be resilient to a complete AWS Region outage. The solutions architect must design a disaster recovery (DR) strategy with a Recovery Time Objective (RTO) of less than 15 minutes and a Recovery Point Objective (RPO) of less than 10 seconds. Which configuration meets these requirements with the least operational overhead?

  1. Deploy an Amazon Aurora PostgreSQL Global Database with the primary cluster in the active Region and a secondary cluster in the recovery Region, using cross-Region replication.Answer
  2. B
    Deploy an Amazon RDS for PostgreSQL DB instance with a cross-Region Read Replica, and configure an AWS Lambda function to automatically promote the replica during an outage.
  3. C
    Deploy an Amazon RDS for PostgreSQL DB instance with Multi-AZ enabled, and configure Route 53 latency-based routing to automatically fail over to a standby instance in the secondary Region.
  4. D
    Deploy an Amazon Aurora PostgreSQL DB cluster in the primary Region and set up AWS Backup to copy snapshots to the secondary Region every 15 minutes, restoring them in the event of a disaster.

Answer

Deploy an Amazon Aurora PostgreSQL Global Database with the primary cluster in the active Region and a secondary cluster in the recovery Region, using cross-Region replication.
The correct configuration uses Amazon Aurora Global Databases, which feature built-in, storage-based replication across AWS Regions. It achieves an RPO of less than 10 seconds because replication lag is typically less than one second. The secondary cluster can be promoted to a primary cluster with full read/write capabilities in less than 15 minutes, satisfying the RTO requirement with minimal operational overhead.

Step-by-Step Solution

1
Analyze RTO and RPO requirements.
RTO must be under 15 minutes, and RPO must be under 10 seconds. This demands a warm standby or active-passive cross-region database replication mechanism with sub-second replication lag.
Traditional backup and restore methods take too long to restore (exceeding RTO) and have data loss intervals larger than 10 seconds (exceeding RPO).
2
Evaluate cross-region capabilities of Amazon RDS vs. Amazon Aurora.
Amazon Aurora Global Database replicates data at the storage layer with typical latency of less than one second, allowing an RPO of under 10 seconds. It also supports quick promotion of secondary clusters (RTO under 15 minutes).
Amazon RDS for PostgreSQL cross-region replicas use logical/physical replication at the engine level, which incurs more overhead, and promotion requires custom orchestration.
3
Select the option that minimizes operational overhead.
Amazon Aurora Global Database is fully managed and requires no custom scripting to handle replication and failover promotion.
Using Aurora Global Database is the most native and reliable way to meet both strict RTO/RPO limits and minimize management complexity.

Key Concept

Disaster recovery replication and failover constraints for managed relational databases on AWS
Rate this question