Examkin

Question

Difficulty: HardHigh Availability and Disaster Recovery Design

A financial services company is designing a disaster recovery (DR) solution for a critical transaction auditing application. The application currently runs in the us-east-1 Region on Amazon ECS tasks using AWS Fargate behind an Application Load Balancer (ALB). The database tier uses a Multi-AZ Amazon Aurora PostgreSQL cluster. Outbound connections to external clearinghouses must go through NAT Gateways. The company needs to establish a DR site in the us-west-2 Region. The business requirements specify a Recovery Time Objective (RTO) of 15 minutes and a Recovery Point Objective (RPO) of 5 minutes, while minimizing infrastructure costs during normal operations. Which of the following architectures meets these requirements most cost-effectively?

  1. Configure Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2. Deploy the Amazon ECS application in both regions, but set the service desired task count to zero in us-west-2. Set up an ALB in each region, and deploy NAT Gateways in each Availability Zone in both VPCs. Configure Amazon Route 53 Failover routing policies with health checks pointing to the ALBs. In a failover scenario, promote the Aurora secondary cluster and scale up the ECS tasks in us-west-2.Answer
  2. B
    Configure daily Amazon EBS snapshots of the ECS instances and copy them along with daily RDS manual snapshots from us-east-1 to us-west-2. In a disaster recovery event, restore the database from the snapshot, recreate the ECS tasks in us-west-2, and update Route 53 Failover routing records. Deploy a single NAT Gateway in each region to minimize baseline costs.
  3. C
    Configure Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2. Deploy the ECS application in both regions with active tasks running at a minimum scale. Deploy a single NAT Gateway in each region's VPC, shared across all Availability Zones, to reduce hourly NAT charges. Configure Route 53 Failover routing with health checks pointing to the Application Load Balancers.
  4. D
    Deploy a Multi-AZ Amazon RDS PostgreSQL database in us-east-1. Configure the ECS tasks in us-west-2 to point to the standby replica in the secondary Availability Zone of us-east-1 to serve read traffic and minimize cross-region latency during normal operations. Deploy NAT Gateways in each Availability Zone in both regions. Configure Route 53 Geolocation routing to direct traffic based on the user's location.

Answer

Configure Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2. Deploy the Amazon ECS application in both regions, but set the service desired task count to zero in us-west-2. Set up an ALB in each region, and deploy NAT Gateways in each Availability Zone in both VPCs. Configure Amazon Route 53 Failover routing policies with health checks pointing to the ALBs. In a failover scenario, promote the Aurora secondary cluster and scale up the ECS tasks in us-west-2.
The correct architecture uses Amazon Aurora Global Database to achieve near-instantaneous cross-region replication, matching the strict 5-minute RPO. Setting the Amazon ECS desired task count to zero in the disaster recovery region avoids compute costs during normal operations while allowing rapid scale-up (well within the 15-minute RTO) when Route 53 detects a failover condition. Deploying NAT Gateways in each Availability Zone satisfies the high availability requirements for outbound API calls.

Step-by-Step Solution

1
Evaluate replication capability for RPO.
Amazon Aurora Global Database provides cross-region physical replication with latency typically under 1 second, fulfilling the 5-minute RPO requirement.
Choosing a replication technology with sub-second lag guarantees that data loss is minimized during regional failovers.
2
Determine compute tier configuration for RTO and cost.
Pre-deploy the ECS task definitions and services in the secondary region but scale the desired task count to zero.
This implements a warm standby/pilot light architecture that incurs no Fargate compute costs during normal operations, yet allows scaling up tasks within a few minutes to meet the 15-minute RTO.
3
Configure highly available outbound connectivity.
Deploy a NAT Gateway in each Availability Zone of the VPCs in both regions.
Deploying NAT Gateways per AZ ensures that the outage of a single AZ does not disrupt outbound traffic to external clearinghouses from the remaining AZs.
4
Establish global DNS routing and health checks.
Configure Route 53 Failover routing policy with health checks pointing to the ALBs.
Active-passive failover routing automatically routes user traffic to the secondary region ALB once the primary endpoint health check fails.

Key Concept

Disaster Recovery strategies (Warm Standby/Pilot Light) balancing RTO/RPO objectives with cost-effectiveness, combined with database cross-region replication and NAT Gateway high availability.
Rate this question

Topics

High Availability and Disaster Recovery Design

More questions from this topic