Question #737
A software-as-a-service (SaaS) provider uses an Application Load Balancer (ALB) to expose APIs hosted on an Amazon Elastic Kubernetes Service (EKS) cluster in the eu-west-1 Region. The APIs utilize non-standard REST methods such as PURGE, CHECKOUT, MERGE, and COPY. Users in Asia and South America report high latency and inconsistent performance. A solutions architect must address the latency issue while minimizing operational complexity. Which solution meets these requirements?
Deploy an Amazon API Gateway regional REST API to front the ALB. Configure integration mappings for the non-standard methods.
Create an Amazon CloudFront distribution with the ALB as the origin. Configure cache behaviors to forward the non-standard HTTP methods.
Use AWS Global Accelerator with the ALB as an endpoint group. Enable client affinity for consistent routing.
Replicate the EKS cluster to ap-southeast-1 and sa-east-1 Regions. Use Amazon Route 53 geolocation routing to direct traffic.
Explanation
The correct answer is C. AWS Global Accelerator uses AWS's global network infrastructure to direct user traffic to the nearest edge location, which then routes requests over optimized paths to the ALB in eu-west-1. This reduces latency for users in Asia and South America. Global Accelerator operates at the TCP layer, supporting all HTTP methods (including non-standard ones like PURGE, COPY) without additional configuration. Enabling client affinity ensures consistent routing, improving performance predictability.
Other options are less suitable:
- A (CloudFront) may not reduce latency for dynamic API calls and requires cache behavior adjustments.
- B (API Gateway) remains regional, not addressing cross-region latency.
- D (multi-region EKS) adds operational overhead.
Key points: Global Accelerator optimizes network paths, supports all protocols/methods, and requires minimal changes.
Answer
The correct answer is: C