Question

A gaming company runs a multiplayer backend on EC2 instances behind Network Load Balancers deployed in three AWS Regions (us-east-1, eu-west-1, ap-southeast-1). The game uses custom UDP and TCP protocols on non-HTTP ports and requires the lowest possible latency for players worldwide, static entry-point IP addresses that can be embedded in the game client, and fast regional failover if a Region becomes unhealthy. Which edge design best meets these requirements?

Accepted Answer

Use AWS Global Accelerator with the three regional NLBs as endpoints, relying on its anycast static IP addresses and health-check-based traffic dial.

Answer

Place an Amazon CloudFront distribution in front of the NLBs and use latency-based Route 53 records to route players to the nearest Region.

Answer

Create latency-based Route 53 alias records pointing to each regional NLB and enable Route 53 health checks for failover.

Answer

Deploy an Application Load Balancer in each Region with CloudFront and Lambda@Edge to steer players to the closest healthy Region.