DOP-C02 Practice Question 97: Global Accelerator Multi Region Failover
AWS Certified DevOps Engineer – Professional (DOP-C02) · Test 10 · question 7 of 10
A latency-sensitive TCP application runs behind Network Load Balancers in two Regions. The company needs client traffic to keep using the same two static IP addresses no matter which Region is serving it, and it needs an unhealthy Region to be removed from rotation within seconds, well under what DNS-based failover with typical client-side caching can guarantee.
Which service meets both requirements?
Show answer and explanation
Answer:
Global Accelerator provisions a fixed set of static anycast IP addresses that never change even as endpoints in different Regions are added, removed, or fail. It performs health checks against each Region's endpoint group roughly every 30 seconds and reroutes traffic away from an unhealthy Region within seconds, at the network layer, without depending on a client re-resolving DNS or respecting a TTL. It also supports TCP and UDP traffic directly, which fits a latency-sensitive TCP application better than an HTTP-oriented service.
Option A keeps the same DNS names but not the same IP addresses, and failover speed depends on resolvers and clients honoring the TTL, which real-world caching often does not do reliably, making the "within seconds regardless of client caching" guarantee unachievable.
Option B is built for HTTP or HTTPS content delivery with origin failover, not for a generic TCP application, and CloudFront does not provide static IP addresses that a client connects to directly.
Option D describes a load balancer topology within or across Regions but does not provide any mechanism for a client to keep using the same static IP addresses while traffic is redirected across Regions during a failure.
Reference: What is AWS Global Accelerator?