Which solution will achieve this goal?
Deploy an additional CloudFront distribution. Create a new Route 53 failover record set with health checks for both CloudFront distributions.
Set the TTL to 4 second for the existing Route 53 record sets that are used for the backend service in each Region.
Create new record sets for the backend services by using a latency routing policy. Use the record sets as an origin in the CloudFront distribution.
Create a CloudFront origin group that includes two origins, one for each backend service Region. Configure origin failover as a cache behavior for the CloudFront distribution.
Explanations:
Deploying an additional CloudFront distribution and adding another Route 53 failover record set with health checks would introduce more complexity. CloudFront itself is designed to improve performance, not provide faster failover by using multiple distributions.
Reducing the TTL to 4 seconds might speed up the DNS resolution but does not address the actual failover process between Regions at the CloudFront level. The TTL only affects how quickly DNS records are refreshed, not how CloudFront handles failover.
Using a latency routing policy in Route 53 with CloudFront as the origin does not specifically address the failover requirement between the primary and DR Regions. Latency routing helps choose the best Region based on latency but does not focus on rapid failover for high availability.
Creating a CloudFront origin group with origin failover allows CloudFront to automatically route traffic to the secondary Region in the event of a failure in the primary Region, which provides the fastest failover time. This solution minimizes downtime.
I weigh that the answer is:
Create a CloudFront origin group that includes two origins, one for each backend service Region. Configure origin failover as a cache behavior for the CloudFront distribution.