Which solution meets these requirements with the LEAST amount of operational overhead?
Replace the software load balancer with an AWS Application Load Balancer. Create an AWS Global Accelerator accelerator. Add an endpoint group for each Region. Configure Route 53 health checks. Add an alias record that points to the accelerator.
Replace the software load balancer with an AWS Network Load Balancer. Create an AWS Global Accelerator accelerator. Add an endpoint group for each Region. Configure Route 53 health checks. Add a CNAME record that points to the DNS name of the accelerator.
Replace the software load balancer with an AWS Application Load Balancer. Use AWS Global Accelerator to create two separate accelerators. Add an endpoint group for each Region. Configure Route 53 health checks. Add a record set that is configured for active-passive DNS failover. Point the record set to the DNS names of the two accelerators.
Replace the software load balancer with an AWS Network Load Balancer. Use AWS Global Accelerator to create two separate accelerators. Add an endpoint group for each Region. Configure Route 53 health checks. Add a record set that is configured for weighted round-robin DNS failover. Point the record set to the DNS names of the two accelerators.
Explanations:
This solution uses AWS Global Accelerator with an Application Load Balancer, providing static IP addresses across multiple AWS Regions. Global Accelerator preserves the client’s source IP address, meeting the auditing requirement. Route 53 health checks enable failover with minimal operational overhead.
AWS Global Accelerator does not support a CNAME record; instead, it uses static IPs for failover and routing. Additionally, Network Load Balancers are required, as they natively support static IPs, unlike Application Load Balancers.
Using two separate accelerators adds unnecessary complexity and operational overhead, as Global Accelerator can natively handle multi-Region failover. This option is not optimal for simplicity and does not meet the requirement of minimal operational overhead.
Weighted round-robin DNS failover does not provide a true active-passive failover setup. Additionally, creating two separate Global Accelerators introduces unnecessary complexity, contrary to the goal of reducing operational overhead.
As I understand it, the answer is:
Replace the software load balancer with an AWS Application Load Balancer. Create an AWS Global Accelerator accelerator. Add an endpoint group for each Region. Configure Route 53 health checks. Add an alias record that points to the accelerator.