Implementing latency-based routing

Many systems make the conscious decision not to run across multiple regions because it is simply not worth the additional effort and cost. This is completely understandable when running in an active-passive mode because the additional effort does not produce an easily visible benefit until there is a regional disruption. It is also understandable when running in active-active mode doubles the monthly runtime cost. Conversely, serverless cloud-native systems are easily deployed to multiple regions and the increase in cost is nominal since the cost of a given transaction volume is spread across the regions. This recipe demonstrates how to run an AWS API Gateway and Lambda-based service in multiple regions and leverage Route53 to route traffic across these active-active regions to minimize latency.

...