I would suggest #13 is updated to make use of Transit Gateway peering in a full mesh configuration, and use AWS backbone/internal routing to provide inter-region latency. This is most likely what large-scale deployments in AWS would make use of and is the more modern way to measure inter-region latency (as opposed to using public endpoints and the public internet).
I would suggest #13 is updated to make use of Transit Gateway peering in a full mesh configuration, and use AWS backbone/internal routing to provide inter-region latency. This is most likely what large-scale deployments in AWS would make use of and is the more modern way to measure inter-region latency (as opposed to using public endpoints and the public internet).