How should an organization design its network architecture to minimize latency and cost when operating across multiple Regions, Availability Zones, and Edge Locations simultaneously?

An optimized global architecture typically places compute resources in the region closest to the majority of users, distributes that workload across multiple Availability Zones within each region for resilience, uses CloudFront edge locations aggressively to cache static and even some dynamic content closer to end users, minimizes unnecessary cross region data transfer since it incurs additional cost and latency, and carefully considers which data truly needs to be replicated globally versus what can remain regionally isolated to control both complexity and cost.

aws cloudfront create-distribution --origin-domain-name myapp.us-east-1.example.com --distribution-config file://global-distribution-config.json