Question

A media company runs identical application stacks in two different public cloud providers to avoid vendor lock-in and improve resilience against a full provider outage. Users are globally distributed. The architecture team needs a way to direct users to the healthiest, closest cloud endpoint and automatically stop sending traffic to a provider that becomes unavailable. Which design approach best meets these requirements?

Accepted Answer

Deploy a global DNS-based traffic management service with health checks and geolocation/latency routing policies that resolve to each provider's endpoint

Answer

Configure a single cloud provider's Layer 4 load balancer to distribute traffic across instances in both providers

Answer

Use a CDN with cached content and point all origins to one provider, failing over manually when it goes down

Answer

Establish a VPN mesh between the two providers and rely on internal routing tables to balance user traffic