How can DNS be used to improve availability across data centers?

How can DNS be used to improve availability across data centers?

• A. DNS can rotate between two load balancers located in different data centers to improve availability if one DC fails.
• B. DNS selects a static IP once and never changes.
• C. DNS bypasses health checks for simplicity.
• D. DNS eliminates the need for load balancers.

Answer: (A)

The concept being tested is using DNS as a lightweight global traffic manager to provide cross-data-center failover. By publishing multiple endpoints for the same service, each behind a load balancer in different data centers, DNS can direct clients to a healthy site. If one data center becomes unavailable, health checks monitored by the DNS service can mark that endpoint as unhealthy and stop returning its IP, so subsequent DNS lookups resolve to the other data center. This gives rapid failover and helps maintain availability even when an entire data center goes down. The choice of TTL affects how quickly failover propagates: shorter TTLs allow faster switching but increase DNS query load, while longer TTLs reduce query load but slow recovery. In practice, rotating between two load balancers in separate data centers leverages DNS for availability, rather than relying on a single static IP, bypassing health checks, or eliminating the need for load balancers inside each data center.

The concept being tested is using DNS as a lightweight global traffic manager to provide cross-data-center failover. By publishing multiple endpoints for the same service, each behind a load balancer in different data centers, DNS can direct clients to a healthy site. If one data center becomes unavailable, health checks monitored by the DNS service can mark that endpoint as unhealthy and stop returning its IP, so subsequent DNS lookups resolve to the other data center. This gives rapid failover and helps maintain availability even when an entire data center goes down. The choice of TTL affects how quickly failover propagates: shorter TTLs allow faster switching but increase DNS query load, while longer TTLs reduce query load but slow recovery.

In practice, rotating between two load balancers in separate data centers leverages DNS for availability, rather than relying on a single static IP, bypassing health checks, or eliminating the need for load balancers inside each data center.