Anycast
Advertising the same IP address from many locations, so the network routes each client to the topologically nearest one.
One address, many physical points of presence; BGP decides which one a given client reaches. This is how CDNs, public DNS resolvers and large-scale DDoS absorption work.
Two properties follow. Latency: clients reach a nearby node without any DNS-based geo logic, and without waiting for a TTL. Failover: withdrawing a route removes a location from service in seconds, which is far faster and far more reliable than DNS failover, since it does not depend on resolvers honouring TTLs.
The constraint is that anycast is naturally suited to stateless or short-lived interactions. Long-lived TCP connections can in principle be re-routed mid-session by a routing change, so stateful services behind anycast need either session-aware steering or tolerance for reconnects.