DNS queries are answered by the nearest healthy node on our anycast network. The same IP is announced from every point of presence — BGP routing handles the rest. Automatic failover, sub-millisecond resolution, no single point of failure.
Anycast works by announcing the same IP address from multiple locations simultaneously. When a DNS resolver sends a query to that IP, the internet's BGP routing infrastructure directs the packet to the nearest node — "nearest" in terms of network hops, not geographic distance.
This happens at the network layer with no application-level redirection. The resolver doesn't know or care that the same IP exists in multiple locations. It simply gets a response from whichever node is closest on the network graph.
Each anycast node maintains a full copy of your zone data. Changes propagate to all nodes within seconds via our internal replication system. There are no stale records — every node serves the current version of your zone.
Each node is continuously health-checked by our monitoring infrastructure. If a node fails (hardware fault, network issue, software crash), it's withdrawn from BGP within seconds. Traffic that was routing to that node automatically shifts to the next nearest healthy node.
This happens transparently. DNS resolvers don't retry or reconnect — the BGP withdrawal causes their next query to route to a different node. From the resolver's perspective, nothing changed except possibly a few milliseconds of additional latency.
When the failed node recovers and passes health checks, it's re-announced via BGP and begins serving traffic again. The system is self-healing with no manual intervention required.
Anycast provides natural DDoS resistance. Attack traffic aimed at the anycast IP is distributed across all nodes proportionally to the attacker's network proximity. No single node absorbs the full force of the attack.
A volumetric DNS flood from a botnet spread across 50 countries would be split across all of our anycast nodes, with each node handling only the traffic from nearby bots. This distribution means each node sees a fraction of the total attack volume.
Zone updates propagate to all nodes within seconds via our internal replication protocol. When you add or modify a DNS record, all anycast nodes begin serving the updated record almost immediately.
Sub-millisecond resolution from every continent.