Global DNS Failure Technology

Global DNS Failover using IP Anycast & BGP

• Data centers on both coasts, San Jose California, Sterling Virginia announce the /23 aggregate.
• Both Data centers announce one /24 out of the larger /23 supernet. For example, San Jose will announce the upper /24 and Sterling will announce the lower /24. If San Jose fails to respond, DNS queries destined for San Jose will be routed to Sterling because Sterling will contain the next longest network match, the /23, and be fulfilled by the name servers there.
• Our Global DNS Failover offeres many advantages including:
  • Fewer DNS queries failures.
  • Scalable to handle multiple name servers in multiple locations.
  • Improved performance and load-handling.
• The Nominum network naming solution provides many advantages including:
  • 6 completely redundant nameserver solutions.
  • Anycast technology, which increases DNS resolution time by directing requests to the nameserver which is the closest to the requestor.
  • Centralized DNS distribution, which allows your DNS changes to be simultaneously broadcast to all nameservers. This allows for solid redundancy, should one of the nameservers stop responding.
  • Hot fail over for all nameservers, which provides an uninterrupted flow of DNS resolution, even if one or more of the nameservers should fail to respond.

Datacenters

Dulles	San Jose
Sterling, VA Verio Facility Multiple OC12s (620 mBps)	California Verio Facility Multiple OC12s (620 mBps)
London	Frankfurt
London, England Verio Facility One STM-16 (2.5 Gbps) and STM-4 (622 Mbps)	Frankfurt, Germany IX Europe Facility Two STM-4 (622 Mbps)