If a node serving your DNS requests goes down due to hard drive failure, power loss, network connectivity, etc. DNSFilter will immediately stop its BGP (anycast) announcement, and your traffic will reroute to the next-nearest location with zero packet loss.
Additionally, DNSFilter achieves redundancy within each datacenter through practices such as load balancing, redundant fiber connectivity, power protection, backup generators, and off-site backups. However, from time to time when a datacenter or catastrophic event occurs, our customers will not be affected. This approach allows us to pro-actively and conveniently take a data center off our global network whenever such an event or maintenance occurs.
Beyond our network, DNSFilter is also engineered to ‘fail open’ – should any element of our architecture fail (such as master storage or APIs), our slaves will take over. We employ tiered slaves under a micro-service architecture. These slaves can continue operating with the master down, and even if they fail we are able to default to open-resolving until the issue is resolved.