Are you using Anycast DNS already? No? It is a great technology that took more than 30 years to get to today’s level. Now you will learn how exactly Anycast DNS was developed.
What is Anycast DNS?
Anycast DNS is a routing method that uses a simple trick – put the same IP address on all of the instances. Using it, you will have a network of Anycast DNS servers with exactly the same IP address. Yes, this is possible, and there is no network problem. What you are getting as a benefit is that whoever is searching your domain name will get the same IP address, no matter where in the world he or she is. Then the request will go on the way to this IP address, and since it is the same IP address, it will get a response from the closest Anycast DNS server there is. This will save time and provide redundancy since if one server is down, the request will simply travel a bit further and find another answer from the next Anycast DNS server on the network.
Try this great Anycast DNS Service!
Anycast DNS history
The DNS creation
In the 70s, there was a huge increase in the number of new computer devices. They were starting to work together in networks. The networks started to get interconnected, and finally, the internet was created. But, at that early stage, everything was complicated. There was no DNS for resolving domains. Everybody was using a centralized Host.txt file, which had a table with hostnames and IP addresses. The number of new devices was increasing and also the complexity to use this method for resolving. This is why the scientist Paul Mockapetris from the Internet Engineering Task Force (IETF) started working on a solution. After a few years of work, he and his team finally introduced the Domain Name System (DNS), a dynamic decentralized solution for domain resolving that was actually possible to manage on a large scale.
Anycast technology has been around since 1989. Then, the principle of delivering a request to any host from a group of hosts with the same IP address was used by different service providers.
In 1993, the same IETF finally documented the Anycast routing method well in its RFC 1546. The document was called “Host Anycasting Service” and explains how the DNS Anycast address that is shared between instances can boost the network performance and provide autoconfiguration of DNS resolvers.
Back in these days, not all hosts supported Anycast addresses, and those were discarding the data packets that were going to Anycast addresses. Another problem of the time later fixed with DNSSEC is that other servers could divert the traffic to themselves, faking their IP address and intercept communications.
In 1999, IETF released the RFC 2526 that explains how Anycast DNS could work with IPv6. At this time, the IPv6 was still new, and it has a different address structure. Some IP addresses were reserved for subnet Anycast addresses, based on subnet prefix identifiers in the IPv6 address space.
Between 2003 and 2006, various root DNS servers started using Anycast to improve redundancy.
In 2006, There were two RFC documents (RFC 4291 and RFC 4786). The first updates the way IPv6 addresses work with Anycast, and the second describes the best practices, at this time about signaling service availability, equal-cost paths, reverse path checking, route dampening, and more. Some security issues are detailed, like service hijacking and Denial-of-Service attacks and mitigation.
The latest document that you can read is RFC 7094 from 2014. It most fully describes Anycast DNS.
Suggested article: Anycast vs Unicast vs Multicast vs Broadcast
If you came to this article searching for a simple solution to improve your uptime, look no further. Anycast DNS is an excellent choice. It is already a mature technology that many, including most of the TLD servers, use. So, what are you waiting for? Start using Anycast DNS today!