Round robin Domain Name System (DNS) refers to a method of load balancing, load distribution or fault-tolerance provisioning of various obsolete Internet Protocol service hosts, such as FTP servers, Web servers, etc., by handling the responses of the DNS. This is for handling requests from client computers as per a proper statistical model.
In this load-balancing technique, as opposed to the standard load techniques, the balance of power is placed in the DNS server rather than in a fully dedicated machine.
Round robin DNS works on a rotation basis during which the IP address of a server is given out, and then it proceeds to the back of the list; the IP address of the next server is given out, and then it proceeds to the end of the list. This process continues with respect to the number of servers utilized. This process is carried out in a looping fashion.
Round robin DNS is mainly used to balance the load of Web servers that are geographically distributed. For instance, if an organization includes three identical home pages that reside in three servers having three separate IP addresses, but only one domain name, then the process will be as follows:
The first user accessing the home page is taken to the first IP address.
The second user accessing the home page is taken the next IP address.
The third user is forwarded to the third IP address.
In every case, as soon as the IP address is handed out, it moves to the end of the list. Therefore, the fourth user is taken to the first IP address, and so on.
Even though it is quite simple to employ, round robin DNS has some drawbacks, including those inherited from TTL times and the DNS hierarchy itself, which results in unexpected address caching that is quite difficult to handle.