DNS Round-robin

What is Round-robin?

DNS Round-robin is a load-balancing technique that requires the use of multiple distinct IP addresses for a single domain name.

It involves associating several IP addresses with a fully qualified domain name (FQDN) to distribute responses to a service across multiple servers, following a round-robin scheduling algorithm.

To distribute traffic randomly across multiple servers, configure multiple A or AAAA DNS records for the same hostname.

Each time the authoritative DNS server is queried, it sends the IP address with which it most recently responded to the back of the queue, operating in a loop, or rotation.

Note: While DNS Round-robin is known for its ease of implementation, it also has significant drawbacks.

 

Advantages

A circular rotation between different IP addresses allows the load caused by high traffic to be distributed among the different machines with these IP addresses.

For each DNS query from a client, the resolution is different. The client then stores the used IP address in its DNS cache, which helps maintain relative stability in managing service sessions.

The servers do not have to be on the same network, nor managed by the same provider.

 

Drawbacks

1) The round-robin method does not always provide effective load balancing by ensuring an evenly distributed load, due to DNS caching and client-side caching.

Because the website's IP address can be cached by recursive resolver servers, it can potentially be heavily used by a significant amount of traffic redirected by the resolver to this cached IP.

2) Another major drawback is that round-robin DNS does not guarantee the reliability and availability of the website.

If one of the servers fails, the authoritative DNS server will still keep the IP address of that inactive server in the round-robin rotation. Therefore, if there are three servers and one is down or disconnected, one out of every three users will be denied service.

Furthermore, DNS round-robin DNS does not take into account server load, transaction time, geographical distance, and other factors that traditional load balancing can address.

 

More articles in this section