Table of Contents
We all know Failover is crucial for maintaining uptime, but did you know about failover’s sidekick, Round Robin? We’re going to show you how you can use the combined powers of both of these services to add redundancy and annihilate downtime for your websites and applications.
Part I: What is Failover
Failover makes sure your web and application servers are still accessible, even in the event of network outages. You can think of it as a form of insurance for your website or application.
Failover operates at the DNS level and is a service often provided by your managed DNS provider.
Failover redirects DNS traffic from your unhealthy webserver to a healthy backup server.
Everyone knows the saying “don’t leave all your eggs in one basket’? Well, the same goes for your network configurations. If you have mission-critical resources, you should never have just one. Yet businesses lose millions of dollars every year because they refuse to add redundant resources to their networks.
What they don’t realize is that while it may be more costly to have double or even triple the resources for a single service, in the long run, it’s significantly cheaper than the cost of downtime.
How does Failover work?
Failover monitors the status of IP addresses in your records. Status checks usually run every 2-4 minutes from numerous monitoring locations.
Failover comes into action only when two of the monitoring nodes report that there is an issue with your primary IP address.
In case your server is down, failover will run another health check to see if your first backup system can accept the incoming traffic.
After confirming its accessibility, traffic will be relocated from the failing server to the healthy redundant backup server.
In the case that the first backup system is also down, the same process Is used for checking the secondary backup server.
When failover gets the green light, the records will be updated to send traffic to the secondary server.
In the meantime, failover will continue to monitor your primary IP address. Once your primary server comes back online, traffic will automatically be redirected back to the original system.
But failover can only do so much on its own… A great addition to the failover process is Round Robin (RR). You can think of it as the sidekick to the failover process
Part II: What is Round Robin?
Round Robin is a method of load balancing that distributes your traffic load across numerous servers. It works on a rotation basis… more on that in a moment.
Round robin’s primary use is to distribute incoming traffic across multiple servers. Let’s say we have our website hosted on three separate web servers which each have their own IP address. And we want our domain to point to all three servers.
Round robin is ideal for any configuration that you have more than one server for one purpose.
For example, multiple web servers that are hosting the same website. RR provides a way to direct traffic load to each of these redundant servers.
How does Round Robin work?
The process of Round robin resembles the motion of a Ferris wheel.
Whenever someone gets into the cabin of a Ferris wheel, the operator spins the wheel so that the next person can get on.
If everyone was to get on the same passenger cart, the ride might experience difficulties. So instead, the Ferris wheel operator distributes the passengers evenly across the carts.
Round robin is similar to this process. Whenever someone searches for www.dnsmadeeasy.com It will give that individual a specific IP address and then rotate it to the end of the list.
Now when another search is done for the same domain www.dnsmadeeasy.com the following IP address (passenger cart) in the queue, will be given to the user.
Every time a search is done, the previous IP address is pushed to the back of the queue, and the following IP in the queue is provided for the next user.
But what happens if you use Round Robin on its own?
In a lot of cases, the sidekick plays a crucial role in helping to save the day. But Round robin does have some drawbacks when implemented without failover.
Round robin can send a user to a downed server; this is like putting a blindfold on the operator of the Ferris wheel.
He won’t know if the next passenger cart is empty, unsecured, or on fire. Now I don’t know about you, but I wouldn’t take that risk.
RR doesn’t take GeoDNS into account either, so it can allocate you to a server that’s far away which will result in slower loading time.
That is why we advise you to use both services for your networks. In conjunction, Failover and Round Robin work brilliantly together.