If your website receives heavy traffic and you are hosted across multiple infrastructures across the world or you are using multiple CDNs, then you ought to use a DNS Load Balancer to reroute traffic for better performance.
While there are some large companies who create their own DNS load balancing system, which is complex and tedious to implement and manage, some companies make life easier by opting for ready-made solutions such as those from Cedexis, Dyn, NSONE, Rage4 etc
Most of the above mentioned companies are backed by Anycast networks spread across global locations.
While there are some differences between the feature sets offered by these companies and the number of POPs they own, it is very important to identify the one that works best for your business.
How can you identify which DNS Load Balancing solution works for you?
You may have already realized by now that at Catchpoint we strongly recommend our customers and prospects have a strategy around DNS. The first step of that strategy is to create benchmark tests to compare the performance of each of these companies by monitoring their DNS servers.
Catchpoint has a global presence spread across 100+ locations with hundreds of backbone & last mile nodes which can help you measure the performance of each of these DNS load balancers.
How can you test?
The above testing methodology allows you to compare apples to apples and choose the best vendor that works for your business.
But wait, it doesn’t end at choosing the right solution.
You should be monitoring it 24/7 to ensure you provide a seamless experience for your end users.
While most DNS load balancers use Anycast, if there is a failure the user doesn’t really notice the failure as they automatically get rerouted to a working DNS server. But the user does have to wait longer, in many cases much longer, for the page to load so this process of rerouting does have a significant effect on the end users’ experience.
Here is an example:
When we dig deeper, we notice that the issue was related to poor DNS Resolution. As explained above while we don’t see many failures, the overall performance of the page—it took up to 13 seconds to load for some users—was impacted by high DNS resolution time:
With the help of our Web tests, we identified the root cause to be the DNS. Now the only question is, which DNS server? So, we monitored the different DNS servers using our DNS tests and voilà!
The graph on the right shows the specific DNS server which is failing across multiple ISPs. In the above scenario it was this specific DNS server, but the next time – who knows?
And this is why a DNS Monitoring Strategy is essential as you continue to make your way on the performance journey.