Multi-CDN Strategies
Blog September 29, 2017

Multi-CDN Strategies For The 99%

Originally posted on Forbes Technology Council

It used to be that only the largest and most well-funded companies could afford to use more than one content delivery network. However, new cloud-based routing techniques are making this strategy a reality for smaller organizations.

There are two different ways you can manage a multi-CDN architecture: automation through a single management provider or DNS load balancing.

Shutterstock

Automated Multi-CDN Management

The first is typically managed through a larger service provider that automatically balances traffic loads across multiple CDNs. Traffic is generally served evenly to each CDN. Automation can save you a lot of time and labor since the provider is responsible for monitoring the CDNs and staying on top of updates. You’ll also enjoy cheaper pricing since the service provider is purchasing its bandwidth in bulk from each CDN vendor.

This method has grown in popularity after a slew of single provider outages last year that brought down dozens of the largest websites on the internet. Having a redundant CDN provider ensures that users will experience no performance degradation even during a single-provider outage.

There are a couple problems with this method. Since everything is managed through a single control panel, not directly through the CDNs, routing decisions are restricted. Also, settings and functionality are limited to what is shared among both providers.

 

DNS Load Balancing

DNS load balancing solves all of these issues since you can choose your providers, you can avoid vendor lock-in and diversify your services. That’s why load balancing is the preferred method among developers.

The secret lies in the DNS. The DNS, or Domain Name System, maps domains to their computer-readable IP addresses. In a multi-CDN environment, the domain points to the hostnames of the CDN providers. The CDN providers then point to the IP address of the web servers that are closest to the end user.

When you steer traffic at the DNS level, there are no additional lookups required. That means speedier resolution and even lower page load times.

DNS providers are now offering services that use integrated monitoring nodes to test which CDN is responding faster. Traffic is then automatically routed to the better performing CDN. That means your traffic is always answered by the best performing server, no matter the vendor.

We mentioned earlier that this method is better for organizations that need greater customization at the provider level. This is invaluable for brands that are looking to expand into new regions that require a particular vendor or vendor-specific feature.

But it can also be more costly since you will need to have contracts with each CDN provider and then a DNS management provider. This method is also more time intensive since you will be working with more than one control panel and then an additional control panel for your DNS load balancing.

 

The Big Picture

Now that we’ve covered the pros and cons of each strategy, how do you know which one is right for your business? It comes down to how much labor and resources you have to dedicate toward your multi-CDN strategy. DNS load balancing is best for companies that have a dedicated team member who can monitor and manage both CDNs. These kinds of companies are typically international enterprises and video streaming services.

Automation is more suitable for startups and growing businesses that don’t have the luxury of time or dedicated staff for multi-CDN management. This method also offers quick implementation with limited maintenance, which could be ideal for smaller teams.

The bottom line is, both of these strategies are affordable for more than just the top 1% of businesses. Take the time to research providers and you’ll see that more and more CDN and cloud-hosting providers are adding multi-CDN management to their service offerings. Now is the time to consider a multi-CDN strategy and how it can not only maintain availability but optimize performance on a global scale.