To make the most out of Multi CDN strategy, it’s common to switch between different CDNs and select the best routing path on a daily basis. There are many different ways to route traffic between multiple CDNs such as:
- Round-robin – Traffic is dynamically divided between all CDNs. It is one of the easiest methods to implement, but it does not value geographical or other factors into account.
- Weighted round-robin – AKA Ratio Load Balancing, you’ll need to evaluate and rank different CDNs first based on their performance, capacity, and different specs that you value the most, then set a ratio to route traffic accordingly. This method, however, does not value geographical or other factors as well.
- Geolocation – Traffic is distributed between CDNs based on the end user’s location. This allows you to optimize end-user performance with optimal performance geographically. While being a practical and efficient balancing method, network conditions are being missed out.
- Performance – Performance-based load balancing has been proven to be the best method overall. This method requires in-depth network conditions measurement, some key metrics including response time, throughput, re-buffering ratio, bit rate, and many more.
Real user monitoring and Multi CDN
Real user monitoring is a tool to monitor performance that can collect and analyze data of each user on a website or application. It can use those data to help illustrate how the overall network is performing, how a user is experiencing your website, and what can be done to resolve performance issue (if any).
As mentioned above, “performance-based” load balancing has been proven to be the best method to route traffic, and RUM is the ideal tool to measure each CDN status.
Using RUM along with Multi CDN
Most RUM software only provides semi-raw data and information, businesses will still need to go through several steps to actually take actions:
- Identify goal – It’s important to understand what you are looking to achieve via RUM and Multi CDN. Identify some key metrics first such as performance, cost, server load bandwidth, usage, overage charges, and etc.
- Interpret – Once you have all the data, information, and key metrics, it’s imperative to interpret these data into actionable strategy.
- Execute – When it comes to executing the strategy, there are many different ways and it really depends on how much time and money are available. Some enterprises are willing to invest money to automate the process, and some just do it manually due to the lack of resources.
« Why is RUM important (for modern web applications)? »