How to measure Content Delivery Network performance

It is widely accepted that content delivery networks are a great way to increase a website’s performance, as providing advanced acceleration services that can improve end user experience dramatically.  Most of the world’s largest websites take advantage of CDN services, and such is their cost effectiveness that many small to medium sized businesses are now seeing the benefits of well implemented CDNs. However, with so many CDN providers entering the crowded and competitive market every year, organisations looking to boost their online performance are facing ever greater dilemmas – which CDN provider is best, how should you measure content delivery network performance and which do you select to accelerate your data?

There are several methods that can be used when analysing the performance of a specific content delivery network provider. The most important benchmarks include:

– Server side performance monitoring:

This method of testing simulates end user data requests, and then measures how quickly a webpage responds to the request. Whilst this is useful for determining response speeds, it isn’t perfect. This is because it often only measures the time it takes the initial document files to be delivered, neglecting files such as images, CSS files and JavaScript. This type of measurement is therefore highly unreliable, and can in many cases be seconds faster than loading speeds are in reality.

– Synthetic transaction monitoring:

To overcome these server side monitoring drawbacks, a second, more accurate method can be used. Synthetic transaction monitoring incorporates emulators and real world browsers to test predefined data requests from many different locations. Whilst this is supposedly as close a method to measuring real world end user performance as possible, it can also be plagued inaccuracies. This is because it does not accurately source requests from the exact locations that your specific customers are based.

– Measuring the performance of a few users:

This method is a more accurate ways of measuring a websites performance, especially when compared to the previous options. By taking a selection of your end users and measuring how long it takes for them to access your pages, you can use the averages of these results to make generalisations about the speed of your content delivery network.

– Measuring the performance of every end user:

This is the most accurate method of measuring the performance of your website. By measuring all actual transaction times across a network and from all users, you can record response times from a server, network and application perspective. This, however, can be very expensive and difficult to administer.

In addition to these benchmarks, there are certain aspects of a CDN’s network configuration that can also be considered, depending upon your exact requirements:

– Number of POPs

The more points of presence a content delivery network provider has, the more bandwidth and customers it can potentially handle

– Network reach

If you’re looking to provide your online service to either a specific continent or around the world, ensure the network has server location in your target markets

– Current customers

This fairly unscientific method is based upon the fact that you can base your CDN provider decisions upon which providers similar sized organisations to yours are using. Furthermore, larger, more reputable organisations with high data acceleration needs tend to use faster, more efficient networks. These are both good things to look out for.

Whilst there is no substitute to trying out a particular content delivery network provider for yourself by utilising many of the trials they make available, these measurements can go some way towards helping you decide which is the best CDN for your organisation’s needs. Taking the time to consider these factors will help make sure that you get the most out of your online acceleration services.