Content Delivery Networks, or CDNs, greatly improve end-user performance (download speeds) when compared to origin servers, due to the geographic proximity of their far-flung infrastructure. Recent performance data shows that images load more than 60% faster from a CDN vs. an origin server, while CSS loads more than 81% faster.
While impressive, these stats can give enterprise customers a false sense that CDNs replace front-end optimization needs. This is not the case, as CDNs can inadvertently introduce various kinds of performance problems. Many enterprises rely on monitoring to compare origin server vs. CDN performance during the evaluation process, but then stop there—not fully understanding how important it is to continue monitoring over the long-term.
In short, enterprises must be responsible for managing the performance of their chosen CDNs just like any element of their own infrastructure. Here are three core approaches for monitoring or optimizing your CDN’s performance:
Monitor Your CDN, With An Emphasis on the Backbone and Last Mile
CDNs can introduce Domain Name System (DNS)-related challenges. When an organization deploys a CDN, their DNS is relegated to their managed DNS provider, who in turn relegates it to the CDN. Over the years we have seen many breaks in such chains leading to errors. This is one reason CDNs need to be monitored continuously.
However, it is also not enough to rely on cloud-based monitoring to test a CDN at regular intervals to ensure it is responding quickly and reliably. While this approach can give peace of mind regarding the overall health of a CDN’s infrastructure, it does not confirm that end users are enjoying a fast experience.
We recently saw an example where a particular CDN was unable to deliver content to mobile end users on the east coast. Last mile monitoring revealed the problem was not within its own infrastructure, but rooted in a peering relationship with one particular mobile network operator. This enabled swift action to be taken in the form of traffic re-routing. Had backbone or last mile monitoring not been in place, no one would have known that the end users were experiencing problems in the first place.
Additionally, CDNs are responsible for mapping end users to the nearest or most optimal CDN edge servers. If end-user load times are slow, this is a sign that mapping may not be optimized—another reason backbone and last mile monitoring is essential.
Ensure CDNs are Mapping Properly
CDN mapping is a highly complex process that depends on various factors at any given second, and organizations generally do not have visibility into their CDN partners’ internal routing dynamics. However, it is possible for enterprise users to detect when a CDN’s mapping patterns may be suboptimal, and raise this to their CDN’s attention.
This capability depends on measuring end-user performance from the largest, most comprehensive number of vantage points possible, to reveal what end users are experiencing. Companies invest in a CDN to deliver good end-user experiences for all their end users across all regions. However if there is a major discrepancy in the time it takes to download similar content, then this is a sign the CDN’s mapping patterns are likely misconfigured.
A well-known user organization recently deployed this technique with impressive results. This company compared end-user performance levels across various city groupings—several in India, and several in Brazil— and noticed significant discrepancies. The user flagged the issue to its CDN, which adjusted the mapping configuration, and as a result, improved average download speeds by over 34% in India, and over 18% in Brazil.
Optimize Your Front End
Organizations using CDNs are often lulled into believing that they can be lax when it comes to front-end optimization—or, packaging website content in a manner most conducive to speedy delivery. The “fatter” a web page is, the slower the content will load, and this is true whether an enterprise is using a CDN or not.
Numerous analyses have shown the greatest positive speed impact comes when these two techniques are used in tandem. Put another way, a CDN helps to fetch, load and deliver content faster, while front-end optimizations helps package the content in a manner most conducive to speedy dispatch by the CDN.
Some of these services are offered by CDN providers where they apply some of these optimizations on the edge which adds to the complexity of managing end-user experiences. This makes it even more important to measure this experience from the backbone and last mile.
Examples of front-end optimization include:
- Image Management: Automatic transformation based on the end-user device.
For all their value, CDNs are not immune to causing end-user performance degradations. Considering that some enterprises are spending millions of dollars on CDNs annually, it becomes imperative for them to consistently monitor their chosen CDNs and identify and leverage opportunities to optimize performance, in order to maximize their investments. Keeping a close eye on end-user performance at the last mile, identifying misconfigured mapping and continuing to emphasize front-end optimizations will go a long way in helping enterprises ensure their CDN investments pay off.