What your CDN provider doesn’t want you to know…
Content Delivery Networks have been around for almost 20 years now. Depending on how you look at it, that fact could be a good thing or a bad thing.
The good: it’s a tested and proven technology.
The bad (and what CDN providers don’t want you to know): it’s not a technology designed for today’s cloud-based, personalized, bi-directional, real-time content.
The ugly (and what legacy CDN providers really don’t want you to know): legacy CDNs aren’t designed for enterprise content which is far less popular (and therefore ignored) compared to content generated by large video and social media companies.
CDN vendors have been closely watching the evolution of the World Wide Web in the hopes of catching up. In fact, the primary factor that has influenced the advancement in CDN technology until now is the nature of web content.
In the 90s, the web used to be a collection of dull-looking, simple websites with static content. The first set of CDNs focused on accelerating static text and images through caching. The goal was simple: duplicate content and place it close to end users.
However, in terms of infrastructure, this was a massive challenge. First-gen CDN providers spent millions of dollars building out this infrastructure, investing in data center points of presence (POPs) at locations where the majority of the web content was being consumed. It was standard practice for CDN vendors to boast about the size of their networks in terms of number of servers or their POPs around the globe because back in the 90s that was the number one way to excel in terms of performance.
The evolution of the web and what it means for business
Fast forward to 2015. The web has been transformed by rich media, dynamic content, personalization, mobility, the cloud, and much, much more. The web is far more ubiquitous than ever before.
Static text and images are now passé. Video and interactive content have taken dominance, among all the diverse content types. Dynamic bi-directional content, web applications, APIs, third-party content, real time data, personalized delivery, SaaS and the cloud – are all part of this transformation.
Even FCC chairman Tom Wheeler has proposed treating the Internet like a utility, so everyone (and every business) can get fair access to all websites, without having to fear being pushed onto slow lanes by Internet monopolies.
From an enterprise viewpoint, CDN realities have changed. As dynamic and interactive content have taken precedence over static data, CDN vendors no longer need a thousand POPs. Current best practices in the industry suggest that newer vendors today can do just as well, with a smaller POP footprint in strategic locations with a higher focus on the optimization software stack – this evolution is partly because of better acceleration techniques, as well as improved broadband speeds and better last-mile ISP networks within local regions.
Moreover, enterprise customers rarely have access to the entire POP footprint and end up getting served by a few POPs or a reduced region; both, for technical and economic reasons. In the case of dynamic/bi-directional content, a bigger POP footprint is not an advantage at all. Since the content needs to be retrieved all the way from the origin server, what really matters is the quality of the end-to-end connectivity and the optimization software stack.
However, given the million-dollar investments that market leaders made into establishing their massive POP presence, these players seldom miss an opportunity to showcase their muscle in terms of the same. This is often misleading for enterprises that need to accelerate non-static content.
Enterprise customers thus need to evaluate the nature of their content prior to evaluating CDN vendors, and then decide what’s best for their needs.
Enterprise customers also need to be wary about CDN vendors who claim to do “dynamic” delivery. Industry analysts consider dynamic delivery to be the future of the CDN market. Multiple CDN providers have hopped on this bandwagon in the hopes of benefiting from the buzz – and that’s what most of their promises about dynamic content delivery are: empty buzz.
It is important to understand what true dynamic delivery is. Real dynamic content is never cacheable. Rather, it needs to be retrieved from the origin server every time. Such content is personalized and is generally seen in web applications, user generated uploads, interactive SaaS platforms, etc.
However, most solutions in the marketplace offer either copycats made to look like the real thing or a piecemeal solution – either low TTL (time to live) settings, which force most requests to go back to the origin, TCP optimization from the origin out to the edge or persistent connections or both. In fact, content with a short TTL is NOT dynamic content at all.
The true judge of the character of Dynamic Content Delivery is the middle mile, its intelligence, and its functionality. This functionality must include middle-mile network quality, either using a private network or routing techniques that bypass Internet congestion. It must also have the proper intelligence, including TCP optimization at each segment of the network and persistent connections.
The intent of this blog is to empower you to make the best choice for your CDN platform. And if you’re an enterprise, that list of possible vendors is quite short.
To learn more about how Aryaka accelerates dynamic, personalized content, check out our CDN data sheet. It will walk you through how we solve the problems other CDN vendors don’t want you to know about.
And the next time a CDN vendor tries to sell you on “dynamic” content delivery, be sure to ask them what exactly they mean by that. If they’re trying to sell you a 1990’s solution packaged in a 2015 wrapper, you’ll know you shouldn’t walk, but rather run away.