At first, CDNs were little more than an interesting intellectual and business venture. ‘World Wide Web inventor Tim Berners-Lee… foresaw the congestion that was soon to become very familiar to Internet users, and he challenged colleagues at MIT to invent a fundamentally new and better way to deliver Internet content'1. This congestion can come in the form of flash crowds - a term for an unprecedented amount of internet traffic that overwhelms server resources. After one such flash crowd caused severe problems for some high-profile sites during the 9/11 terrorist attacks, awareness and investment into CDNs increased dramatically2.
Technological innovations such as proxy caching - which allows intermediate servers between users and content providers - and server farms - which allow server functionality far beyond a single server - laid the foundation for first-generation CDN infrastructure. Most websites at this point had pages that contained content that was the same each time the page loaded, e.g. HTML and images. By the late 1990s, significant improvements had been made in web acceleration for this kind of static content3. There was a focus on the optimisation provided by physical proximity to servers, since the transition from copper to fiber wiring was only just beginning, and only large corporations were able to afford these services4.
In the course of just several years, however, the strain on first-generation CDNs escalated rapidly. Internet users in general gained faster connections, more powerful hardware, and a stronger demand for quicker downloads. Likewise, the very media that they demanded became richer and more dynamic - qualities that the old system simply couldn’t handle anymore. Second-generation CDNs emerged in response, having a much greater storage footprint in fewer locations as opposed to the scattered network topology that was prevalent before5. This generation also made forays into cloud computing and peer-to-peer techniques for content distribution.
The third generation, where pretty much anybody with a website could realistically obtain a CDN’s services, featured the main challenge as distributing mobile content ‘because it is, by definition, more personal, more dynamic, and harder to cache’6. Different CDN providers focused on different solutions: one called Cotendo, before being acquired by Akamai, collaborated with Google on a more efficient protocol to replace the problematic and decade-old HTTP6. Server topology became even more highly consolidated. Likewise, security became far more of a priority as the use of CDNs became global.
Based on this deployment history, it would seem that CDNs are going to be playing an ever-increasing role in future internet content delivery, and are showing no signs of going obsolete.
Images from http://openclipart.org, sourced from the public domain. This page was written by Ali Yepifanova.