The Machine Learning Wars: CDNs vs Google vs AWS


Is there an innovation so disruptive that it poses an existential threat to the CDN business model? So disruptive that it can implode an entire sector in one year? Fei-Fei Li and Google have spoken, and in doing so they have declared war on the CDN industry. Unfortunately, some industry players don’t see it because ML is still a vague concept that has yet to make any meaningful impact.

AWS dominates the cloud market, but Google dominates the AI field. The progress Google has made in the last two years in the field of AI is the most revolutionary feat ever achieved in the technology industry. Machine learning, which is a subfield of AI is a game changer and threat to the CDN feature set. The learning algorithm, deep neural net, recurring neural net, convo neural nets, supervised learning, unsupervised learning, classifiers, and everything else in the ML field is the new CDN feature. Today, content delivery performance is taking a back seat to functionality. What is better? Shaving 2 seconds off of page download times or having a learning algorithm that serves a highly customized page to a visitor that canĀ  double sales per visit. That’s the power of machine learning. Can CDNs really co-exist in a world dominated by Google AI?

Google Cloud vs CDN Infrastructure

Today, Google has an advantage in the ML software stack because it has developed lots of it. However, they don’t have an advantage in the PoP infrastructure. We would argue that its a level playing field in the infrastructure. Fortunately, a lot of the software tools, frameworks, libraries, datasets and code that make ML possible are open sourced. Thus, CDNs can use the open source ecosystem to building out an ML ecosystem for the future. And a Fei-Fei Li is not a staff requirement to build up core competencies in ML. We’re sure even those with Masters in Mathematics and Computer Science can do a fine job in creating and/or optimizing the required algorithms. From hear on out, we are going to add machine learning to our research efforts.

Scroll to Top