Part 2: Deep Dive into the Apple CDN

Apple has a few challenges in building out a global content delivery network. Having the massive financial resources at its disposal for CDN build-out definitely helps, however, that satisfies only part of the equation. AT&T had massive resources at their disposal, when they built out their CDN, plus they had decades of experience building out global networks, but in the end they decided to go in another direction, and partner with Akamai. Apple’s first challenge is the Talent Gap, as previously stated. Apple’s next challenge is building the optimal CDN POP count model for distributing its content to a massive global audience. There are 195 countries, and 300+ Internet Exchange Points (IXC) globally. Selecting the right mix is going to be a daunting task.

2nd Challenge – The Optimal POP Count

For Starters, Apple can mimic the POP count model of Limelight, EdgeCast and Level 3. These three CDNs have selected the optimal CDN POP count through years of trial and error. These three CDN have around 30 POPs in different regions of the globe, including the major IXC in Frankfurt, Amsterdam, London, and so on. Apple needs a higher CDN POP density than 30, due to the sheer number of iPhone/iPad/Mac users. As of December 2013, Apple was estimated to have 300M active iPhones, with ABI estimating 44% annual smartphone growth. In addition, more than 170M iPads have sold since October 2013. In 2014, it’s likely that the Apple mobile device count will range anywhere from 550M to 700M devices.

That means having 1 or 2 CDN POPs in mainland China, India, Brazil, Russia, and other countries isn’t going to cut it. In the US CDN POP count model, ten is sufficient for high volume content delivery. In the BRIC (Brazil, Russia, India & China) countries, we can assume that 10 CDN POPs should suffice, with a minimum of 5 CDN POPs per country. Of course, these are just my high level guesstimates. The only CDN with enough high value market intelligence as to proper POP count per region is Akamai, and I doubt very much they will reveal this data to the public. Below are some POP counts I came up with.

Number of CDN POPs per Region
  • USA – 10
  • (France, Spain, UK, Germany, Nordics) – 25
  • Middle East – 5
  • Mainland China – 10
  • Australia – 2
  • Singapore & Indonesia – 10
  • Japan – 5
  • India – 5
  • Brazil – 5
  • Eastern Europe – 10
  • LatAm – 10
  • Russia – 5
  • Africa – 10
  • Total POPs = 112

Apple wants more control over the end user experience, and more control over their infrastructure, just like Google. So based on the estimates above, an ideal POP count for Apple is 112. This leads to the question, should Apple build out infrastructure at 100+ POPs? If Apple wants total control of their ecosystem, than yes, but it’s going to be 5 year deal. Or else, Apple can maintain some control of their ecosystem, and continue partnering with their existing CDNs for the BRIC countries, and other regions that are difficult to build in. Continued in next post.