Every disruptive change has its pioneers. Someone must be the first to think of it, someone must be the first to launch it, someone else could be the first to use it at scale. While most observers agree that Amazon Web Services are an undisputed pioneer of cloud computing on the supply side (i.e., among vendors and service providers), year 2010 saw the emergence of Netflix as one of the pioneers in the use of cloud computing by enterprises.
Nextflix, many say, is not a typical enterprise. Even though it’s a big company (employs more than 2,000 people according to its press kit) that is publicly traded on NASDAQ with current market capitalization of slightly over $11B, it’s relatively young (founded in 1997 according to Wikipedia), caters primarily to consumers and its IT is their business (as opposed to a typical enterprise where its IT supports its business - when Netflix was a DVD mailing company, its IT supported its business; as Netflix is transforming itself into a content streaming company, its IT is becoming its business).
Netflix has always been a bit geeky. Their recommendation algorithm has been their prized asset for some time. Also recall, for example, a competition they ran in 2006 offering a big reward to anybody who would develop a better one.
I started following Netflix’s cloud use in 2010. Netflix is a big operation, possibly even regarded by AWS as their “reference” customer. If you follow cloud computing, you couldn’t have missed it.
I watched several presentations given by Adrian Cockcroft (see his interview with Randy Bias) and subscribed to techblog.netflix.com. The latest post there is full of cloud wisdom of a cloud practitioner.
For example, we learn that Netflix went to cloud in search for high availability and agility was almost like a side effect. We learn that Netflix does not want to be a datacenter expert because they regard it as an accidental complexity. Being a consumer brand in a relatively new market segment, they did not want to worry about getting their capacity forecasts right.
If you look around the web, you will find that Netflix is running a Java application stack, which makes it close to many other enterprises out there. But the key components that sets them apart is their internal operations platform tailored for Amazon EC2 - from development to QA, to deployment, to monitoring, to trend analysis, to troubleshooting. (It’s important to realize that an “internal operations platform” is not only software - it’s also a set of processes, standard operating procedures, mind set and operations philosophy).
And here is the point of this post (finally!). Netflix essentially has built an enterprise-friendly world-class PaaS for Java. They probably built it without thinking of selling it as a standalone product to other enterprises one day, but I would like to ask - why not? If the world's biggest web retail operation managed to build a hosting business, why can't one of the world's geekiest enterprises build an IT ops platform business as well?
Think about it - Netflix Web Services, Java application hosting, for enterprise by enterprise…
Can’t say for sure if it will happen or when but my gut feel is that it will happen eventually, and if not Netflix then some other cloud customer will enter application hosting (PaaS) business with a platform originally developed for internal needs.
Netflix folks, are you reading this?