When individuals release open source projects, their motivations are often altruistic. Their intention is to give back to the community, get as many people as possible to help improve the project, or just to release what they think is a useful tool to others.
Although the general perception of open source has definitely advanced sinceMicrosoft's "un-American" comments, the best companies are not open sourcing things for the altruism. There are real, strategic reasons hidden behind the warm and fuzzy glow of open source.
Google and Microsoft are good examples of this.
Google Cloud and open source
Google has a major advantage when it comes to container tooling because it effectively invented the concept and has been running containers in production for almost a decade. As a result, Kubernetes is perhaps the most famous of Google's recent open source releases. For the uninitiated, Kubernetes is a cluster and container orchestration framework that allows you to schedule workloads across many nodes.
The goal for this project seems to be to provide a standard platform for deploying and managing containerized environments. Developers will build their infrastructure using Kubernetes because it's free, and it has good documentation, well-designed APIs, and a growing ecosystem.
This works well for experiments and personal projects, butin the words of AWS's CTO, Werner Vogels, all this management is "undifferentiated heavy lifting" that you really should be outsourcing to an experienced provider.
AWS has a massive head start with its huge cloud portfolio, but Google's Container Engine happens to be what amounts to a hosted SaaS version of Kubernetes. This means that if Kubernetes becomes the default way of running containers, Container Engine could become the default managed, supported service for it. Managed, supported services are things that businesses like.
The recently open sourced Seesaw load balancer follows exactly the same approach. Sure, you can run the load balancer yourself, but you can get the exact same functionality with Google Cloud Platform's load balancer products. They come with elastic scalability, support, an easy-to-use-control panel, and APIs.
Another Google project that may follow the same example is TensorFlow, the machine learning framework. This used inside Google's massively distributed data centers but the open source version is limited to a single node. It seems inevitable that Google will release an elastic machine learning service running on Google Cloud Platform that utilizes TensorFlow.
Microsoft and open source
Google's competitive advantage is its advanced infrastructure. Cloud Platform is its attempt to productize that, and open source is a great way to drive adoption. You can run it yourself, but a cheaper, easier and more flexible approach would be to let Google run it for you.
Sign up for MIS Asia eNewsletters.