Docker And The Rise of Open Source

There’s never been a phenomenon like Docker. Eighteen months ago, the company took its core technology, which enables IT people to move software easily between different machines by enclosing it in “containers”, and made it open source. Almost immediately, things took off: since last summer, Docker has been downloaded 43 million times; there are over 30,000 “Dockerized” apps in the Docker Hub Registry, and 13,000 third-party projects on GitHub using Docker. At AWS Re:Invent last week, its little booth on the show floor was overrun with people wanting to learn more about the company.

While its history and trajectory are unique, Docker is a more general example of a new breed of open source companies, which are building long-term, sustainable businesses. Companies like Cloudera, MongoDB and MuleSoft have revenue run-rates between $50-100M, strong growth, and huge market opportunities ahead of them. That’s new, in that historically only Red Hat (the grand-daddy of open source) has managed to scale a profitable business; many of the best companies that followed in the 2000s had successful acquisitions, but struggled to grow revenue.

Why Now?
So, what’s changed? Why is it possible to build a big open source company in a way that it wasn’t before? There have been 3 big developments over the past 10 years:

1. Open source technologies are finally good enough:
“Free” is a compelling value proposition, but it’s not enough. Open source also has to be better, and that’s what’s happened over the past 10 years as open source technologies have improved at a faster rate than proprietary software products. People choose Hadoop, Kafka or Mongo because it’s the best tool for them to achieve their objectives, and not just because it is free. A big enabler of this has been support for open source at leading firms, especially Google and Facebook. Projects like Android, the Chrome browser, and Open Compute have given open source legitimacy, making it easier for mainstream enterprises to embrace it and pay for it.

2. Distribution through developers:
Open source appeals to developers as a better way to build software. In recent times, developers have become more skilled at leveraging open source and have grown to have a greater voice in what technologies to use than ever before. They are also very collaborative, openly sharing experiences online. This has enabled open source vendors to go around traditional corporate IT buyers, and instead get viral, grass roots adoption from the community.

3. New approach to licensing:
As Peter Thiel recently explained to a class at Stanford, it’s not enough to create value; if you want to build a successful business, you also have to capture it. That’s been a challenge for open source companies because historically they have licensed their software under a general public license (GPL), which prevents them from charging for any software. It’s also one reason why HortonWorks, which licenses under GPL, had lower revenue than many expected. By contrast, most of today’s open source companies license under an Apache software license (ASL), which lets the vendor decide whether to share code changes with the community or keep them as proprietary. Under this hybrid or dual-license, open source companies can charge for proprietary product extensions, and not just for services and support, accelerating revenue growth and improving margins.

So What?
There are two big implications of all this. One is that we should expect to see more open source startups. Recent examples include Confluent, built around Kafka (message broker), and Databricks, built around Spark (data processing). The second big implication is that almost every category of infrastructure software will come under pressure from open source. As the table above shows, aside from in security, there are open source vendors snapping at the heels of almost every large proprietary vendor.