Some Caveats on The Case for Supporting Open Source Infrastructure
|Sep 29, 2020|
For Lincoln Network, I recently published a piece on The Case for Supporting Open Source Infrastructure. While the reception of the piece was positive, one person brought up the question of Red Hat. What about Red Hat indeed?
Corporate Sponsored Open Source
While my piece was mainly focused on how the government could support open source software more, the private sector in some places does support open source software strongly, most obviously seen with companies like Red Hat.
As seen with a quote from how they see themselves in the open source universe, they develop their product using an open source model. Their emphasis though, is less on the open source model, then it is running a traditional enterprise business to support the open source software, so people don’t have to go to Red Hat for support as much
"We’re an enterprise company with an open source development model."
Red Hat produces a distribution of Linux (Red Hat Linux), which they maintain themselves. While the software itself is free, support and on-premise Red Hat Linux will cost you. The support open source software model that Red Hat pioneered has been taken up by other companies as well. A sample of other companies that do this or a similar model below:
MongoDB | MongoDB
Cloudera | Hadoop
Docker | Docker
Confluent | Kafka
Variations on this pattern include that the project is open source on GitHub (MongoDB on Github for example). However, doing all the installation and related work is hard, and people may not have the time to do it. So for a fee, the company will offer to do the hard parts for you (hosting), so you can focus on using the database for whatever purpose you need be, instead of setting up the environment for the database and making sure things to don’t go terribly wrong.
While the model for Red Hat and other companies (MongoDB, Cloudera, Confluent for example), is profitable, there are several downsides to this model. For starters, there is always a chance that the open source product could be copied by a competitor. This has happened to companies like MongoDB, where the database itself was copied by Amazon when Amazon rolled out DocumentDB.
Secondly, having a corporate run open source project often requires the open source project to be of a certain size. Could Vue.js be run by a corporation? Probably, but it would be a challenge to figure out how to monetize it. Something the size of left-pad or event-stream for example, would be even harder to monetize at the level of a corporation.
There is nothing wrong with the corporate model of sponsoring open source. In fact, it has resulted in many innovations, mainly in the realm of databases. However not every open source project, even if it is highly critical, may fit under the model of being primary developed on by a corporation. In this case, as I argued in my piece, using the OTF as a vehicle to support critical projects that may be difficult to monetize may fill some gaps.
For a newsletter on visa issuance data I am working on, I filled a FOIA request to the State Department to ask about some discrepancies in the dataset.
My friend Grant Dever recently wrote about 1000 true fans the phenomena with his blog post 999 Real Fans.
I recently discovered The March of the Siberian Riflemen, a march from the Russian Civil War. It’s a very good instrumental to do work to.