Sustainable Open Source Development

Current trend in software deployments is to rely on open source software for entire production stacks. You can find open source software in the core technical stacks of every startup out there, I’m told. If you’re using Cloud based offerings, most of Cloud providers are running Free/Libre Open Source Software as their foundation.

This article is a deep dive into the economic models behind successful open source projects and communities, and how as a professional, enterprise grade user, you depend on the long-term sustainability of all the open source projects you’re using. And because you depend on the projects you’re using to be successful, how to contribute and guarantee their success.

You can choose to deploy your web application or services using django, rails, nodejs, some Go based framework such as Iris or Revel. Or PHP frameworks Symfony or the new kid in the block Laravel. Or you might roll out Java™ based technologies such as Tomcat, JBoss or GlassFish. Or you might prefer to use a .Net environment.

There’s something common in the technologies listed above, and many more you’re most certainly relying on: all of them are Open Source Software. It means you can access the source code, run it locally yourself, play with it and hack it away either to learn how it’s done or even to fix bugs… or to add new features. Then you’re welcome to contribute your improvements to the community, so that everyone can benefit from your awesome work!

Why is Everybody Using Open Source?

Open Source Software has not always been the first choice of enterprise as pillars of their production infrastructure. It used to be that company decisions would favor using proprietary software, closed source, with expensive maintenance contracts. What it meant is that every single customer of a given vendor would separately pay for the same bug fixes, unknowingly.

There was such a time when Open Source Software quality was not there yet, so that you could not rely on them for running your production. And before that, maybe the software wasn’t available at all, or for your particular hardware choice… yeah it used to be that people had a choice to make amongst more than just 3 hardware architecture…

So how did we arrive at the current situation, where it doesn’t make sense anymore to use proprietary and closed source software for most of the things? Investment is the answer. Individual hackers and contributors have investing humongous amounts of time into their little Open Source Software alternatives, up to a point where our Free Software catalog is running the world.

Who paid for this initial investment? Public research labs and facilities such as those running with universities, individuals on their free time, and then slowly some private companies who allowed some of their staff to use their paid time to contribute, up until when it has become obvious that the best solution both in terms of economics and in terms of quality would be to contribute to the common goods, our beloved Open Source Software.

Many people are still investing their own free time, by the way, and that’s how we now have those advanced core technologies to rely our products on. Standing on the shoulders of giants, as they say.

Sustainable Open Source

More and more Open Source Software contributors are paid to improve Open Source Software, and most of them to that in the open and actually contribute to the software that is available to everybody. That’s the right way to do it when handling Common Good.

That said, with the rise of Cloud based computing, more and more people are also paid to improve Open Source Software without contributing their improvements back to the communities. That can been seen as a loop hole in the license terms, or a plain refusal to understand the rules. Most importantly, it’s all about taking for yourself what was freely offered to be shared with you.

Open Source Software nowadays is mainly maintained in one of those possible organisations:

• Open Source Editors

  Some Open Source Software are now developed, maintained and released by commercial editors for whom it makes sense to offer their solution under FLOSS licencing.

• Independant Communities

  Most Open Source Software projects don’t have an enterprise editing them. Instead a community of individual contributors are working on the project, either as part as their daily job or most often, on their free time, because they like the project.

• Single Person Projects

  And a huge number of smaller software projects are still taken care of by a single individual who is devoted to his project, for some reason or another.

For some projects, the sustainability isn’t much of a concern, because of a very small and motivated user group, who will either find another solution or learn how to maintain the software if needed. That’s mainly for software you would use as a hobby.

For other projects though, I’d say as soon as when the project is used in an enterprise setting, then ensuring a mid to long term sustainability is crucial to the project users. Because now an enterprise business depend on the availability of some external software.

As an enterprise using Open Source Software, you should learn enough about the project’s organisation to assess if it’s reliable for you to use it. When it’s not, then you have two ways to make the software reliable for you:

• Contribute time to the software

  That’s as easy as allowing proper paid time to your staff to contribute to the Open Source Software you’re using, in particular when the project does not have the backing of a professional editor behind it.

• Contribute money to the software

  If you can’t have your staff devote time to learn the project’s code and how to contribute to it, then you need to make it so that the project is in a good position to do it themselves when you need them to. That means contributing money to the project so that they are able to pay for other hacker’s time — after all time is money, and you’re not willing to put time on the table, so put money on it, to much the same effect, right?

Open Source Editors Business Models

When an Open Source project is edited by a professional editor, then we can observe a short list of business models that are often used. The French free software work group for research and development investments Systematic has published a paper about those business models.

The most common open source business models for software editors are:

• Foundations

  Foundations typically involve several companies working together on the same software, in order to lower their R&D costs. They’re funded by a group of enterprises who would each need to maintain the same source base anyway.

• Double Licencing

  The software is available either as under FLOSS license terms, using either the GPL or the Affero GPL license, or under a prorietary license for those users who don’t want to publish their own software as a FLOSS project. Revenues from selling proprietary licenses allows to invest in the FLOSS version of the software.

• Open Core

  The core of a product is Open Source and available under FLOSS licencing terms, and the editor also builds a proprietary product on-top of the FLOSS offering, with an additional feature set. Revenues from selling the extended product allows to invest in the basic FLOSS version.

• Cloud

  The editor maintains an Open Source Software and also offers a managed SaaS solution. Again, a part of the revenues from the offer is invested back into the Open Source Software maintenance and improvements.

• Professional Open Source Services

  The editor contributes to an Open Source Software and then sells professional services to its users, such as follow the sun 24/7 support that includes upstream bug fixes, specific software builds and maintenance, on-demand feature development of either extensions or new features.

PostgreSQL itself is very lucky in that several companies are contributing to the project while entertaining at least one of the previous business models. That’s how PostgreSQL came to be the world’s most advanced open source database and continues to strive at a very high pace, releasing every year a new major version packed with improvements, fixes, and new features.

Single Person Projects Business Models

So with people getting paid to contribute to Open Source Software, we could think that the times or yore are over now, and all is good in our world.

While some highly visible Open Source Projects have the opportunity to attract direct contributions from smart corporations who understood the economics behind it, lots of smaller projects, often one-person projects, are not so lucky, and are actually dependant on the willingness of a single contributor that handles the project on their free time.

As an enterprise user of such a project, you should wonder how reliable the situation is, and how to make it actually reliable. As mentioned before, it’s all about you being able to contribute to the project either time, money, or both.

When developing and maintaining an Open Source Software as a single individual, how do you make it so that it’s possible for you to sustain your activity?

The least you have to do is make it so that you can pay the bills and support your family, and maybe it woud be best for the project to benefit from its maintainer to be able to spend more than just free time on fixing bugs and improving the software.

Now, how to make that happen is still a gray area, and I think we need to find better answers than are currently broadly available. I like The Varnish Moral License idea a lot, and another widely adopted solution consists of using a crowd funding solution for sponsoring a specific feature set.

The Varnish Moral License

The Varnish Moral License idea from Poul-Henning Kamp is quite simple. In his own words, here’s how it works:

• Happy varnish users tell me that they want a Varnish Moral License
• I send them an invoice from my company
• They pay the invoice
• I develop Varnish

A Moral License to make any kind of big or small money running Varnish on your website, without feeling the least bit bad about the poor sucker who created the software, his two kids, his old cat etc etc.

On a more concrete level, the Varnish project gets more of my time and attention which I use to improve the tool which makes your website work so well.

Somebody started the rumour that the name was chosen to sneak this expense into your IT departments budget, along with other software licenses, rather than put it under your marketing department, where all sponsorships properly belong and where it would compete with the local soccer team.

That rumour is very possibly true. I might even have started it myself. Right here even.

Let’s Talk about pgLoader now

As you might know already, I am personnaly involved in a Single Person Open Source Software Project: pgloader. This project implements loading data into PostgreSQL in a fully automated way, to enable the following two use cases:

• Load data from files with automatic error handling.

• Implement Continuous Migration from your current database to PostgreSQL.

As you can imagine, the pgloader project is mainly used in enterprise setups, and so we now have a bunch of companies who depend on my abilities to single-handedly maintain, improve and release the software.

On a personal level, I’m lucky enough to get paid to work on an awesome PostgreSQL Extension — check it out at Citus Data, the Worry-Free Postgres for SaaS.

This is a classic Open Source paradox: my contributions to PostgreSQL allow me to have a very good job at a great company, and I also have a bunch of projects I have started along the years and that I maintain on my own. Some of those projects certainly deserve more attention that what I can offer on my free time. Those projects include pgloader of course, and also the PostgreSQL prefix extension that I know is used in production in several telecom companies, or the PostgreSQL Extension White Listing that is used in production at several PostgreSQL SaaS providers, and which should now be deprecated by the more ambitious project pginstall… which never took off because I couldn’t make it easy enough to deploy, basically.

What I would like to achieve is an organisation where enterprises that are using pgloader can help sustain the project in the mid to long term. Contributing time to the project is already easy, but seldom happens. After all pgloader is not an infrastructure level project, it’s more like a background job that you script around then mostly forget about. So if the product works well enough to solve your immediate problem, that’s good, and now you’re already doing something else.

And if you don’t have to contribute to pgloader in order to successfully use it, then why would you bother investing time on the project? While that’s understandable, it leaves the whole idea of the project being sustainable unanswered.

That’s why I’ve been setting up funding solutions for the project. The idea is that I could then hire fellow Open Source Hackers to work on the project with me and improve it, fix bugs, develop new features. We have a long list of improvements to make to the project, and I often talk with people from different companies who would like to see them happening.

Of course, I don’t know anyone who would hack Oracle™ support into pgloader on their own free time so that some big companies can now easily migrate their legacy systems to PostgreSQL, cancelling huge amounts of license fees in the making. That said, I know several Common Lisp hackers who are for hire and would be happy to be paid to contribute to enterprise grade Open Source Software solutions…

The pgLoader Moral License

As I really like The Varnish Moral License idea, this funding option is of course available to pgloader. If you’re an enterprise using pgloader, consider contributing to the project either time or money, or both:

The pgloader Patrons Membership

The pgloader project is fully Open Source and released under The PostgreSQL License so that anyone can easily contribute to the project. The whole project management (issue tracking, feature proposals, etc) happens on the pgloader github page, in public.

The pgloader Moral License is a one-off payment — that you can do as many times as you like. That’s a good way to show support for the project and allow specific feature development, but maybe that’s not enough for me to hire Common Lisp hackers and improve the many little things that need strong focus and attention in the project.

That’s why I’ve also setup a recurring payment solution, the pgloader Patrons Membership, that allows a monthly to yearly commitment to sustain the project improvements in the long term:

Conclusion

In this very long article we’ve been talking about how to sustain Open Source Software development both at the editor level and at the individual level. While several business models are available for both situations, it’s still not that easy in 2018 for enterprise users of Open Source Software to ensure that the projects they depend on are sustainable.

We could wait until people like GitHub or GitLab integrate Open Source project funding options, or we can try to solve the problem at our individual scope, one project at a time, like Poul-Henning Kamp is doing with The Varnish Moral License.

Let’s build a bridge between companies using Open Source Software and individual contributors using their own personal time to maintain and improve innovative solutions!