As the head of policy for Open-Xchange, I spend most of my time travelling around the world and advocating for open source. In the last year or so, I was often met with the question: “So, is Dovecot going proprietary?” The answer is clearly a no, but the experience made me think that an explanation of what is happening from a policy and business perspective was due.
In the open source community, we regularly see developer teams move from a set of volunteers to a company as they grow. This step becomes almost unavoidable once the project becomes more complex and requires more and more time, as volunteers often cannot get their employers or customers to sponsor their work beyond a certain extent. There are intermediate options, such as getting a foundation to host your project and relieve you of most of the non-technical work, yet in the end, good code needs focus and quality time; unless you can find a sponsor or a grant, you need to start making revenues from the project itself.
Dovecot made this step long ago: its namesake company, Dovecot OY, was established in 2011. Since the beginning, Timo Sirainen announced that he would adopt an “open core” model. This meant that most features would still be included in the open source codebase, but a few advanced ones, the ones most needed by large commercial users, would be kept under a proprietary license and provided in exchange of, guess what, money. In turn, this money would be used to pay the salaries for the developers of the entire code, proprietary and open source, ensuring that the free version could continue to thrive over a solid base and receive all the energy and care that it needs.
This model remained unchanged when, in 2015, Dovecot became part of the Open-Xchange family. Our guiding principle was not just the openness of the code, but the openness of messaging as a whole. We continued to offer to all Internet users the option to run their own email server without having to pay a license or buy proprietary software; we also ensured that even large commercial deployments, while sometimes requesting proprietary extensions, would keep using interoperable open standards. The commercial synergies between Dovecot and the OX App Suite front-end platform allowed the number of Dovecot developers to grow; the community edition of Dovecot, the one you can and will still find as the default IMAP server on almost all Linux distributions, has remained free, relevant, and cutting-edge, and has continued to expand steadily.
Then, as in any well-managed project, the moment came when a deep refactoring of the codebase was necessary. In fact, the deployment models for large server-side applications have evolved so much in the last ten years that a brand new architecture was necessary; at the same time, preserving backward compatibility with older deployment models and maintaining the related code became an unjustifiable burden, which distracted us from working for the future.
This is why, while developing the new Palomar architecture for the commercial version, we also decided to clean up some of the old features that were available in the community edition and that were not any more in line with state-of-the-art technical practices, nor used internally or requested by any paying customer. Our assessment is that this will still make the community edition the best in class for almost all its current users; it will guarantee its long-term availability, free as in free beer and free as in free speech, as always. On the other hand, business users that really need the features designed for larger deployments should use the commercial version and contribute financially to the development of both Dovecot versions.
All in all, we think that the open core business model is the only sustainable one for projects that, like Dovecot, have reached global ubiquitousness and need to prove their reliability not only at the technical level, but also in terms of support and business continuity, that is, in terms of the continued existence of a successful company leading and promoting the project and funding most of its development.
This last element is vital if we want open-source offerings to compete in the top-tier installations market with large proprietary software makers and big tech walled garden operators. For their most crucial services, top-tier operators do not want to rely on non-commercial projects or even on small companies. They want to deal with companies that are sufficiently large and financially stable to ensure that the software will not disappear or become unsupported in the blink of an eye.
If we want mass consumer services to be based on open source software, if we want alternatives to big tech and big oligopolies to emerge, it is vital that large commercial users pay the developers for the most advanced features they need; we think that it is also fair.
This does not betray the spirit of free software in any way. Other companies have taken their open codebases and made them proprietary; we do not do this. In fact, even the code that was removed from newer versions of Dovecot CE is still free and available. We are trying to steward the Dovecot project in a responsible way while keeping it open source in the depths of its soul. We intend to live up to a world of expectations, and we thank you for almost three decades of adoption and support.
SHARE