Want to know the future of FOSS? You can look it up in a database
If you make it here, you can make it anywhere
Opinion In IT, there is sexy tech, there is fashionable tech, and there are databases. Your average database has very little charisma, however. Nobody's ever made a movie about one.
They should. They should make lots of movies. (The Reg must note at this point that we're not counting the vendors in this. Some of them have, indeed, spent a bit of money on just such a project.)
You don't have to spend long in any aspect of IT to discover that databases are the soul of IT, its constant animating force. From one perspective, everything digital is a specialized database: word processors, spreadsheets, shoot-em-ups, streaming services, from Google to your disk filing system. The storing, sorting and retrieval of data? That's it. That's the whole game. It has been the case ever since Herman Hollerith designed the punched card tabulating machine in the late nineteenth century.
As for databases that call themselves that, they're the engine of corporate computing. Their capability, reliability and maintainability are essential, and the metrics of performance and expense are unambiguous. Corporate decisions about databases are one of the purest indicators of how IT is sourced and deployed. Hype is quickly exposed, as is the good stuff.
So when you look at the databases developers actually choose, you're seeing a market model with wider implications. Open source versus proprietary, hosted versus on-prem, innovation versus maturity: all primary concerns across IT, all crystallized in DB decisions.
But there's an equally important flip side: how the developers and suppliers of DB software manage to stay in business themselves. That's the other great question of IT in the 2020s: how do you make money either fighting or flaunting FOSS.
That's the first lesson from a feature discussing today's FOSS databases and their respective licensing terms: open source has won. It's about time too. Before FOSS was a corporate option, the big guys were ruthless at monetizing their position in the heart of IT.
Licensing models were set at what clients could bear, not what was equitable. Random audits could turn accidental license breaches into very expensive mistakes, and it could be very hard to manage those licenses if you were trying to scale. Or if license management was curiously difficult.
- Open source databases: What are they and why do they matter?
- Python tops programming love list – but if you want a job, learn SQL
- MariaDB buys geospatial specialist CubeWerx
- Oracle, IBM losing ground to local databases in China, says IDC
Why did anyone put up with this? They had to: these were the costs of mitigating the risk of ushering in unknowns to the galactic center of your company business model.
Times change, but memories abide. It is hard to overstate the organizational resentment towards what looks, feels and costs like extortion, or the readiness to explore options that do not have that particular pistol to point. Momentum has built for FOSS, as more people use and develop it. The quality and variety of the code has increased and deployment edged deeper into risk-averse, and rich, areas of the market.
There is a lag between what developers choose and what is actually deployed, but the trend is unambiguous and continuous. Proprietary has lost and is losing market share, open source has and is gaining. By some measures Oracle was just about equal to MySQL in 2021. Guess which is sliding down the snake, and which is climbing the ladder.
This is it. This is the canonical proof that open source can achieve everything needed for corporate software, when there's a big enough community of motivated developers. Can it in turn support that community?
Again, looking at databases in particular gives a good lens for the bigger picture. FOSS was born of idealism, frustration, opportunity and optimism. It recognized the inequity of centralized control of software, born of a time when entry costs to making software were very high and distribution very difficult, in an environment when neither situation was still the case. Like so many successful revolutions, the very act of winning changed the dynamics that made the win possible.
The ideal FOSS license is completely unrestrained: take the code, do what you like, just ask those who come after to do the same. That works in many cases, where those who do most of the work can parlay their expertise into business relationships.
However, it doesn't work so well in the age of hyperscalers, where hosted services can craft deals that require minimal interaction and risk for clients, based on FOSS running behind an API. Hence the advent of ideas like BSL, the Business Source License, that fulfils part of the FOSS ideal by making source open, but restricts commercial use. That can be any commercial use, or specific cases like selling a hosted service - something that databases are very well suited for.
Is this betrayal of FOSS? Many think so, and in a model that relies on community as a proxy for closed-door development, that could be fatal. Or is it a sensible evolution, absorbing a very well-tested model of free for non-commercial use, subsidized by production use, that's been part of proprietary for decades?
MongoDB loses its mind with marketing budget movie mania: Yep, it's choose-your-own-adventure Hackers with drop-down menusREAD MORE
The real danger isn't some dilution of FOSS ethics, but the resurgence of lock-in. While BSL and its ilk has that danger, so does any FOSS project too dependent on a powerful sponsor. The fact that the code is open is a strong safeguard: that which can be rewritten cannot be constrained. Ask IBM about its proprietary but visible PC BIOS.
This is an evolving market, but it's evolving into a more just, more sustainable and more flexible one as FOSS ideas change the landscape.
You'll have problems if you change your model in ways your initial supporters didn't expect – so be aware of how the evolution is progressing and build in your long-term options at the start. If you're as open about your plans as you are about your software, that's good enough.
The evolution of the dull old database not only predicts the future, it's helping to define it. And that's as sexy as any tech gets. ®