Why the future of enterprise data is open source
Commentary: Open source databases spent years trying to get taken seriously for enterprise workloads. It worked.
Your next enterprise database is almost certainly going to be open source. No, I’m not saying proprietary databases like Oracle are dead. Nor am I suggesting that developers who have a command of SQL Server don’t get paid (and quite well). What I am suggesting is that there’s clearly a perfect storm brewing for open source databases, as revealed in Stack Overflow’s 2020 survey of over 65,000 developers.
Indeed, what’s particularly interesting in the data is just how much open source databases have already become the de facto enterprise databases.
Losing stickiness every day
Databases have tremendous staying power within the enterprise. Once data finds its way into a particular database, there’s little appetite for ripping it out and replacing it. Of course, database migrations happen. But more often than not, companies stick with legacy databases for legacy applications even as developers embrace new databases for new applications.
This isn’t news. It’s been happening for years.
SEE: How to build a successful developer career (free PDF) (TechRepublic)
As such, it’s not surprising to see legacy databases holding firm at the top of the DB-Engines database popularity rankings. Those who have watched for years have watched these leads get roughly halved, but that’s over nearly a decade. Looking at the Stack Overflow survey data, each of the big three proprietary database giants makes the list of most popular databases, though with wildly divergent levels of “popularity.”
In other words, you can expect to see the database incumbents in use for a long time.
Loving your data(base)
This isn’t the whole story, however. Nor is it the most interesting aspect of the database market. After all, if we look at which databases developers actually love to use, open source databases lead the pack, and by a massive margin. Asked which databases developers use and want to continue using, the top five results are:
Not all of these are open source (Firebase never has been, and MongoDB replaced its open source license a few years back), but the most in-demand databases are. What about those database technologies that developers aren’t using today, but hope to use in the future? It’s the same list as above, though with rearranged ordering:
For PostgreSQL, its popularity may come down to rejuvenating the relational data model with a fresh, enterprise-class approach. It also helps that it’s a true community-driven database, and not owned by a proprietary database company.
MongoDB, Elasticsearch, and Redis, by contrast, are popular because they have introduced innovation into the database market, enabling new kinds of applications that simply weren’t possible with an RDBMS. Across the spectrum of open source (or quasi-open source) databases, the fact that they’re already in heavy use within the enterprise suggests that this trend toward open source databases isn’t slowing down anytime soon. And given that even the erstwhile open source databases now rely on cloud, not licensing, to generate revenue, I’d imagine we’ll see an accelerating trend toward more open source among the database elite.
For years open source databases worked hard to be taken seriously as enterprise-class alternatives to the proprietary incumbents. Based on these survey results, it seems to have worked. The future of enterprise data is open source.
Disclosure: I work for AWS, but the views here are my own.