The April 2022 DB-Engines database popularity rankings are out, and the big news is … that there really isn’t much big news. Today’s top 10 databases are largely the same as they were last year, or even five years ago. Not that things never change in database land. In fact, Matthias Gelbmann, co-founder of Solid IT, the company behind DB-Engines’ multifaceted ranking system, put together a video (best at 2x speed) that shows how the rankings have changed over time.
Yet it’s still true that change in databases happens slowly, similar to what we see in programming languages, as Redmonk’s Steve O’Grady regularly reminds us. Why? Because both involve significant enterprise investments, which change infrequently, if at all.
Databases over time
Though Gelbmann’s video is worth watching, it’s perhaps even easier to portray the relative stasis in database rankings through a few screenshots. Here’s how the top 10 databases stand in April 2022:
Now let’s rewind a year to April 2021, using the Wayback Machine:
Same as it ever was, right? Well, what about April 2017, five years ago?
Okay, now we’re seeing some changes, but still not many. Apache Cassandra fell off the top 10 since April 2017, but it didn’t fall far: In April 2022 it’s ranked #11. Not much of a change. And Elasticsearch, which now sits in the top 10, was just outside (at #11) back in 2017.
What about if we go back a full 10 years to October 2012 (the first time DB-Engines started publishing its rankings)?
Memcached (now #30 in 2022) made the top 10, but otherwise there are very few absolute changes in database popularity, though of course there are significant changes in relative popularity. That is, Oracle still tops the list but its relative position is less secure.
Of course, if you want to see significant movement in rankings, positions 11 to 25 are in constant motion, though even there, perhaps not as much as you might think. It’s really in the long tail of databases (50 – 300+) where there’s a seething cauldron of database popularity changes. Google BigQuery, for example, was launched in 2011 but didn’t even meet the DB-Engines thresholds for ranking until late 2014. In April 2015 BigQuery couldn’t muster a top 50 finish, but now sits at #24. Given how slowly data moves, that kind of ranking improvement suggests dramatic adoption over seven years, measured in terms of jobs, search interest and more.
This shouldn’t surprise us. Companies are loath to change their databases, given how sensitive data can be. Even Amazon, which had profound incentives to move off Oracle, took years before it finally accomplished that feat. Nor are databases alone in getting the sticky-fingers treatment.
SEE: Hiring kit: Android developer (TechRepublic Premium)
Programming language perpetuity
I’m a fan of how Redmonk ranks programming languages, using GitHub and Stack Overflow data. The Redmonk team has been analyzing the evolution of programming language adoption for over a decade. Sometimes there are exciting trends that O’Grady and team uncover. Not recently, though. As O’Grady recently wrote:
The story of this quarter’s run — as it has been for a few runs now — is stability. Outside of a few notable exceptions that we’ll discuss momentarily, the rule of language movement in recent years has been that there is little movement. Seventeen of the twenty languages here, in fact, have been stable for three consecutive quarters.
SEE: Cheat sheet: How to become a database administrator (free PDF) (TechRepublic)
The reason for this seeming stasis is different but also similar to the stability in database choices: There’s significant friction in moving to something new. In languages, it takes a vendor with massive market power (like Apple) to convince developers to move to something like Swift or Objective-C. Or it takes dramatic improvement in safety or other concerns, as with Rust, to motivate change.
But most of the time, for most enterprises, “boring” is a feature, not a bug. Which is why for databases, programming languages, ERP systems and more, rapid technology innovation doesn’t necessarily equate to rapid enterprise adoption. After all, companies have businesses, not science fairs, to run.
Disclosure: I work for MongoDB but the views expressed herein are mine alone