How to objectively measure the popularity of a DB engine? Good question! And there’s an Austrian company (Solid IT) who claims to have the answer. The company focuses on “Big Data und NoSQL”, but this focus does not seem to have biased the result of the measurement. Among the top 10 database engines, there is only MongoDB, which is not an RDBMS. And it’s astonishing just how popular MongoDB seems to be (although, they must be doing something right)!
Now, I’m not surprised by the top 3. I am definitely surprised by the fact that PostgreSQL and SQLite are not more popular. I am also surprised, that there aren’t more “wide-column stores” among the top 10. Maybe, Michael Stonebraker has to review his claims about the traditional RDBMS wisdom being all wrong?
And what about the other databases supported by jOOQ? Where are the Java databases? Here’s a condensed view of the ranking, consisting only of the 15 databases currently supported by jOOQ 3.1:
It turns out that Java databases (Derby, H2, HyperSQL) are not so popular compared to all the others. It also turns out that MariaDB still has a lot of grounds to gain, compared to MySQL.
The ranking considers a lot of data from various somewhat authoritative sources as is explained here. These include:
- Number of mentions of the system on websites. Measured through search engine results.
- General interest in the system. Measured through Google Trends.
- Frequency of technical discussions about the system. Measured through Stack Overflow and similar.
- Number of job offers, in which the system is mentioned. Measured through Indeed and similar.
- Number of profiles in professional networks, in which the system is mentioned. Measured through LinkedIn.
This ranking is certainly something to keep an eye on!
There’s no way H2 is not more popular in reality. I have yet to see a product *not* embedding H2 by default. All application servers, all tools, anything that uses a DB (at least in the Java world, which just happens to be gigantic) ships with H2.
H2 just happens to have a very bad name for any such analytics… As opposed to MySQL, which is a very unique term
Another great example of NoSQL databases is CryptonorDB (cloud – mobile database). It delivers high availability, fault tolerant database service accessible via a RESTful HTTP/JSON API. Moreover, CryptonorDB safeguards the data by encrypting it: it manages the storage of encrypted data, but only the client manages the key.