All your database are belong to us: Snowflake named DBMS of the year by DB-Engines
Homespun database trackers also see PostgreSQL and MongoDB as hotshots for 2021
Snowflake, that cloud-only data warehouse much loved by investors, has been named DBMS of the Year for 2021 by tracker site DB-Engines.
Once valued at $120bn, Snowflake has not only produced the first software-as-a-service to win the accolade, but it is also the first data-warehouse focused product that makes it into the top three of this award.
It is joined in the top three by three-times previous winner PostgreSQL, which came second, and third-placed MongoDB, which took the top prize in 2014 and 2015.
AWS Babelfish for PostgreSQL: A chance to slip the net of some SQL Server licensing costs?READ MORE
Rather than measuring database popularity on revenue — which would skew against FOSS systems — or basing its findings on downloads — which would include software downloaded for hobbyists, pilots, and tinkerers — DB-Engines bases its popularity score by amalgamating metrics.
Such measures include mentions on websites, Google search trends, appearance in online technical discussions, job offers, professional profiles and social media feeds.
DB-Engines standardises across these metrics, which might in themselves be flawed but, taken together, offer a fair indication of a DBMS's popularity, according to Paul Andlinger, DB-Engines co-founder.
Andlinger founded specialist Austrian database consultancy Solid IT in 2011 with co-founder Matthias Gelbmann. They launched DB-Engines in 2012 in response to client requests for basic information about databases coming onto the market and their relative popularity. Finding no such resource easily available, they decided to build it themselves.
The DBMS of the year is calculated by subtracting the previous year's popularity score from the current years, thus avoiding skewing in favour of systems just entering the market and growing from a very low base.
As DB-Engines tracks popularity every month, Andlinger was not surprised by Snowflake's success but said he was still impressed by how its popularity has increased over the last couple of years. "At the beginning of the year, and over the months, it was a surprise that took such attention and increased its numbers so heavily. Of course, the IPO at the end of 2020 contributed to this success. But it's not only that, the number of questions on Stack Overflow where people have experience with Snowflake has also increased. It's not only investors. It is developer and operators and so on."
- WTF? Value of Finnish open-source-as-a-service startup Aiven jumps $1.2bn in 7 months
- You only love me for my cache: New modules try to make NoSQL Redis more of a general-purpose database
- Google plays catch-up with JSON support for distributed RDBMS Spanner
- EDB tries to crowbar graph, JSON, and time-series data models into PostgreSQL – but can they pull it off?
Meanwhile, PostgreSQL's sustained popularity was down to the strength of the open-source community behind it, Andlinger said. "This is dramatically different to Snowflake. They have an extremely active community and the community is very dedicated to the product. We see each year with each version, that they improve the product and introduce new features. On the other hand, it's their type of license. They have a very free open-source license, which makes it easy for developers and designers to use their product in applications," he said.
MongoDB re-entered the top three databases of the year because of developers, Andlinger argued. "If you're new to database systems, and if you are not stuck in the relational system thinking, and decided to quickly select the database system, it may be a good choice to go with MongoDB especially if developers decide they don't need a database administrator," he said.
But there was no need to be drawn into a dichotomy of RDBMS Vs NoSQL. There are simply more database use cases, and different IT professionals choose different systems for different reasons, he said.
Plus, the two worlds are getting closer together: many RDBMS can handle unstructured data (in JSON document stores for example) while many NoSQL databases offer schema-like structure and support SQL or something similar.
While Snowflake has clearly made an impression on the database market in terms of its increasing profile, some things never change. Oracle has led the ranking since the service was launched more than nine years ago. ®