Blog > Post
NoSQL systems continue to be on the rise
by Matthias Gelbmann, 3 December 2013
These opposite extremes are caused by the fact that in our ranking method all the scores are relative to each other. This means that a drop of a top ranked system automatically causes an increase of the average of the other scores.
Why do we do that? The reason is, that we cannot measure absolute popularity. When we count for example the number of Linkedin profiles mentioning a certain DBMS, and we see that this number has doubled within a year, then this increase is to a certain extend caused by the increase in the usage of LinkedIn and only partly due to a change in the popularity of that DBMS. But when we see that another system has gained only say 70% of LinkedIn profiles, then we can make conclusions about the relative popularity of these two systems. If we would use the absolute numbers, we would end up measuring the popularity of LinkedIn, or the size of the Google index or the overall shape of the job market.
The systems that reached their all-time high this month include the open source RDBMS MySQL and PostgreSQL, and also the leading NoSQL systems MongoDB, Cassandra and Redis are stronger than ever before.
The trend towards NoSQL systems is most clearly visible in our ranking category report, where the RDBMS category still dominates, but compared to the beginning of the year the dominance is a bit weaker. At the same time, NoSQL systems increased their popularity by 50% or more.
While this trend is not new, it is somewhat amplified this month by a change in one of our ranking sources, namely the Bing search engine. The Bing API now returns the same number of hits than an online search on the Bing website, while previously it returned ony a relatively small subset. The number of search results is now two orders of mangnitude higher than in the months before. As explained above, the way we calculate the scores is immune against such changes, but it looks like the filtering of the Bing API was not perfectly uniform in the past. This caused a bit of distortion, which seems to be eliminated now. We carefully monitor all changes in our ranking sources to make sure changes in one of our sources don't cause a disruption of our ranking.