Blog > Post
Oracle and Cassandra increase their DB-Engines Ranking score
by Paul Andlinger, 2 January 2013
Tags:
The DB-Engines Ranking score is an average of several measurements, see ranking method. These measurements are then converted to a score by dividing them by the average of the top 5 DB engines in each category. That gives us a relative score rather than an absolute one.
Why don't we simply add up the raw measurements? Mainly for two reasons:
- The values of the measurements differ substantially in the various categories. There are many millions of results in a Google query for certain systems, whereas for example the number of job offers in Indeed is much smaller. Just adding these two figures would not make much sense. A relatively small increase in Google results would easily outweigh any significant change in the Indeed figures.
- In order to make sure that the trends are relevant for our goal to measure popularity of DB engines, we have to filter out other trends. For example, an increase of 10% in the number of LinkedIn profiles mentioning a specific system could well mean that the number of LinkedIn members has increased by that much. Only if at the same time the number of LinkedIn profiles for other systems has increased by less than 10%, then we consider this to be a significant change in the popularity ranking of the DB engines. Otherwise we would end up measuring the popularity of LinkedIn.
Using a relative score means, that if the popularity of one system increases, the score of the other systems decrease, even if their absolute popularity is the same. This is what we have seen this month.
Oracle significantly increased the popularity score, and at the same time most other systems lost points, some more and some less. The only other system in the top 10 that could keep up with Oracle was Cassandra, gaining some points at rank 10.
When looking at changes in ranks, we continue to see the fight for rank 2 between MySQL and Microsoft SQL Server, with MySQL being slightly ahead this time.
Outside the top 10 Memcached gained one rank, coming quite close to Redis, which remains the most popular key-value store. Also HBase and, a bit surprising, Ingres advanced in the ranking.
We have added several systems to our survey, which now covers well over 100 tools. You can find all of them in our complete ranking.
Share this page