DB-Engines Ranking Identifies Popular NoSQL Systems

DB-Engines ranking

The DB-Engines ranking shows the popularity of database management systems. Popularity is measured by 5 factors: number of results shown on Google and Bing when searching for a system, number of search queries on Google Trends, number of job offers on Indeed, number of LinkedIn profiles mentioning the system, and number of related questions and followers on Stackoverflow.

The ranking lists over 100 systems, both traditional relational DMBS and NoSQL systems. The overall ranking is dominated by the established RDBMS: Oracle, MySQL, SQL Server, MS Access, PostgreSQL and DB2. These are still the big players in the database scene, however, the majority of systems in the ranking are now NoSQL systems and they are getting more popular every month.

The ranking of the NoSQL systems is split up by database models.

Document Stores

MongoDB is the clear leader in the document store ranking. It is also the most popular NoSQL system, at rank 7 overall. CouchDB, one of several systems maintained by the Apache Foundation, comes second in that list. After another big gap we have the next two systems: the .NET-based RavenDB and Couchbase, a promising evolution of CouchDB, which both increased their score in the last month. None of the other document stores has gained enough public attention to be assigned a score for the ranking.

Key-value stores

Key-value stores are the most crowded NoSQL category. The complexity of these systems varies a lot, from the simplest version of Berkeley DB to the feature-rich, distributed Riak and the powerful Amazon-hosted DymamoDB, it’s quite a menagerie of systems.

The most popular key-value store is Redis, a fast little in-memory system supported by VMWare with some unique features, ranking 11th in the overall list. Just behind it is Memcached, which is widely used as a caching system. The next systems, with a distinctly lower score, are Riak, Oracle’s Berkeley DB, the two Amazon systems SimpleDB and DynamoDB, and Oracle NoSQL, which is based on a version of Berkeley DB and which almost doubled its score in the last month.

Wide Column Stores

Wide column stores are generally considered the third important NoSQL category. Cassandra, which has been handed over from Facebook to Apache, is the leading system here. Standing at rank 10 in the overall ranking makes it the second-most popular NoSQL system. The Hadoop-based HBase, another Apache system, follows closely and with Hypertable there is only one more system listed in that group. Google’s BigTable, the system that originally defined that category, is not listed, because it is not publicly available.

Graph DBMS

The Java-based Neo4j is by far the most popular Graph DBMS. Its main competitors DEX, OrientDB, InfoGrid and InfiniteGraph follow at a distance.

RDF Stores

RDF stores, a specialized variant of Graph DBMS’s, are led by Apache Jena, just ahead of Sesame and some more systems with a clearly lower score.

Other categories

The other categories defined at DB-Engines are generally not seen as NoSQL systems, partly because they have been around for a long time, and also because they don’t necessarily address the Big Data problems that are often associated with NoSQL. These categories are native XML DBMS, object oriented DBMS, content stores and multivalue DBMS.

Future

The DB-Engines ranking is updated monthly. While the number of NoSQL systems might shrink in the coming years due to some consolidation process, we can expect their popularity to rise further.