The term NoSQL dates back to 1998 when it was used for a particular RDBMS that did not support SQL. It was not until 2009 that it was used with approximately the same meaning that it has today. NoSQL databases differ from RDBMSs mainly in that they use non-relational data models, lack explicit schemas and scale horizontally. The two seminal papers on Google’s Bigtable and Amazon’s Dynamo seem to have been the starting point for the NoSQL movement. Nowadays NoSQL is a hot topic in big data and enterprise architecture, as organizations look for solutions to the massive new requirement for scale, performance and analytical capability. A definition of NoSQL databases does not exist, due to a lack of organizations in this particular field. The first attempt to define the term NoSQL databases is published in Friedland et al. (2010).
Databases following some of the principles below are generally referred as NoSQL databases:
- schema free or at least weak schema restrictions non-relational data model;
-
distributed and horizontal scalability;
-
simple application programming interface;
-
different consistency approaches: eventual consistency and basically available, soft state eventually consistent (BASE) but not ACID;
-
open source;
-
support for a simple replication.
