A comparative evaluation of NoSQL database
systems

Abstract. With the quick emergence of the largest web sites in the mid-2000s, and the adoption of the cloud-based computing model, traditional relational database systems could not keep up with the requirements of high throughput and distributed operation. As a result, major web companies developed their own, inherently distributed, lightweight solution to act as a database back-end for their services. These developments spun interest in the open source world and numerous products appeared under the term NoSQL — not only SQL. In the present work we compare a few NoSQL systems (MongoDB, Cassandra, Riak) according to a wide set of aspects. We conclude, that while NoSQL systems offer much less functionality than traditional relation database management systems, especially in transaction isolation and scan operations, they can be successfully used when complex database logic is not, but large-scale, distributed operation is an objective.