Data modeling for analytical queries on document-oriented DBMS
R. A. S. N. Soransso, Maria Cláudia Cavalcanti · 2018
NoSQL database management systems have emerged as an alternative to increase performance and decrease the hardware costs of applications that use traditional relational databases. However, there are not many works on how to guide data modeling for such DBMS to gain query performance. Specially, in the context of Business Intelligence (BI) applications, data modeling should take into account analytical queries performance. This work highlights the importance of data modeling for this kind of application on NoSQL DBMS. It shows how much alternative modelings can significantly impact on the query performance. Experiments were performed on MongoDB, a popular document-oriented NoSQL DBMS, and show some significant results. In addition, a modeling heuristic is presented for this DBMS, and suggests that more than one document collection, based on alternative data modelings, should be maintained in order to improve query performance. Future work points to query redirection mechanisms for such systems.