Graph Schema Storage in SQL Object-Relational Database and NoSQL Document-Oriented Database: A Comparative Study

Abstract

Over the past years, the main objective of database management systems was to store data and give users the possibility to manipulate those data. SQL databases were sufficient in the first time, but with the evolution of computer science technologies, other factors start to appear; such as the volume of the stored data, the velocity of the transactions between the user and data, the possibility to store heterogeneous data (variety) and many other Big Data challenges. This paper presents the major differences between the SQL and NoSQL databases in term of variety, velocity and ease of programming. For SQL, we used the Oracle object-relational database and for NoSQL, we used MongoDB document-oriented database. The comparative study was applied to data modeled by graphs, where we consider a set of graphs and for each graph, we measure the time needed to insert the graph in the database, the size of the graph in the database and other factors. Measurements presents in this process are generally automatic and supported by a set of developed algorithms.