An evaluation of data models, replication and cache in MongoDB: a case study with Enem in Brazil

In Brazil, grades from the National Secondary Education Examination - Enem are used for access to federal public universities. In addition to the grades, the microdata of this exam has a large amount of information about the students enrolled in the tests that can be used to carry out various analyzes of student performance. Due to the tens of gigabytes of data available, the data must be stored properly, to reduce the processing time for information retrieval. This work proposes a study to analyze whether different data models for storing Enem information in a document-based database can impact the response time of queries to this database. For this, data were stored using different strategies for data modeling, different replica sizes and cache usage. The experiments were performed on the MongoDB database and their response times were compared. The results showed relevant differences in response times.