MongoDB performance analysis: A comparative study between stand-alone and sharded cluster deployments with open data from Brazilian Bolsa familia program

摘要

In the "Big data Era", new trends like "Internet of Things", Social Networks and Open Data have contributed to a real explosion in data availability. To face these phenomena, the so-called NoSQL Databases arise, MongoDB being an important representative of this category of solutions. The present work aims to evaluate performance aspects of MongoDB when applied to the analysis of large volumes of open data related to Bolsa Família, and to evaluate the gains derived from the adoption of the cluster architecture versus the standalone mode. The results showed significant performance gains in the execution of aggregate queries with the cluster adoption and that the performance loss during the load can be minimized with workarounds.