ABSTRACT
Companies have migrated data from relational databases to NoSQL
databases to improve their business through more active services at
a lower operating cost, especially by the adoption of cloud services.
This process is called Data Migration and is considered by some
authors one of the biggest challenges in systems engineering today.
Although it is advantageous, the process of migrating data from
the relational model to NoSQL models is not trivial and has led to
the development of different methodologies for this purpose. The
objective of this work was to analyze and compare three different
migration methodologies between Relational and NoSQL Document
Oriented databases under the following aspects: algorithm
input, method documentation, migration process and generated
documents. For that, different relational models were empirically
migrated using such methodologies, allowing the analysis of the
evaluated aspects. The results show that there is no consolidated
way to perform the migration and that the method to be chosen
depends on the context of the application. So, scenarios that show
when to use each method are presented. Although not performing
computational tests, this work provides suggestions and insights
through the evaluation of the migration processes under the theoretical
models. It expected that the results presented here will help
IT managers decide on the best data model migration methodology
to follow in their actual projects.