Análise de desempenho sobre camadas de persistência em banco de dados relacional e não-relacional

Abstract

Non-relational databases emerged as a solution to problems that need great scalability and availability to store data. This data storage paradigm is capable to provide great performance in data storage large scale. For this reason, the use of this paradigm has grown a lot in corporate applications. Nowadays there is a demand for systems migrations feasibility testing applications from relational to non-relational databases to verify the performance of the two distinguished data storage paradigm, verifying the advantage of this migration, besides the feasibility of implementation time and complexity for this migration. The motivation to execute this project is to help the developers community to verify the efficiency ans easiness of the migration between databases with different data store paradigms without the need to rewrite the source code, aiming to save time and costs. The objective of this project is to test the persistence layer for the relational and non-relational databases to identify the viability of migration between database in legacy systems. The methodology used in this project was the spiral, in which repeats the search, implementation, tests and analysis steps until the results are as specified. An application was initially developed to run on a relational database, such as most existing corporate systems, using the persistence layer to perform operations in the initial database. After this step, was performed a connection with a non-relational database without modifying the software code. The only change was made in the persistence layer, so it can perform operations in the new database. When verifying that the persistence layer is able to handle the two storage paradigm, tests were performed to verify if the persistence layer is able to maintain a satisfactory performance with the new database. An analisys was made with the data generated in the tests to verify the viability to migrate databases with diferents paradigms only changing the persistence layer with a focus on performance.