Uma avaliação experimental entre MongoDB e Mongochain utilizando o benchmark YCSB+T

• Ano de defesa: 2024
• Autor(a) principal: Morceli, Caio Lazarini
• Orientador(a): Ciferri, Ricardo Rodrigues
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• Palavras-chave em Português: Consistência; MongoDB; Mongochain; Análise de desempenho
• Palavras-chave em Inglês: Consistency; Performance analysis
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/19941

The Atomicity, Consistency, Isolation and Durability (ACID) properties found in a Database Management System (DBMS) are normally related to relational Databases (DB). On the other hand, the emergence of the Not Just SQL (NoSQL) paradigm arrived to meet the demands that relational DBs left something to be desired, such as scalability, availability and ease of handling a larger volume of data. Consequently, the NoSQL ecosystem has thrived by taking advantage of the Basically Available, Light State and Eventual Consistency (BASE) concepts that prioritize scalability and availability. However, the use of these concepts also brought disadvantages, the most considerable of which was that it crucified the consistency of the ACID model to prioritize availability and performance. This contrast generated new research niches on this issue. Therefore, Mongochain was created in an attempt to couple the ACID and BASE models to enjoy their benefits. This software used an innovative technology, blockchain, which has the necessary characteristics to guarantee ACID and BASE, as it can execute reliable and transparent transactions in a distributed and decentralized network. In addition to Mongochain, MongoDB has also brought improvements over time. One of them was support for transactions that preserve ACID properties in multiple documents. In addition, MongoDB provides several consistency levels for working with data persistence, but there is no specific study that analyzes the performance of these levels. Therefore, this work seeks to carry out a study on MongoDB’s consistency levels, checking which of them guarantee the consistency property included in the ACID model and evaluating the correctness of Mongochain. Therefore, it was necessary to use a benchmark that would allow performance and consistency tests to be carried out, which are found in YCSB+T. Finally, the results show that the stronger consistency levels of MongoDB and Mongochain were able to guarantee data consistency.