Análise de sistemas de gerenciamento de banco de dados para armazenamento de dados climáticos

• Ano de defesa: 2016
• Autor(a) principal: Kuhnen, Igor Antonio
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Mato Grosso; Brasil; Instituto de Física (IF); UFMT CUC - Cuiabá; Programa de Pós-Graduação em Física Ambiental
• Palavras-chave em Português: Dados micrometeorológicos; NoSQL; Benchmark; CNPQ::CIENCIAS EXATAS E DA TERRA::FISICA; Meteorological data
• Link de acesso: http://ri.ufmt.br/handle/1/5697

The monitoring and study of weather data can help public and private institutions to draw up plans for maximizing your production or even preserve life and public property in areas such as civil security, agricultural, energy. Brazil, especially the state of Mato Grosso, in this context, has greater responsibility by having in its territory three different biomes: Amazonia, Cerrado and Pantanal. A major problem for the development of research in climatology area is the availability and access to data, and also the reliability of the data, since tampering of the data remains for further research. Data Base Management Systems allow access to historical data from different sources for different users, integrating the data and providing security. The objective of this study is to identify efficient ways of storing and time series data recovery obtained from meteorological towers. The methodology used was based on the preparation of the computacional test environment and execution of tests on two different database: PostgreSQL and Cassandra using the Benchmarck YCSB with different workloads configurations commonly required for management of weather data. The results showed that the DB Cassandra has greater efficiency in the processing operations of insertion and selection of data with equivalent structure to the time series obtained from meteorological towers, when compared to the PostgreSQL relational database, with lower latency and increased data flow for all volume of operations tested . However, when dealing with operations of reading the DB PostgreSQL proved to be more efficient both for the latency and the data stream.