| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.41 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A necessidade crescente de se armazenarem grandes quantidades de dados, de forma a prover
serviços escalåveis, obrigou à procura de novas soluçÔes, ao longo do tempo. No decurso da
histĂłria podem enumerar-se diversos sistemas de dados: o hierĂĄrquico, o em rede, o relacional,
o orientado a objetos, o objeto-relacional e, mais recentemente, o NoSQL. Todos estes sistemas
tentaram dar resposta a diferentes realidades do armazenamento de dados, indo ao encontro dos
problemas de cada Ă©poca.
Devido Ă necessidade de se aproveitarem as vantagens que cada tipo de base de dados oferece,
surgiu o conceito de PersistĂȘncia Poliglota, que traduz a integração de vĂĄrios tipos de bases de
dados num sĂł sistema. Esta abordagem tem como objetivo retirar o melhor de cada base de
dados, apresentando uma solução fiåvel e alternativa aos sistemas com apenas um tipo de base
de dados.
Como tal, este trabalho visa a anĂĄlise da abordagem de PersistĂȘncia Poliglota para comparar
sistemas compostos por diversos sistemas de gestĂŁo de base de dados versus os que utilizam
apenas um motor de base de dados, de modo a verificar se esta abordagem Ă© Ăștil e vantajosa.
Desta forma, elaborou-se uma prova de conceito, com base num problema proposto, com o
objetivo de se analisarem dois sistemas, um Ășnico e outro poliglota, tendo por base trĂȘs volumes
de dados diferentes. Para isto, foi necessĂĄrio proceder-se Ă anĂĄlise e escolha dos sistemas de
gestĂŁo de base de dados a utilizar e montar os ambientes de teste, para ambos os sistemas. Com
recurso a vĂĄrias consultas individuais (a cada base de dados) e globais (conjunto das bases de
dados que compÔem o sistema poliglota), foram analisados os resultados obtidos com recurso
à métrica de medição do desempenho relativa aos tempos de consulta.
O trabalho e os resultados obtidos evidenciaram um aumento do desempenho, quanto Ă
utilização individual das bases de dados. Perante o conjunto das bases de dados, apesar de um
ligeiro aumento, nota-se que os resultados não são claros e que carecem de uma investigação
mais profunda. Por fim, Ă© possĂvel afirmar que a abordagem poliglota Ă© principalmente Ăștil em
sistemas complexos, onde o volume de dados Ă© elevado, e onde se pretende armazenar
diferentes tipos de dados.
ABSTRACT: The increasing need to store large amounts of data in order to provide scalable services has forced the search for new solutions over time. In the course of history, several data systems can be enumerated: hierarchical, network, relational, object-oriented, object-relational, and, more recently, NoSQL. All these systems tried to respond to different realities of data storage, meeting the problems of each era. However, due to the need to take benefit from all the advantages that each type of database offers, the concept of Polyglot Persistence has emerged, which allows the integration of several types of databases in a single system. This approach aims to get the best out of each database, presenting a reliable and alternative solution to systems with only one type of database. As such, this work aims at the analysis of Polyglot Persistence approach to compare systems composed of several database management systems versus those using a single database engine, in order to verify if this approach is useful and advantageous. In this way, a proof of concept was elaborated, based on a proposed problem, with the objective of analyzing two systems, a single and another polyglot, based on three different data volumes. For this, it was necessary to proceed to the analysis and choice of the database management systems to be used and to assemble the test environments, for both systems. Using a number of individual queries (for each database) and global queries (set of databases that make up the polyglot system), the results obtained were analyzed using the performance metric relative to the query times. The work and the results obtained showed an increase in the performance, regarding the individual use of the databases. In spite of a slight increase, the results are not clear and need further investigation. Finally, it is possible to affirm that the polyglot approach is mainly useful in complex systems, where the volume of data is high, and it is intended to store different types of data.
ABSTRACT: The increasing need to store large amounts of data in order to provide scalable services has forced the search for new solutions over time. In the course of history, several data systems can be enumerated: hierarchical, network, relational, object-oriented, object-relational, and, more recently, NoSQL. All these systems tried to respond to different realities of data storage, meeting the problems of each era. However, due to the need to take benefit from all the advantages that each type of database offers, the concept of Polyglot Persistence has emerged, which allows the integration of several types of databases in a single system. This approach aims to get the best out of each database, presenting a reliable and alternative solution to systems with only one type of database. As such, this work aims at the analysis of Polyglot Persistence approach to compare systems composed of several database management systems versus those using a single database engine, in order to verify if this approach is useful and advantageous. In this way, a proof of concept was elaborated, based on a proposed problem, with the objective of analyzing two systems, a single and another polyglot, based on three different data volumes. For this, it was necessary to proceed to the analysis and choice of the database management systems to be used and to assemble the test environments, for both systems. Using a number of individual queries (for each database) and global queries (set of databases that make up the polyglot system), the results obtained were analyzed using the performance metric relative to the query times. The work and the results obtained showed an increase in the performance, regarding the individual use of the databases. In spite of a slight increase, the results are not clear and need further investigation. Finally, it is possible to affirm that the polyglot approach is mainly useful in complex systems, where the volume of data is high, and it is intended to store different types of data.
Description
Keywords
PersistĂȘncia Poliglota Modelo de Dados Bases de Dados Sistemas de GestĂŁo de Bases de Dados NoSQL Modelo Relacional