NewSQL: características e análise da nova categoria de armazenamento de dados

Abstract

In the current context of computing, there is a growing number of applications that require high data throughput. However, this usually implies sacrificing features such as consistency in data handling. The NewSQL category emerges as an alternative that seeks to combine the predicates of SQL and NoSQL databases, providing efficiency, scalability, and consistency in data handling. However, this category lacks a definition of its characteristics, which makes it difficult to identify whether a particular solution is really classified as NewSQL. This work proposes to define the essential characteristics for a technology to be part of the NewSQL category. It also aims to choose and evaluate a storage solution in terms of these characteristics. The evaluation was conducted by applying a benchmark tool in conjunction with Python programming language codes developed to observe the specified characteristics. The most evident were support for complex queries, where the Database Management System was able to respond to multiple joins. In addition, the solution stood out for ensuring strong consistency, preserving data integrity even in a concurrent access scenario. Experiments have shown that the analyzed storage solution, Google Cloud Spanner, confirms its classification as a high-performance and reliable NewSQL system.