Machine Learning for Optimizing Database Performance

Due to the developing requirement of data storage, today’s database cannot meet the highly developing world. Besides, the wind of Machine Learning (ML) showed a possible solution to the dilemma. Therefore, many researchers have proposed strategies to improve the performance of databases in terms of security and efficiency. This paper discusses the methods of how to use ML to improve the performance of databases. In the methodology part, an introduction is proposed to interpret the workflow of ML first. Then, security problems and relevant solutions are presented in this paper, including improved performance to deploy the database locally to avoid attacks during data transportation, ignore dangerous configurations, and detect traffic data by building a new system. Besides, efficiency issues are also resolved by ML. A strategy is introduced to filter dimensional queries, and sorting functions, to accelerate scheduling time which rapidly constructs the database, a framework and a method to find the relationships under queries also contributed. Lastly, the paper finds the underlying issues that are still under challenges. For instance, reduce manpower on building or managing a database, unified system, and real-world limitations.