Qiushi Bai: Improving SQL Performance Using Middleware-Based Query Rewriting

Abstract:

Query performance is critical in database-supported applications where users need answers quickly to make timely decisions. Traditional databases rely on rewriting queries to improve SQL performance. With the emergence of business intelligence and interactive visualization applications, databases often miss opportunities to rewrite their queries, due to reasons such as failure to adopt high-accuracy time estimators to choose efficient plans, and missing domain-specific rewriting rules valid only for specific datasets. We focus on middleware-based query-rewriting solutions to address the problem since, in many cases, both the application and database layer are black boxes. First, we develop Maliva, a machine-learning-based technique that leverages high-accuracy query-time estimators to rewrite queries under time constraints. Second, we present QueryBooster, a human-centered query rewriting framework that provides an easy-to-use language for users to formulate rewriting rules based on their domain knowledge. Finally, to make it easy for users to optimize their application queries, we propose a middleware-based system called Squidster that provides query rewriting as a service. Our experiments in real and synthetic datasets show the effectiveness of the proposed solutions in improving end-to-end SQL query performance.

In this talk, we will focus on the QueryBooster and Squidster work. We will show a demo to motivate the problem and demonstrate the effectiveness and convenience of using the QueryBooster system to improve SQL query performance.

Bio:

Qiushi Bai is a final year Ph.D. candidate in the Computer Science Department at UC Irvine. He received his Master’s and Bachelor’s degrees in CS from Northeastern University in China. His research interests have focused on improving query performance for big data analytics.