NTU Singapore and OceanBase will initiate collaborative research on “Learning to Rank for Parametric Query Optimization” based on OceanBase 4.0 Paetica.
OceanBase and Nanyang Technological University, Singapore (NTU Singapore) signed an agreement to explore technology innovations in enhancing database system performance and advancing green computing practices. The agreement will support advanced research on the optimisation of databases, with a particular focus on improving the performance of parameterised queries.
Prof. Cong Gao, principal investigator of the project and a tenured professor at NTU Singapore’s School of Computer Science and Engineering, said, “The incorporation of machine learning technologies into database systems offers a significant opportunity to overcome the performance constraints of parameterised queries, enhancing efficiency and reducing energy consumption. Our collaboration with OceanBase is a step towards advancing database and green computing technologies into a new era of innovation.”
Yang Chuanhui, Chief Technology Officer of OceanBase, said, “We are excited to be joining forces with NTU Singapore to explore cutting-edge innovations in machine learning for databases. This venture goes beyond enhancing database technologies and is crucial for the environmental sustainability of technology systems that underpin our digital economy. We are confident that this university-industry collaborated research will hasten advancements in this area and reveal significant opportunities as database systems being more widely deployed.”
Parameterised queries, widely used in database systems, is a kind of data-intensive computing tasks that typically require substantial computing resources and energy. Consequently, database researchers are continuously exploring optimisation strategies to reduce CPU and memory consumption, shorten query times, enhance the overall performance of database systems, and foster more eco-friendly database operations. Although machine learning technologies have been applied to parameterised queries as part of this research direction, the current methods face limitations when dealing with complex queries.
To address this technical challenge, NTU Singapore and OceanBase will initiate collaborative research on “Learning to Rank for Parametric Query Optimization” based on OceanBase 4.0 Paetica. Researchers from both institutions will investigate the application of innovative technologies to resolve performance bottlenecks associated with parameterised queries, including the development of new algorithms that dynamically adapt to ever-changing database environments, leveraging advanced machine learning technologies to generate near-optimal cache plan sets for parameterised queries, and more accurately pinpointing and forecasting the effects of subtle parameter changes on query plan efficacy to identify the optimal option.
