EAGER: III: Divergent Design Index Tuning Using Machine Learning on Quantum Computers for Large-Scale Database Applications

With advances in quantum computing and quantum machine learning, which have been shown to be able to solve certain problems much faster than their classical counterparts, this project uses this to develop machine learning-based divergent design index tuning algorithms on quantum computers for large-scale database applications. The developed quantum algorithms are implemented and comprehensive performance studies are conducted using the IBM quantum simulator and quantum computer to compare the developed quantum algorithms with their classical counterparts.The project ideas are novel as there is no existing algorithm on quantum ISP and quantum divergent design indexing tuning. Additionally, no existing study analyzes the benefits and limitations of quantum computing versus classical computing on this topic. Given a database consisting of tables with columns and rows residing on a computer node, a workload of queries, and a space budget, the Index Selection Problem (ISP) involves selecting an index configuration composed of one or more columns in the database tables that would minimize query processing time. Divergent design index tuning extends the ISP for a single node to index selection for a database replicated on multiple nodes. Instead of creating one uniform index configuration for all nodes, it creates a set of index configurations, each tailored to a subset of the queries in the workload. This technique has been shown to reduce query processing time for small-scale database applications on classical computers. Divergent design index tuning is a difficult problem to solve within a reasonable amount of time for large-scale database applications in the era of Big Data, where the numbers of computer nodes, database tables, rows, columns, and queries are very large—common in many application domains such as e-commerce, finance, and healthcare. Additionally, the frequent occurrence of update queries, alongside retrieval queries, adds complexity to the problem. Developing a divergent design index tuning algorithm that can work efficiently for such large-scale applications is essential. Quantum computing offers new possibilities from anew algorithm on quantum measures is the goal of this project. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.