Google llc (20240311356). Workload-Driven Index Selections simplified abstract
Contents
Workload-Driven Index Selections
Organization Name
Inventor(s)
Haoyu Huang of Fremont CA (US)
Vincent Zhuang of Mountain View CA (US)
Sam Idicula of Santa Clara CA (US)
Gaurav Jain of Mountain View CA (US)
Workload-Driven Index Selections - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240311356 titled 'Workload-Driven Index Selections
The method described in the abstract is for workload-driven index selection in a database system. It involves analyzing a set of queries executed in the database to recommend an optimal index configuration based on workload cost.
- The method starts by collecting a variety of queries executed in the database.
- It then identifies a set of candidate indexes from the available indexes.
- For each candidate index, the method calculates a workload cost based on the queries executed.
- A first candidate index is selected as part of the recommended index configuration based on the workload cost.
- Additional candidate indexes are also chosen for the recommended index configuration if their combined size meets a specified threshold.
- The recommended index configuration is then transmitted for implementation in the database.
Potential Applications: - Database optimization for improved query performance. - Automated index selection for database administrators. - Enhancing overall database efficiency and speed.
Problems Solved: - Simplifying the process of index selection in a database. - Improving query performance by selecting the most suitable indexes. - Reducing manual effort in optimizing database configurations.
Benefits: - Increased database performance and speed. - Reduction in manual workload for database administrators. - Enhanced overall efficiency of database operations.
Commercial Applications: "Workload-Driven Index Selection Method for Database Optimization" could be utilized by database management companies to offer enhanced performance optimization services to their clients. It could also be integrated into database management software tools to automate the index selection process for users.
Questions about Workload-Driven Index Selection Method: 1. How does this method differ from traditional index selection approaches? 2. What are the key factors considered in determining the workload cost for candidate indexes?
Original Abstract Submitted
a method for workload-driven index selections includes receiving a request for a recommended index configuration. the method includes obtaining a plurality of queries executed at the database. the method also includes selecting a set of candidate indexes from the plurality of indexes. the method includes for each respective candidate index of the set of candidate indexes, determining, based on the plurality of queries, a respective workload cost for the respective candidate index. the method also includes selecting, based on the respective workload cost, a first candidate index from the set of candidate indexes for the recommended index configuration. the method includes selecting one or more additional candidate indexes from the set of candidate indexes for the recommended index configuration. the method includes determining that a size of the selected candidate indexes satisfies a size threshold and transmitting the recommended index configuration.
