Background

With the increasing amount of data to be processed and diversified application scenarios, databases are facing more and more scenarios with large capacity and diversified data. In the past 20 years, the data volume has gradually increased from MB- and GB-level to TB-level. Facing such a large amount of data, the database management system (DBMS) has higher requirements on data query and management. Objectively, the database must support multiple optimization search policies and O&M methods.

In classic algorithms of computer science, people usually use the Divide and Conquer method to solve problems in large-scale scenarios. The basic idea is to divide a complex problem into two or more same or similar problems. These problems are divided into smaller problems until they can be solved directly. The solution to the original problem can be regarded as the combination of the solutions to all small problems. In a large-capacity data scenario, the database provides a Divide and Conquer method, that is, partitioning. The logical database or its components are divided into different independent partitions. Each partition maintains data with similar attributes logically. In this way, the large amount of data is divided, facilitating data management, search, and maintenance.