This topic shows the results of online transaction processing (OLTP) performance tests on ApsaraDB PolarDB for MySQL 5.6.

The PolarDB MySQL 8.0 optimizer supports the followingmodes:

  • Cost-Based Optimization (CBO): chooses the mostefficient way to optimize queries by creating a cost-based model base onstatistical information or samples collected in real time.
  • Rule-Based Optimization (RBO): optimizes queries based on the preset rules.

The PolarDB for MySQL 8.0 optimizer uses CBO as the primary mode, and also supports RBO under some circumstances. You can run the ANALYZE command to collect the required statistical information, or allow the optimizer to automatically collect the information. The required information includes the table size, number of records, density vector, density histogram,Number of Distinct Value (NDV), and null ratio.

Note For more information about how to test the OLTP performance of ApsaraDB PolarDB for MySQL, see OLTP performance tests.

Test results

Figure 1. Read-only performance of different specifications
Read-only performance of different specifications
Figure 2. Write-only performance of different specifications
Write-only performance of different specifications
Figure 3. Read-write performance of different specifications
Read-write performance of different specifications
Figure 4. Use cluster endpoints that are addressed to different numbers of read-only nodes
Use cluster endpoints that are addressed to different numbers of read-only nodes