This topic describes the relationship between the number of Data Transmission Service units (DUs) allocated to a task that runs in a Data Transmission Service (DTS) dedicated cluster and the performance of the task.
Usage notes
- The performance specifications provided in this topic are only for reference and are not used as a criterion for service level agreement (SLA) evaluation.
- The latency of a task in a DTS dedicated cluster may be higher due to multiple factors, such as the workload of the source instance, the bandwidth of the transmission network, the network latency, and the write performance of the destination instance.
Terms
- Cluster: a DTS dedicated cluster that consists of 2 to 100 nodes of the same specifications.
- DU: the resource unit of a DTS dedicated cluster. A DU indicates the minimum hardware resources that are required to run a DTS data migration, data synchronization, or change tracking task. The hardware resources include the CPU and memory.
- Rows Per Second (RPS): the number of table rows processed per second. For example, 5000 RPS for incremental data migration indicates that 5,000 rows of incremental data can be migrated from the source database to the destination database per second.
Performance specifications
The following table describes the relationship between the maximum performance of a task that runs in a DTS dedicated cluster and the number of DUs allocated to the task. The actual performance may vary based on factors such as the CPU utilization, memory usage, and disk usage.
| Number of DUs | Maximum performance (RPS) |
| 1 | 200 |
| 2 | 2,000 |
| 3 | 5,000 |
| 4 | 10,000 |
| 5 | 20,000 |
| 6 | 30,000 |
| 7 | 40,000 |
| 8 | 50,000 |
| 9 | 60,000 |
| 10 | 70,000 |
| 11 | 80,000 |
| 12 | 90,000 |
| 13 | 100,000 |
| 14 | 110,000 |
| 15 | 120,000 |
| >15 | No upper limit Note The number of DUs can exceed 15 only after you configure the overcommit ratio. For more information about how to configure the overcommit ratio for a DTS dedicated cluster, see Set an overcommit ratio for a DTS dedicated cluster. |
Note The actual performance may vary due to multiple factors such as the network environment, the performance of the source and destination instances, and the latency. The maximum performance values provided in the preceding table are only for reference and are not used as a criterion for SLA evaluation.