A serverless instance provided by Data Transmission Service (DTS) provides scalable resources. DTS automatically adjusts the resource specifications of serverless instances to meet changing business requirements. This avoids resource waste and reduces O&M costs.
Background information
Serverless billing is a dynamic billing method that allows you to dynamically adjust the resource specifications of an instance at a one-minute interval based on the workload of the instance. You are billed for the instance per hour based on your resource usage. This significantly lowers your costs. Instances that are purchased based on the serverless billing method are called serverless instances.
DTS can dynamically adjust the resource specifications of a serverless instance based on factors such as records per second (RPS), CPU utilization, memory usage, and network traffic. The resource specifications are adjusted by changing the number of DTS units (DUs). In 60 seconds after the number of DUs is changed, DTS checks whether the current resource specifications can meet the performance requirements.
The following figure shows how the resource usage and specifications of a regular instance and a serverless instance change when the amount of data transmitted significantly fluctuates.
The preceding figure provides the following information:
The regular instance wastes a large number of resources during off-peak hours. As a result, resources are insufficient during peak hours, which affects the performance of regular instances.
The resource specifications of the serverless instance are dynamically adjusted based on workloads. The serverless instance can fully meet your business requirements during both off-peak and peak hours to guarantee instance performance.
Benefits
High scalability
The resource specifications of serverless instances are adjusted at a one-minute interval based on the amount of data transmitted. This improves resource utilization and reduces resource usage.
Zero O&M
You do not need to manually adjust the resource specifications of serverless instances. This improves O&M efficiency and reduces O&M costs for administrators and developers.
Cost-effectiveness
The serverless billing method is a pay-as-you-go method that allows you to strike a dynamic balance between workloads and resource usage at lower costs.
Scenarios
The amount of data to be transmitted is large in the early stage and levels off in later stages.
The amount of data to be transmitted fluctuates widely between peak and off-peak hours and the operating costs must be reduced.
Customers have existing DTS tasks and are willing to purchase serverless instances or change regular instances to serverless instances.
The amount of data to be transmitted is unpredictable.
Individual developers.
Limits
You can create a serverless instance only for a one-way data synchronization task.
You cannot create a cross-region or cross-border serverless instance.
A serverless instance supports only the pay-as-you-go billing method.
Valid values for the minimum number of DUs are 1, 2, 4, 8, 16, and 32. Valid values for the maximum number of DUs are 2, 4, 8, 16, and 32.
The following types of databases cannot be used as the source database in a serverless instance: Tair/Redis, PolarDB-X 1.0, PolarDB-X 2.0, and MongoDB Sharded Cluster.
The destination database in a serverless instance cannot be an Tair/Redis instance.
Billing
For more information, see Billing of DTS serverless instances.
Supported regions
DTS serverless instances are supported in the following regions: China (Hangzhou), China (Shanghai), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Ulanqab), China (Shenzhen), China (Guangzhou), China (Wuhan - Local Region), China (Chengdu), China (Hong Kong), US (Silicon Valley), US (Virginia), Singapore, Malaysia (Kuala Lumpur), Indonesia (Jakarta), Japan (Tokyo), and Germany (Frankfurt).
Performance
The following table describes the mapping between the number of DUs and maximum RPS of serverless instances. The actual RPS may vary based on factors such as the network environment, performance of the source and destination databases, and latency.
DUs | Maximum RPS |
1 | 200 |
2 | 1,000 |
4 | 3,000 |
8 | 9,000 |
16 | 17,000 |
32 | 35,000 |
References
FAQ
Can I create cross-region or cross-border serverless instances?
Cross-region or cross-border serverless instances are not supported.
Can I create a serverless instance for a data migration or synchronization task?
You can create a serverless instance only for a data synchronization task.
Can I create a serverless instance for a two-way data synchronization task?
No, you cannot create a serverless instance for a two-way data synchronization task.