A Data Transmission Service (DTS) dedicated cluster consists of multiple Elastic Compute Service (ECS) instances of the same specifications in a region. A DTS dedicated cluster is used to manage and configure DTS data migration, data synchronization, and change tracking tasks. Compared with DTS public clusters, DTS dedicated clusters are characterized by exclusive resources, better stability, better performance, and reduced costs.
Benefits
Compared with DTS public clusters, which are used to manage and configure the regular DTS instances that you purchase, DTS dedicated clusters have the following benefits:
Exclusive resources with high stability
DTS dedicated clusters provide independent resources that are isolated from the resources used by DTS instances of other users. Your DTS instances that are run in DTS dedicated clusters can never be affected by DTS instances of other users. In addition, you can control the loads of DTS dedicated clusters to enhance the stability of your business.
High performance
In a DTS dedicated cluster, you can allocate the resources of the entire cluster to a single DTS instance to improve the performance of the DTS instance. This allows you to write large amounts of data to the source database and synchronize the data to the destination database in real time.
Reduced costs
You are charged only for the resources of a DTS dedicated cluster. You do not need to pay any fee when you create a DTS instance in a DTS dedicated cluster. The cost is further reduced if you enable the overcommitment mode for a DTS dedicated cluster compared with that of purchasing a large number of DTS instances.
Limits
DTS dedicated clusters do not support the classic network.
To allow a DTS dedicated cluster to access a database over the Internet or an Express Connect circuit, you must add the public and private CIDR blocks of the region in which the dedicated cluster resides to the whitelist of the database.
If the usage of resources such as the CPU, memory, or disks in a DTS dedicated cluster exceeds 80%, you must scale up the cluster at the earliest opportunity. The Service Level Agreement is not applicable to task failures or switchover failures caused by insufficient resources.
DTS dedicated clusters support only the subscription billing method. You must pay when you create a DTS dedicated cluster.
You are charged only for the resources of a DTS dedicated cluster. You do not need to pay any fee when you create a DTS instance in a DTS dedicated cluster.
If you want to configure a DTS data synchronization task in a DTS dedicated cluster, the source database, destination database, and DTS dedicated cluster must reside in the same region.
The number of nodes and the size of disks in a DTS dedicated cluster cannot be reduced.
When you purchase a DTS dedicated cluster, you can select the number of nodes within the range of 2 to 100. When you determine the number of nodes, we recommend that you estimate the resources that are required to run your DTS tasks and keep the usage of DTS units (DUs) in the cluster lower than 50%.
DTS dedicated clusters support only the node specifications of 15 DUs, which indicates that you can configure up to 15 concurrent DTS instances on a single node. If the overcommitment mode is enabled, you can obtain up to 30 DUs of instance configuration resources. For example, if the overcommit ratio is set to 200%, you can configure up to 30 concurrent DTS instances on a single node.
A region can contain up to five DTS dedicated clusters. If you need to create more than five DTS dedicated clusters in a region, contact Alibaba Cloud technical support to add your account to the whitelist for creating more DTS dedicated clusters.
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.
Cluster node specifications: the specifications of a node in a DTS dedicated cluster. Only 15 DUs are supported. This indicates that on a cluster node that is not overcommitted, you can run up to 15 data migration, data synchronization, or change tracking tasks at the same time.
Cluster node storage: the storage capacity of a single node in a DTS dedicated cluster. The storage capacity of a single node ranges from 1,024 GB to 30,720 GB. The system reserves an amount of storage space on a node. The remaining storage capacity is used to cache the data generated within the previous seven days by the data synchronization, data migration, and change tracking instances that are run on the node. DTS automatically clears the data that was generated seven days ago. You must make sure that the total storage capacity of a DTS dedicated cluster is greater than the total size of the existing data in the DTS instances and the available storage capacity is greater than the total size of the incremental data in the DTS instances. Otherwise, an instance exception may occur. You can specify a threshold for the Disk Usage metric on the Cluster Alert Management page to monitor the storage usage of a node in a DTS dedicated cluster. If the disk space cannot meet your business requirements, you can extend the disk capacity for the DTS dedicated cluster.
Resource isolation: the mode in which resources are isolated for different DTS tasks. If a DTS dedicated cluster is not overcommitted, the hardware resources allocated to each DTS task are isolated from each other. For example, DTS Instance A requests four DUs of hardware resources and DTS Instance B requests 10 DUs of hardware resources. The hardware resources of the two DTS instances are isolated from each other.
Cluster overcommit ratio: the ratio of overcommitting cluster resources. The cluster overcommit ratio ranges from 100% to 200%. For example, if you set the overcommit ratio to 200%, the number of DUs on a node is changed from 15 to 30. In this case, you can configure up to 30 DTS instances on a node. If the overcommitment mode is enabled, hardware resource contention occurs among DTS instances. For example, you create 30 DTS data synchronization instances based on 30 DUs. Although each DTS instance requires only one DU of resources, a DTS instance may fail to preempt one DU of resources. Therefore, the overcommitment mode is applicable only if the amount of incremental data to be synchronized is small or the DTS instances are run in different time periods.
Billing
For more information about the billing method, billable items, and billing rules of DTS dedicated clusters, see Billing of DTS dedicated clusters.
Supported regions
DTS dedicated clusters are supported in the following regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Shenzhen), China (Guangzhou), Singapore, Japan (Tokyo), and US (Virginia).
References
For more information about how to configure a DTS dedicated cluster, see Create a DTS dedicated cluster.
For more information about how to configure data migration, data synchronization, and change tracking tasks in a DTS dedicated cluster, see Configure a data migration task, Configure a data synchronization task, and Configure a change tracking task.