You can use deployment sets to deploy RDS Custom instances on different physical servers to ensure the high availability (HA) of your workloads and disaster recovery (DR) of the physical servers. You can also use deployment sets to deploy RDS Custom instances in the same network topology in a centralized manner to reduce the network latency between the instances and improve the performance of business applications.
When to use a deployment set
You want to improve service availability. You can use deployment sets to deploy RDS Custom instances on different physical servers to prevent single points of failure (SPOFs) and ensure service stability and reliability.
You want to reduce the network latency. If your applications are sensitive to the network latency, such as real-time trading systems and online gaming services, you can use deployment sets to deploy RDS Custom instances in the same network topology of the same zone to reduce the network latency between the instances.
Billing rules
You can use deployment sets free of charge. However, you are charged for your RDS Custom instances.
Limits
Regions and zones: RDS Custom instances and the deployment sets to which the RDS Custom instances belong must reside in the same region. RDS Custom instances in a deployment set that uses the low latency policy must be deployed in the same zone.
Deployment sets cannot be merged.
Deployment policies
A deployment policy determines the distribution and deployment mode of RDS Custom instances on physical servers to meet specific business requirements and performance metrics.
Deployment set policy | Description | Scenario | Maximum number of instances |
HA (Availability) | All RDS Custom instances in a deployment set are strictly distributed to different physical servers in a region. | This policy is suitable for scenarios in which RDS Custom instances must be isolated from each other to prevent SPOFs. |
|
HA group (AvailabilityGroup) | A deployment set can be divided into up to seven groups. RDS Custom instances can be distributed to the groups of a deployment set as needed. RDS Custom instances in different groups are strictly distributed to different physical servers in a region to prevent SPOFs. RDS Custom instances in the same group may be deployed on the same physical server to reduce the network latency. | This policy is suitable for large-scale deployment scenarios and scenarios that require the prevention of SPOFs and low network latency between RDS Custom instances. |
|
Low latency strategy (LowLatency) | All RDS Custom instances are centrally deployed in the same network topology in the same zone to reduce the network latency. In this case, multiple RDS Custom instances may be deployed to the same physical server, and HA cannot be ensured. | This policy is suitable for latency-sensitive HA application clusters. |
|
Using deployment sets
For more information about how to create a deployment set, see Create an RDS Custom deployment set.