All Products
Search
Document Center

ApsaraDB RDS:Global Active Database

Last Updated:Jul 04, 2025

Overview

To ensure business continuity and high availability (HA), ApsaraDB RDS for PostgreSQL supports Global Active Database (GAD) to provide disaster recovery capabilities. This feature delivers a one-stop, cost-effective cross-region disaster recovery solution based on Data Transmission Service (DTS) and the native physical replication feature of RDS for PostgreSQL. In the solution, a primary instance and a secondary instance are added to a GAD instance group. Data is synchronized from the primary instance to the secondary instance in real time by using DTS. This achieves cross-region disaster recovery and allows users to access data from the nearest region.

Architecture

image

Scenarios

GAD is applicable for the following scenarios:

  • Cross-region disaster recovery

    Some businesses have extremely high requirements on availability. To avoid region-level failures, the database system in these businesses must implement cross-region disaster recovery to ensure availability and recovery time objective (RTO).

  • Cross-region read-only request handling

    In some businesses, users need to query data from the nearest region. The data system in these businesses must handle the cross-region read-only requests.

Benefits

  • Secure and reliable

    • Implemented based on the native replication feature provided by RDS for PostgreSQL, GAD offers stable performance and is fully compatible with the database services of Alibaba Cloud.

    • Supported by the guaranteed Service Level Agreement (SLA) provided by RDS and DTS, GAD can ensure the availability and security of databases and synchronization tasks.

  • Convenient and easy to use

    • You can quickly create and configure a GAD instance group for disaster recovery in the product-ready RDS console.

    • GAD provides various capabilities to help you implement disaster recovery through the complete lifecycle, such as disaster recovery assessment, one-click promotion, and latency display.

  • Cost-effective

    • The secondary instance (disaster recovery instance) in GAD instance groups supports cost-effective RDS series and editions, such as Yitian RDS instances and RDS instances that run Basic Edition .

    • The data synchronization in GAD instance groups is implemented by DTS. The bandwidth used for data synchronization can be automatically scaled using a serverless approach. You are charged only for the resources that are actually used when you use GAD.

Limits

Limits on instances

The primary instance (original instance) and secondary instance (disaster recovery instance) in a GAD instance group must meet the following conditions:

  • Major engine version: The instance must run PostgreSQL 11 or a later version.

    Note

    The major engine version of the primary and secondary instances must be the same.

  • Edition: The instance must run Basic Edition or High-availability Edition.

  • Billing method: The instance must use the pay-as-you-go or subscription billing method.

Limits on regions

The primary instance and secondary instance must be located in one of the following regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Shenzhen), China (Hohhot), China (Heyuan), and China (Chengdu).

Billing

When you use GAD, you are charged for the RDS instance that you newly purchase to create the GAD instance group and the data synchronization capability provided by DTS. If you create a GAD instance group by using instances that you have already purchased, you are charged only for DTS.

  • Fees for RDS instances newly purchased for GAD

    If you newly purchase an RDS for PostgreSQL instance and use it as the primary or secondary instance of a GAD instance group, you are charged fees for the billable items of the instance, such as instance type and storage capacity, on a subscription or pay-as-you-go basis.

  • Fees for data synchronization

    When you add a secondary instance to a GAD instance group, a one-way data synchronization task between the primary and secondary instances is created. You are charged fees for the tasks based on the region of the instances, task type, the number of tasks, and the usage duration of the task. You can calculate the fees by using the following formula:

    Fees for data synchronization = Unit price for data synchronization task x Number of tasks x task usage duration

    The unit price for one-way synchronization task is USD 0.187969924812/hour/task.

How to use