All Products
Search

This Product

    ApsaraDB RDS:Disaster recovery solutions

    Document Center

    ApsaraDB RDS:Disaster recovery solutions

    Last Updated:Jun 17, 2025

    RDS PostgreSQL provides multiple disaster recovery solutions to ensure business continuity and high availability.

    Comparison of disaster recovery solutions

    Note

    • RTO (Recovery Time Objective) refers to the time between when a failure occurs and when the system is restored to an available state.

    • RPO (Recovery Point Objective) refers to the maximum acceptable data loss duration during a disaster or system failure.

    Solution

    RTO

    RPO

    Price

    Advantages

    Disadvantages

    Scenarios

    Global Distributed Cache (GAD)

    Typically 10 to 30 seconds

    Within 5 minutes, typically 30 seconds to 1 minute

    Medium

    • Instance-level disaster recovery based on RDS PostgreSQL native physical replication, efficient and stable.

    • Network fees are charged based on the usage of one-way synchronization links.

    • Disaster recovery instances are readable.

    • The entire process can be operated in the console.

    Both source and disaster recovery instances must be RDS PostgreSQL with consistent major versions.

    Suitable for cross-region disaster recovery between RDS PostgreSQL instances and cross-region read-only scenarios.

    One-click disaster recovery setup

    Typically 10 to 30 seconds

    Within 5 minutes, typically 30 seconds to 1 minute

    Slightly high

    • Source instances can be self-built local instances or instances hosted by other cloud service providers.

    • The entire process can be operated in the console.

    Users need to ensure network connectivity themselves and bear the corresponding network costs.

    Suitable for scenarios where self-built local instances or instances hosted by other cloud service providers use RDS PostgreSQL as disaster recovery instances.

    Cross-region backup

    Typically hours

    Within 5 minutes, typically 30 seconds to 1 minute

    Low

    • Based on snapshot backup, with no impact on the source database.

    • Backups are stored in Object Storage Service (OSS).

    • RTO time is longer.

    • The minimum upload unit for logs is 20 MB. When business changes are small, RPO is higher.

    Suitable for scenarios where the source database is RDS PostgreSQL and cross-region disaster recovery does not have high requirements for recovery time.

    Data synchronization (DTS)

    Typically 10 to 30 seconds

    Within 5 minutes, typically 30 seconds to 1 minute

    High

    • Can synchronize specific databases and tables.

    • Through two-way synchronization technology, it can effectively meet the needs of multi-active scenarios.

    • Uses asynchronous replication with a storage layer in the middle. When the disaster recovery system resumes after an interruption, there is no need to completely rebuild the link.

    Data synchronization is at the database level. When the number of instance databases is large, the cost is higher.

    Suitable for non-full database disaster recovery or multi-active scenarios.