Supported DTS synchronization topologies and usage notes - Data Transmission Service

DTS supports multiple synchronization topologies. Choose a topology based on your business requirements.

One-way synchronization

To ensure data consistency, perform only read operations on synchronized objects in the destination instance.

Topology type	Topology diagram	Description
One-way one-to-one synchronization		None
One-way one-to-many synchronization		Requires multiple synchronization instances. Example: synchronizing from Instance A to Instances B, C, and D requires three instances.
One-way cascade synchronization		Requires multiple synchronization instances. Example: synchronizing from Instance A to B, then B to C requires two instances.
One-way many-to-one synchronization		Requires multiple synchronization instances. Example: synchronizing from Instances B, C, and D to Instance A requires three instances. Note In multi-table merge scenarios where changes from multiple source objects target the same destination row, concurrent execution may cause data inconsistency. When you add a synchronization instance to this topology, the precheck may return a warning for the check for objects with the same name. You can skip this check in the Precheck step in the console or by calling an API operation. Skipping the check does not affect the synchronization instances that are already running. To handle primary key conflicts in multi-source aggregation scenarios, go to Task Details > Parameters and set the `sink.conflict.strategy` parameter of incremental synchronization to `replace`. This enables the overwrite write policy and keeps the data in the destination database eventually consistent. The default value is `insert`, which does not overwrite existing rows.

The following databases support two-way synchronization.

Note

If the Database Type is MongoDB or Tair/Redis, two-way synchronization is supported only for Alibaba Cloud database instances. In this case, the Access Method must be Alibaba Cloud Instance.
DTS supports two-way synchronization only between two databases.
You can upgrade a one-way synchronization task to two-way. Upgrade a one-way synchronization task to a two-way synchronization task.

Source database type	Destination database type	Configuration document
MySQL	MySQL	Two-way synchronization between MySQL databases
MySQL	PolarDB for MySQL	Two-way synchronization between a MySQL database and a PolarDB for MySQL cluster
PolarDB for MySQL	MySQL	Document in preparation
PolarDB for MySQL	PolarDB for MySQL	Two-way synchronization between PolarDB for MySQL clusters
PostgreSQL	PostgreSQL	Two-way synchronization between PostgreSQL databases
PostgreSQL	PolarDB for PostgreSQL	Two-way synchronization between RDS for PostgreSQL and PolarDB for PostgreSQL
PolarDB for PostgreSQL	PostgreSQL	Two-way synchronization between PolarDB for PostgreSQL and RDS for PostgreSQL
PolarDB for PostgreSQL	PolarDB for PostgreSQL	Two-way synchronization between PolarDB for PostgreSQL clusters
PolarDB (Compatible with Oracle)	PolarDB (Compatible with Oracle)	Two-way synchronization between PolarDB for PostgreSQL (Compatible with Oracle) clusters
PolarDB-X 2.0	PolarDB-X 2.0	Two-way synchronization between PolarDB-X 2.0 instances
Tair/Redis	Tair/Redis	Two-way synchronization between Tair (Enterprise Edition) instances
MongoDB (ReplicaSet architecture)	MongoDB (ReplicaSet architecture)	Two-way synchronization between ApsaraDB for MongoDB ReplicaSet instances
MongoDB (sharded cluster architecture)	MongoDB (sharded cluster architecture)	Two-way synchronization between ApsaraDB for MongoDB sharded cluster instances

Topology type	Topology diagram	Description
Two-way one-to-one synchronization		To ensure data consistency, update records with the same primary key, business primary key, or unique key in only one instance. Note If a record is updated in both instances, DTS resolves the conflict based on the conflict resolution policy configured for the synchronization task.