What is PolarDB for PostgreSQL Distributed Edition?

Updated at:
Copy as MD

PolarDB for PostgreSQL Distributed Edition is a distributed database built on PolarDB for PostgreSQL. It scales horizontally to handle sub-petabyte data volumes, ensures consistency in distributed transactions, and preserves all the capabilities of the centralized edition—making it the right choice when your data or write throughput exceeds what a single node can handle.

How it works

The cluster uses a two-layer architecture with compute nodes (CNs) and data nodes (DNs):

LayerComponentRole
1Compute node (CN)Manages cluster metadata and generates distributed query plans
2Data node (DN)Stores actual data shards

Each CN and DN internally uses compute-storage separation, the same design as PolarDB for PostgreSQL. Every node independently supports single-writer, multiple-reader mode, database proxy, read-only nodes, and failover mechanisms—while also participating in the distributed cluster.

image

Key capabilities

  • Horizontal scaling with manual sharding — Add CNs or DNs to handle data volumes below petabytes (PB).

  • Distributed transaction consistency — Ensures consistency in distributed transactions across all nodes.

  • Dual-zone deployment — The active cluster runs in the primary zone. If the primary zone fails, the hot standby cluster in the secondary zone seamlessly takes over to ensure uninterrupted service.

  • Zero-downtime operations — Maintenance, upgrades, and configuration changes run 24/7 without service interruption. Heterogeneous CNs and DNs can be added at any time.

  • Centralized management — Manage the entire distributed cluster—including cross-node Operations and Maintenance (O&M)—from a single console interface.

Use cases

PolarDB for PostgreSQL Distributed Edition is designed for scenarios where data volume or write throughput exceeds what a single centralized database can handle. Typical use cases include:

  • SaaS multi-tenancy — Isolate tenant data across shards while keeping a unified management view.

  • Smart driving training — Process continuously growing datasets generated by training pipelines.

  • Blockchain — Handle high write volumes and large data accumulation from on-chain transactions.

If your data volume fits within a single PolarDB for PostgreSQL node, the centralized edition is the simpler choice.

Benefits

Integration of centralized and distributed architectures

Each CN and DN can operate independently as a centralized cluster or as part of the distributed cluster. This lets you retain centralized cluster advantages—transparent compression, plug-in support, and single-writer, multiple-reader scalability—without giving them up when you scale out.

Simplified management

A unified distributed view lets you perform all cross-node O&M from one interface, removing the need to manage each node separately.

Billing

Fees for a PolarDB for PostgreSQL Distributed Edition cluster include:

  • Computing resource fees

  • Storage usage fees

  • SQL Explorer fees (optional)

For details, see Distributed Edition billing.

Get started

Purchase a cluster: Purchase a Distributed Edition cluster.

Manage your cluster using any of the following interfaces. All interfaces support the same set of operations—creating clusters, databases, and accounts.

InterfaceDescription
ConsoleWeb-based GUI
CLIAlibaba Cloud CLI
SDKProgrammatic access
APIDirect API calls

Connect to your cluster after it is created:

  • DMS (Data Management) — Connect and develop databases directly in the browser. See Use DMS to connect to a cluster.

  • Client — Connect using common database clients, such as pgAdmin.