Purchase a PostgreSQL Distributed Edition Cluster - PolarDB

PolarDB for PostgreSQL Distributed Edition is a distributed database service built on PolarDB for PostgreSQL (which runs in centralized mode). It uses a two-layer architecture of separated compute nodes (CNs) and data nodes (DNs) to achieve distributed scaling capabilities. This topic describes how to purchase a PolarDB for PostgreSQL Distributed Edition cluster in the PolarDB console.

Prerequisites

Before you begin, make sure that you have a registered and active Alibaba Cloud account.

Purchase a cluster

Go to the PolarDB for PostgreSQL Distributed Edition buy page.
Note
If the buy page displays abnormally or the PostgreSQL Distributed tab is missing, contact us.
Configure the parameters described in the following sections, then click Buy Now.
On the Payment page, confirm your order and payment method, then click Subscribe.

The cluster is available 10 to 15 minutes after payment. View it on the Clusters page.

Note

The cluster is unavailable while any node is in the Creating state. The cluster is ready only when its status changes to Running. Make sure to select the correct region on the Clusters page; otherwise, the cluster will not appear.

Configuration reference

Billing method

Option	Description	Best for
Subscription	Pay upfront for a fixed term. Longer terms offer larger discounts.	Predictable, steady-state workloads
Pay-as-you-go	Pay for actual resource usage, billed hourly. No upfront commitment.	Fluctuating or short-term workloads

You can change the billing method at any time after purchase.

Region and database engine

Parameter	Value
Database engine	PolarDB For PostgreSQL Distributed Edition
Region	The region where the cluster is deployed. Cannot be changed after the cluster is created.
Creation method	Create Primary Cluster (default)
Database engine (version)	PostgreSQL 14
Database edition	Enterprise Edition
Edition	Cluster Edition (Recommended)
CPU architecture	X86

Choosing a region: Deploy the cluster in the same region as the ECS instances running your applications. For example, if your application runs in the ap-southeast-1 region, create your cluster in ap-southeast-1 as well. Clusters and ECS instances in different regions can only communicate over the Internet, which increases latency. Clusters in the same region communicate over the internal network.

For details, see Regions and zones.

Specification

Option	Description	Best for
Dedicated	Compute resources (CPU) are exclusively allocated to the cluster, not shared with other clusters on the same host.	Production workloads requiring consistent performance
General-purpose	Idle compute resources (CPU) are shared among clusters on the same host.	Development, testing, or cost-sensitive workloads

Network and zone

Parameter	Description
Network type	VPC (required)
VPC	Select a VPC in the same region as your ECS instances. Use an existing VPC if it meets your requirements. Otherwise, use the default VPC.
Zone and vSwitch	Select the zone and vSwitch for the cluster. The selected zone becomes the primary zone. Deploying the cluster in the same zone as your ECS instances reduces network latency.

Default VPC and vSwitch: If you do not have a custom VPC, use the default VPC and vSwitch. Each region has one default VPC (CIDR block: 192.168.0.0/16, up to 65,536 private IP addresses) and each zone has one default vSwitch (CIDR block: 192.168.0.0/20, up to 4,096 private IP addresses). Neither counts against your VPC or vSwitch quota.

To create a custom VPC, see Create and manage a VPC.

High-availability mode

Mode	Replicas	SLA	Storage cost	Use when
Double zones (hot standby storage cluster enabled)	6 replicas (3 per cluster) across two zones	Higher	Full price	Production workloads requiring maximum availability
Single zone (hot standby storage cluster disabled)	3 replicas in one zone	Lower	Half price	Development or cost-sensitive deployments

Important

Once enabled, the hot standby storage cluster cannot be disabled. Evaluate your availability and cost requirements before selecting this mode.

When you select Double zones (hot standby storage cluster enabled), also configure the Secondary zone — the zone where the hot standby storage cluster is deployed.

Compute node specifications

A Distributed Edition cluster has two types of computing resources:

Compute nodes (CNs): Manage cluster metadata and generate distributed execution plans.
Data nodes (DNs): Store actual data shards. Each DN automatically creates two replicas.

Configure the Compute node specification and Data node specification for your expected workload. For billing details, see Computing resource billing rules.

Optionally, add one read-only node for CNs or DNs under Read-only node for CNs/DNs.

Storage

Parameter	Description
Storage type	PSL5: Higher performance, reliability, and availability. Supported in historical PolarDB versions. PSL4: Uses Alibaba Cloud smart-SSD technology to compress and decompress data at the physical SSD level, reducing storage cost while maintaining disk performance.
Storage cost	Charged based on actual usage, billed hourly. No need to pre-allocate capacity — storage scales automatically as data grows or shrinks. See Billing rules for storage.