All Products
Search

This Product

    PolarDB:What is PolarDB for PostgreSQL (Compatible with Oracle)

    Document Center

    PolarDB:What is PolarDB for PostgreSQL (Compatible with Oracle)

    Last Updated:Mar 30, 2026

    Your existing Oracle stored procedures, data types, SQL statements, and built-in functions work in PolarDB for PostgreSQL (Compatible with Oracle) with minimal changes. Built on a compute-storage separation architecture, this cloud-native database delivers Oracle-level reliability and performance without Oracle licensing costs.

    All compute nodes share a single copy of data through PolarFileSystem, a distributed file system backed by PolarStore. This architecture enables resource scaling in minutes and fault recovery in seconds. A cluster has one primary node (read/write) and up to 15 read-only nodes. PolarProxy, the internal proxy layer, handles security authentication, parses SQL statements, routes writes to the primary node, and load-balances reads across read-only nodes—so your application connects to the cluster just like a single-node database.

    Benefits

    Large capacity

    Default storage capacity is up to 500 TB. Add storage without sharding or purchasing additional instances.

    PolarStore (PSL4/PSL5) supports petabyte-level storage. To reserve petabyte-level resources, contact us.

    Low cost

    • Shared storage: Adding a read-only node charges only for compute resources. Traditional databases charge for both compute and storage on each added node.

    • Elastic storage: No pre-provisioned storage needed. Storage scales automatically with your data volume, billed by the hour.

    High performance

    Delivers over 500,000 read requests per second and over 150,000 write requests per second for online transactional processing (OLTP) workloads.

    Scaling within minutes

    The compute-storage separation architecture and container virtualization let you add or remove a node in 5 minutes. Storage scales out online without interrupting your services.container virtualization

    Read consistency

    Cluster endpoints use Log Sequence Numbers (LSNs) to guarantee global read consistency, eliminating inconsistencies caused by replication delays between the primary and read-only nodes.

    Millisecond-level replication latency

    Physical replication based on redo logs replaces binary log-based logical replication. Even Data Definition Language (DDL) operations on large tables—such as adding an index or a column—do not cause replication latency.

    Fast backup within seconds

    A full backup completes in 30 seconds regardless of data volume. Backups run without locking the database and have almost no impact on running applications.

    Cross-node parallel execution

    SQL statements execute in parallel across multiple compute nodes, using all available CPU, memory, and network resources to accelerate analytical query performance. For details, see Cross-node parallel execution.

    Pricing

    See Billable items or visit the PolarDB buy page.

    Manage clusters

    Manage your PolarDB for PostgreSQL (Compatible with Oracle) clusters—create clusters, databases, and accounts—through any of the following interfaces:

    • Console: A graphical web interface.

    • CLI: You can perform all console operations using the CLI.

    • CLI: Run all console operations from the command line.

    • SDK: Integrate all console operations into your application using the software development kit (SDK).

    • API: Call the API to automate cluster management.

    After creating a cluster, connect to it using:

    • Data Management Service (DMS): for web-based database development.

    • Client tools: Use pgAdmin or any PostgreSQL-compatible client.

    Key concepts

    Term Definition
    Cluster One primary node plus up to 15 read-only nodes.
    Region A physical data center. Place your cluster in the same region as your Elastic Compute Service (ECS) instance for optimal latency.
    Zone A physically isolated area within a region with independent power and networking. Zones in the same region are functionally equivalent.
    Specification The resource configuration of a node, such as 2 vCPUs and 8 GB memory.

    Related services

    • ECS: Access your PolarDB cluster from an ECS instance in the same region over an internal network for optimal performance.

    • Redis: Combine ECS, PolarDB, and ApsaraDB for Redis to handle high-traffic read workloads and reduce response times.

    • MongoDB: Store structured data in PolarDB and unstructured data in ApsaraDB for MongoDB when your application has diverse data structures.

    • DTS: Use Data Transmission Service (DTS) to migrate on-premises databases to a PolarDB for PostgreSQL (Compatible with Oracle) cluster.

    • OSScloud storage: Object Storage Service (OSS) is a massive, secure, low-cost, and highly reliable cloud storage service from Alibaba Cloud.