PolarDB for PostgreSQL(Compatible with Oracle) is a cloud-native database built for teams migrating away from Oracle. It combines Oracle syntax compatibility with a distributed shared-storage architecture that delivers elastic scaling, high availability, and serverless storage.
Oracle compatibility
PolarDB for PostgreSQL(Compatible with Oracle) is highly compatible with Oracle syntax. Migrate your services from Oracle databases with no code changes, or at most minor modifications.
Cost efficiency
Compute-storage separation — Compute nodes share a single storage pool. When you add read-only nodes, you pay only for the additional compute, which greatly reduces scale-out costs.
Serverless storage — Storage capacity scales automatically with your data volume. No capacity planning required; you pay only for the storage you actually use.
High performance
An enhanced database kernel with physical replication, Remote Direct Memory Access (RDMA) protocol, and shared distributed storage drives substantial throughput gains.
A cluster supports one primary node and up to 15 read-only nodes, making it well-suited for read-heavy, high-concurrency workloads.
Because all nodes share the same storage pool, a data change only needs to be written once across the primary and read-only nodes.
Storage capacity for hundreds of terabytes
PolarDB uses distributed block storage with a file system. Storage scales up automatically independent of any single node's capacity limit, supporting databases up to hundreds of terabytes.
High availability, reliability, and data security
Zero data loss on node failure — Shared distributed storage eliminates the data inconsistency risk introduced by asynchronous primary-secondary replication. A single point of failure in the cluster results in zero data loss.
Multi-zone redundancy — Data replicas are distributed across multiple zones, providing disaster recovery and backup coverage without manual replication setup.
Layered security controls — Access protection includes IP whitelist-based connection filtering, Virtual Private Cloud (VPC) network isolation, and multiple data replicas for storage durability.
Elastic scaling in under 5 minutes
Configuration changes — Scale CPU and memory up or down within 5 minutes using container virtualization combined with shared distributed block storage.
Node addition and removal — Add or remove nodes within 5 minutes. Cluster endpoints abstract the underlying topology, so applications are unaware of node changes.
Lock-free backup
PolarDB snapshots TB-scale databases in minutes using the snapshot technology of the underlying distributed storage layer. No locks are acquired during backup, which delivers higher efficiency and minimizes the negative impact on your workload.
Cross-node parallel execution
SQL queries run in parallel across multiple compute nodes, fully utilizing the CPUs, memory, and network bandwidth of every node in the cluster. This improves the performance of analytical queries. For more information, see Cross-node parallel execution.