Help Center/ GaussDB(DWS)/ Service Overview/ Data Warehouse Types
Updated on 2025-03-03 GMT+08:00

Data Warehouse Types

Product Type Overview

  • Computing In-Memory(CIM): The storage-compute coupled data warehouse provides enterprise-level data warehouse services with high performance, high scalability, high reliability, high security, low latency, and easy O&M. It is capable of data analysis at a scale of 2,048 nodes and 20 petabytes of data and is suitable for converged analysis services that integrate databases, warehouses, marts, and lakes.
  • Decoupled Storage and Compute: The storage-compute decoupled data warehouse is designed with a cloud native architecture that separates storage and compute. It also features hierarchical auto scaling for computing and storage, as well as multi-logical cluster shared storage technology (Virtual Warehouse or VW). These capabilities allow for computing isolation and concurrent expansion to handle varying loads, making it an ideal choice for OLAP analysis scenarios.
  • GaussDB(DWS) data warehouses cannot access each other. You can create an OBS foreign table to associate two databases in the same data directory for data query.

Features

Table 1 Features

Module

Function

Storage-Compute Coupling

Storage-Compute Decoupling

Dashboard

Resources

Yes

Yes

Alarms

Yes

Yes

Recent events

Yes

Yes

Cluster monitoring metrics (DMS)

Supported (excluding standalone systems)

Yes

Data

-

Yes

Yes

Cluster management

SQL editor

Yes

Yes

Monitoring panel (DMS)

Supported (excluding standalone systems)

Yes

Monitoring metrics (Cloud Eye)

Supported (excluding standalone systems)

Yes

Restart

Yes

Yes

Start

Yes

Yes

Stop

Yes

Yes

Scaling

Supported (excluding standalone systems)

Yes

Scale-in

Supported (excluding standalone systems)

Yes

Redistributing data

Supported (excluding standalone systems)

Yes (Note 1)

Viewing redistribution details

Supported (excluding standalone systems)

Yes

Changing the node flavor

Yes

Yes

Changing all specifications

Supported (excluding standalone systems)

No

Resetting passwords

Yes

Yes

Creating snapshots

Supported (excluding standalone systems)

Yes

Canceling read-only status

Yes

Yes

Deletion

Yes

Yes

Managing CNs

Supported (excluding standalone systems)

Yes

Storage space scaling

Yes

Yes

Basic Information

Basic information

Yes

Yes

ELB

Supported (excluding standalone systems)

Yes

Resource Management

Supported (excluding standalone systems)

Yes

Intelligent O&M

Supported (excluding standalone systems)

Yes

Logical cluster

Supported (excluding standalone systems)

Yes

Snapshot

Supported (excluding standalone systems)

Yes

Parameter modifications

Yes

Yes

Security settings

Yes

Yes

MRS data sources

Supported (excluding standalone systems)

Yes

Tags

Yes

Yes

Node management

Supported (excluding standalone systems)

Yes

Upgrade management

Supported (excluding standalone systems)

Yes

Logging

Supported (excluding standalone systems)

Yes

User management

Supported (excluding standalone systems)

Yes

Integration

Data migration

Supported (excluding standalone systems)

Yes

DR management

DR management

Supported (excluding standalone systems)

No

Snapshot management

Restoration

Supported (excluding standalone systems)

Yes

Deletion

Supported (excluding standalone systems)

Yes

Copy

Supported (excluding standalone systems)

Yes

Incident management

Event management (general)

Yes

Yes

Alarm management

Alarm management

Yes

Yes

Client connections

Client connections

Yes

Yes

Others

Inspection

Supported (excluding standalone systems)

Yes

Intelligent O&M

Supported (excluding standalone systems)

Yes

Node restoration

Supported (excluding standalone systems)

Yes

Warm backup on the tenant side

Supported (excluding standalone systems)

Yes

  • Note 1: The storage-compute separated table stores data on OBS, eliminating the need for redistribution. However, metadata and indexes are stored locally and must still be redistributed. Redistributing a table with separated storage and compute nodes only allows for read operations, and metadata redistribution is fast. However, creating an index on the table can affect redistribution performance, with completion time increasing with index data volume. During this period, the table is read-only and cannot be modified.
  • A storage-compute coupled data warehouse deployed in standalone mode cannot be changed to the distributed mode. Scaling, logical cluster operations, and resource management are not supported.
  • A storage-compute coupled data warehouse deployed in standalone mode does not support functions such as snapshots, monitoring panels, and MRS data sources.
  • In storage-compute coupled data warehouse (standalone), the original syntax for specifying distribution columns will be ignored. Scheduled jobs, the sequence type, HDFS/OBS foreign table import and export, multi-temperature storage, and auto-increment partitioned tables are not supported. OBS foreign table import and export are supported in 8.2.0 and later versions.
  • Only storage-compute decoupled clusters of 9.0.2 and later versions support the snapshot function.