All Products
Search
Document Center

Hologres:Hologres instance selection

Last Updated:Jul 04, 2025

This topic describes the instance types of Hologres and their applicable scenarios.

Selection recommendations

The following table describes various business scenarios suitable for different types of Hologres instances. You can select an instance type based on the information in this table and your actual business requirements.

Instance type

Resource type

Core requirement

Applicable scenario

Virtual warehouse instance (recommended)

Exclusive computing resources and storage resources

  • Supports high-performance, high-concurrency, and large-scale data processing.

  • Implements a read/write splitting architecture in a single instance.

  • Large-scale data services with data volumes at the TB or PB level, and frequent complex queries.

  • High-concurrency online services with millisecond-level response requirements (above 1000 QPS).

  • Mixed workload isolation, such as OLAP queries and real-time writing, or resource isolation requirements for multiple lines-of-business.

  • Elastic scaling requirements, such as significant traffic fluctuations (periodic peaks) that require rapid scaling.

Serverless instance

Shared computing resources, and exclusive storage resources

  • Supports high-performance, large-scale, and low-frequency data processing and data queries.

  • Development or testing environments.

  • Low-frequency and high-performance query scenarios.

  • Scenarios with significant business load fluctuations, unpredictable peaks, and difficulty in reserving exclusive computing resources in advance.

General-purpose instance

Exclusive computing resources and storage resources

  • Smaller data volume (below TB level).

  • Lightweight analysis, such as low-frequency ad hoc queries, and small to medium-scale BI report generation.

  • Cost-sensitive businesses, such as limited budget requiring high cost-effectiveness solutions. Non-core business trial stages.

  • Low-concurrency services, such as internal data service platforms. Scheduled task processing.

Read-only secondary instance (to be deprecated)

Exclusive computing resources and storage resources

  • General-purpose primary instances serve as data writing and data processing instances, while read-only secondary instances serve as data analysis instances, ensuring read/write splitting.

Shared Cluster (Lakehouse Acceleration Edition) (to be deprecated)

Shared computing resources, and no storage resources

  • Supports low-cost acceleration for lakehouse data queries and BI integration.

  • No need for read and write operations on Hologres internal tables.

  • Scenarios where data is stored in OSS or MaxCompute and you want to use Hologres for accelerated queries.

  • Scenarios with burst large queries where reserved exclusive instance resources are insufficient.

  • Scenarios requiring BI tools to accelerate lakehouse data queries. Hologres Shared Cluster can be used to connect with Tableau, Quick BI, FineBI, and other BI tools.

Capability comparison

The following table compares different Hologres instance types across four dimensions to help you better understand and select the right instance type.

Note

已开通 indicates the feature is supported, and 未开通 indicates it is not supported.

Comparison item

Virtual warehouse instance (recommended)

Serverless instance

Primary-secondary instance (with read-only secondary instances to be deprecated)

General-purpose instance

Lossless scaling

已开通

Not required

未开通

未开通

Time-specific scaling

已开通

Not required

未开通

未开通

Workload isolation

已开通

已开通

已开通

未开通

Automatic throttling

已开通

已开通

未开通

未开通

Serverless Computing

已开通

已开通

General-purpose instance: 已开通

Read-only secondary instance: 未开通

已开通