AnalyticDB for MySQL offers five editions — Enterprise Edition, Basic Edition, Data Lakehouse Edition, Data Warehouse Edition in elastic mode, and Data Warehouse Edition in reserved mode — each designed for different workload profiles and cost requirements.
At a glance
| Enterprise Edition | Basic Edition | Data Lakehouse Edition | Data Warehouse Edition (elastic mode) | Data Warehouse Edition (reserved mode) | |
|---|---|---|---|---|---|
| Architecture | Storage-compute decoupled + coupled | Storage-compute decoupled + coupled | Storage-compute decoupled | Storage-compute decoupled | Storage-compute coupled |
| High availability | Yes (multi-replica) | No (single-replica) | Cluster mode | Cluster Edition only | — |
| Spark engine | Yes | Yes | Yes | No | No |
| Auto scaling | Yes | Yes | Yes | No | No |
| Best for | Production workloads across all enterprise sizes | Low-cost hot data storage without HA | Large-scale batch processing + real-time analytics | Real-time writes, ETL, log analysis | Query acceleration, interactive reports |
Edition evolution
AnalyticDB for MySQL has evolved through five editions over five years.
Data Warehouse Edition in reserved mode launched first, using a storage-compute coupled architecture for high-performance queries. Because it does not allow independent scaling of compute and storage, Data Warehouse Edition in elastic mode followed with a storage-compute decoupled architecture, enabling elastic compute scaling and physical resource isolation for multi-tenancy and hybrid workloads.
As demand for semi-structured and unstructured data analysis grew, Data Lakehouse Edition built on elastic mode to add the Spark compute engine for large-scale batch processing. Enterprise Edition and Basic Edition then unified the benefits of all prior editions, combining reserved resources for predictable performance with elastic scaling for variable workloads.
Enterprise Edition
Enterprise Edition is designed for teams that need production-grade reliability and the full feature set — real-time analytics, batch processing, elastic scaling, and tiered storage — in a single edition.
It includes all capabilities of both Data Lakehouse Edition and Data Warehouse Edition:
-
From elastic mode: resource group isolation, elastic resource scaling, tiered storage of hot and cold data
-
From reserved mode: high throughput, real-time writes, high-concurrency real-time queries
Enterprise Edition runs in multi-replica mode with a multi-replica storage architecture, providing distributed capabilities and high availability. It suits development, testing, and production environments across enterprise sizes.
Basic Edition
Basic Edition is designed for teams that need the full Enterprise Edition feature set at lower cost and can accept the tradeoff of no high availability.
Basic Edition runs in single-replica mode and provides the same features as Enterprise Edition. Because it uses a single-replica storage architecture, it does not support high availability. It suits business scenarios that require low-cost hot data storage without a high availability requirement.
You cannot change a Basic Edition cluster to Enterprise Edition.
Data Lakehouse Edition
Data Lakehouse Edition is designed for data engineering teams running large-scale batch processing alongside high-performance real-time analysis.
It uses a storage-compute decoupled architecture with fully enhanced data collection, storage, computing, and application capabilities compared to Data Warehouse Edition in elastic mode. A single copy of data at the underlying storage layer serves both batch processing and real-time analysis, eliminating the consistency and timeliness issues that arise from separate data synchronization paths. Computing resources are physically isolated for batch and real-time workloads.
Key capabilities:
-
Scheduled and automatic scaling of compute and storage resources
-
Spark multi-language programmable compute engine with standardized APIs for large-scale batch processing
-
Real-time data synchronization to Hudi tables on Object Storage Service (OSS) or C-Store tables, configured through a visual interface
Data Lakehouse Edition runs in cluster mode and suits: data processing (cleansing and standardization), multi-source aggregate analysis, wide table development, and prediction and insights (machine learning and AI).
Data Warehouse Edition
Elastic mode
Data Warehouse Edition in elastic mode is designed for teams writing large volumes of data in real time and running complex analytical queries or extract, transform, load (ETL) pipelines.
It uses a storage-compute decoupled architecture with physical isolation of computing resources for batch processing and real-time analysis. Compute and storage resources scale independently. Tiered storage of hot and cold data reduces storage costs.
Typical use cases: real-time large-volume writes, complex ETL, complex queries on large datasets, and historical data and log analysis.
Data Warehouse Edition in elastic mode is available in two deployment options:
Standalone Edition
Standalone Edition is deployed on a single node. It does not provide distributed architecture benefits or high availability. No service-level agreement (SLA) is provided, and a failover requires 4–8 hours. Do not use Standalone Edition in production environments.
It supports tiered storage of hot and cold data but does not support resource group isolation or scheduled scaling. Standalone Edition suits individual developers running tests and startups or small enterprises with basic business needs that do not require large data volumes, high queries per second (QPS), or high availability.
Cluster Edition
Cluster Edition is deployed across multiple nodes. It delivers distributed architecture benefits, supports high availability, and provides more powerful features for enterprise development, testing, and production workloads.
Reserved mode
Data Warehouse Edition in reserved mode is designed for teams with steady, predictable workloads that require high-concurrency query performance and fast response times.
It uses a storage-compute coupled architecture and delivers high-throughput real-time writes, high concurrency, and quick response. It does not allow independent scaling of compute or storage resources.
Typical use cases: query acceleration, user profiling, interactive reports, and real-time data services.
Feature comparison
The following table compares features across all editions.
| Category | Feature | Enterprise Edition | Basic Edition | Data Lakehouse Edition | Data Warehouse Edition in elastic mode | Data Warehouse Edition in reserved mode |
|---|---|---|---|---|---|---|
| Computing | XIHE analytical compute engine | Supported | Supported | Supported | Supported | Supported |
| Spark programmable compute engine | Supported | Supported | Supported | Not supported | Not supported | |
| Storage | XUANWU analytical storage engine | Supported | Supported | Supported | Supported | Supported |
| Cost-effective Hudi storage | Supported | Supported | Supported | Not supported | Not supported | |
| Resource management | Resource group management | Supported | Supported | Supported | Cluster Edition only | Not supported |
| Scheduled scaling | Supported | Supported | Supported | Cluster Edition only | Not supported | |
| Auto scaling | Supported | Supported | Supported | Not supported | Not supported | |
| Tiered storage | Tiered storage for hot and cold data | Supported | Supported | Supported | Supported | Not supported |
| Data ingestion | Real-time data import | Supported | Supported | Supported | Supported | Not supported |
| Automatic metadata discovery | Supported | Supported | Supported | Not supported | Not supported | |
| Job development | SQL job development | Supported | Supported | Supported | Not supported | Not supported |
| Spark job development | Supported | Supported | Supported | Not supported | Not supported |
Specifications
Enterprise Edition and Basic Edition
| Category | Specifications |
|---|---|
| Single-node specifications of reserved resources | 8 ACUs, 12 ACUs, 16 ACUs, 24 ACUs, 32 ACUs |
Data Lakehouse Edition
| Category | Specifications |
|---|---|
| Reserved computing resources | Minimum: 16 ACUs / Maximum: 4,096 ACUs |
| Reserved storage resources | Minimum: 24 ACUs / Maximum: 2,064 ACUs |
To purchase more than 512 ACUs of reserved computing resources or more than 256 ACUs of reserved storage resources, submit a ticket.
Data Warehouse Edition in elastic mode
| Category | Specifications |
|---|---|
| Computing resources — Standalone Edition | 8 cores and 32 GB memory; 16 cores and 64 GB memory |
| Computing resources — Cluster Edition |
32 cores and 128 GB memory or more |
| Storage resources — Elastic I/O resources | 24 cores and 192 GB memory; 36 cores and 288 GB memory; 48 cores and 384 GB memory |
Data Warehouse Edition in reserved mode
| Model | Specifications | ||
|---|---|---|---|
| CPU | Memory (GB) | Storage (GB) | |
| C8 | 24 cores | 192 | Minimum: 100 / Maximum: 2,000 |
| C32 | 96 cores | 768 | Minimum: 100 / Maximum: 8,000 |
Edition changes
The following edition changes are supported:
-
Data Warehouse Edition or Data Lakehouse Edition to Enterprise Edition or Basic Edition (how to change)
-
Data Warehouse Edition in reserved mode to Data Warehouse Edition in elastic mode (how to change)
-
Basic Edition to Enterprise Edition: not supported
FAQ
How do I view the edition of an AnalyticDB for MySQL cluster?
Log on to the AnalyticDB for MySQL console and go to the Cluster Information page of a cluster. In the Cluster Attributes section, view the edition and deployment mode.
When can cluster availability be affected?
Availability may be affected when failures occur, or when the cluster undergoes configuration changes or version updates.
How are AnalyticDB for MySQL clusters billed?
For billing methods and billable items, see Billing overview.