Starting August 6, 2025, DuckDB-based analytical read-only instances are available for ApsaraDB RDS for MySQL. Powered by columnar storage and vectorized execution, these instances run complex analytical queries up to 100 times faster than the InnoDB engine.
How it works
A DuckDB-based analytical read-only instance runs a built-in DuckDB engine alongside your existing RDS for MySQL setup.
Data synchronization happens in two phases:
-
Historical data: When you create the instance, RDS automatically copies existing data from the primary instance and converts it to DuckDB's columnar format.
-
Incremental data: After the instance is created, changes from the primary instance are replicated in real time using MySQL native binary logging replication — entirely within RDS, with no external sync tools required.
Resource isolation: The analytical instance handles complex queries independently. Your primary instance and standard read-only instances continue handling transactions without interference.
Key capabilities
| Capability | What it means for you |
|---|---|
| Columnar storage | Column-oriented data layout reduces I/O for large-scale scans and aggregations |
| Vectorized execution | Processes data in batches rather than row by row, cutting CPU cycles for analytical workloads |
| Just-in-time (JIT) compilation | Compiles query plans to native code at runtime, eliminating interpreter overhead for complex queries |
| Efficient memory management | Optimized memory usage for handling large analytical workloads |
| Parallel processing | Distributes computation across CPU cores, reducing query latency as data volume grows |
| MySQL compatibility | 100% compatible with the MySQL protocol and data types; over 99.9% compatible with SQL syntax and Data Definition Language (DDL) — you can obtain results efficiently without changing your original search statements |
Use cases
Aggregation and analysis
Analytical instances excel at aggregation-heavy workloads — log analysis, time-series aggregations, and reports across large-scale datasets. If your current queries are slow on InnoDB, the DuckDB engine can significantly reduce query time.
Multi-table JOIN queries
Services with wide schemas and complex JOIN paths — such as order analytics, user behavior analysis, or financial reporting — benefit from the columnar layout, which reads only the columns each query needs instead of full rows.
Replace or simplify your OLAP pipeline
If you currently maintain a separate OLAP database with a Data Transmission Service (DTS) data synchronization link, a DuckDB-based analytical read-only instance can eliminate that pipeline. Data flows within RDS at no extra synchronization cost.
Comparison
| DuckDB-based analytical read-only instance | Standard read-only instance | OLAP database | |
|---|---|---|---|
| Primary use case | Complex analytical queries | Transaction processing | Complex analytical queries |
| Analytical query performance | High | Low | High |
| Data synchronization | Native binary logging replication | Native binary logging replication | DTS data synchronization link |
| MySQL data type compatibility | Fully compatible | Fully compatible | Incompatible (requires field mapping) |
| SQL syntax compatibility | Highly compatible (over 99.9%) | Fully compatible | Incompatible (requires SQL rewrite) |
| DDL compatibility | Highly compatible | Fully compatible | Partially compatible |
| O&M costs | Low (integrated instance) | Low (integrated instance) | High (requires extra maintenance for the database and synchronization link) |
Billing
Billing for DuckDB-based analytical read-only instances follows the same model as standard read-only instances. The cost depends on the product series, instance type, storage type, and storage capacity. The actual price is shown on the purchase page.