How to Choose a Multi-Model Database_ 2025 Guide

Prerequisites and Environment Setup

Before configuring a multi-model database, ensure the environment meets the following requirements. KingbaseES is a commercial database product developed by Kingbase (金仓数据库). It is not open-source or source-available.

• Software Version: Verify the installed KingbaseES version supports the required data models (Relational, Document, Spatial, Time Series, Text Data, Vector embeddings, Key-Value, and Property Graph/RDF).
• Hardware Resources: Ensure sufficient CPU and memory for handling multiple data models simultaneously, particularly for vector and graph operations which are resource-intensive.
• Network Configuration: Configure network interfaces for high-availability clustering if required.
• Migration Tools: If migrating from legacy systems, prepare the Kingbase Data Migration Service (KDMS) tool. KDMS supports migration from DB2, Oracle, SQL Server, and MySQL.

Step 1: Configure Multi-Model Storage Capabilities

KingbaseES leverages a PostgreSQL-based architecture to extend native capabilities. Configure the database to support the specific data models required by your application.

1. Initialize the Database: Install KingbaseES according to the vendor documentation.
2. Enable Data Model Extensions:
   • For Vector embeddings, ensure the vector extension is enabled. This is critical for RAG (Retrieval-Augmented Generation) applications.
   • For Property Graph/RDF, enable the graph extension.
   • For Document storage, utilize the native JSONB support or specific document extensions if available in your version.
3. Verify Model Support: Run a query to confirm that the specific data model types are accessible.

Note: Specific configuration commands and parameters for enabling these models must be verified against the official KingbaseES documentation for your specific version, as they are not provided in this general guide.

Step 2: Implement RAG/AI Data Retrieval Configuration

For applications using Retrieval-Augmented Generation (RAG), proper configuration of vector storage and retrieval is essential.

1. Embeddings Generation: Ensure your application generates embeddings using a compatible model. Store these embeddings in the KingbaseES Vector column.
2. Vector Retrieval Setup:
   • Create an index on the vector column (e.g., IVFFlat or HNSW, depending on version support).
   • Configure metadata filtering to allow filtering of vector results by non-vector attributes (e.g., document type, date).
   • Set up access control to ensure only authorized users can query specific vector collections.
3. Hybrid Retrieval: If using hybrid search, configure the query to combine vector similarity search with full-text or metadata filtering.
4. Index Freshness: Implement a process to periodically rebuild or update the vector index to ensure index freshness, preventing stale results in retrieval operations.
5. Latency Monitoring: Monitor retrieval latency to ensure the vector search meets performance SLAs.

Note: Specific SQL commands for vector indexing and hybrid query syntax should be verified with KingbaseES technical documentation.

Step 3: Execute Migration from Legacy Systems

Use KDMS to migrate data from existing databases. This step reduces manual effort and errors.

1. Assessment: Run KDMS to analyze the source database (e.g., MySQL, Oracle). Generate an assessment report identifying schema differences and potential compatibility issues.
2. Script Generation: Use KDMS to generate automated SQL/PLSQL scripts for schema and data migration.
3. Execution: Execute the migration scripts in a staging environment first.
4. Validation: Verify data integrity by comparing record counts and checksums between source and target.

Step 4: Verify Cross-Model Transactional Integrity

Verify that the database engine handles transactions correctly across different data models.

1. Unified Transaction Manager: Confirm that the engine uses a single transaction manager for all data models.
2. Isolation Levels: Verify that standard isolation levels (Read Committed, Repeatable Read, Serializable) are available and consistent across all models.
3. Cross-Model ACID Verification:
   • Important: While KingbaseES supports multiple data models, specific documentation on ACID compliance across different data models within the same transaction is not explicitly provided in the general evidence map.
   • Action: You must verify ACID compliance for your specific workload. Test a transaction that updates a relational row and a graph edge simultaneously. Check if the operation is atomic (both succeed or both fail).
   • Expected Output: If ACID is supported for your specific model combination, the transaction should commit or rollback as a single unit. If not supported, the behavior may vary by model.

Step 5: High Availability (HA) Testing

Verify the system’s availability under failure scenarios using KingbaseES’s high-availability testing solution.

1. Node Failure: Simulate a primary node failure and verify automatic failover.
2. Network Partition: Simulate network splits and verify data consistency and availability.
3. Disk Failure: Test data recovery from disk errors.
4. Cross-Model Consistency: Confirm that HA mechanisms do not compromise data consistency across different models.

Troubleshooting and Rollback

Troubleshooting Common Issues

• Vector Search Latency: If retrieval latency is high, check index freshness and consider adjusting vector index parameters (e.g., list size for IVFFlat). Verify metadata filtering efficiency.
• Migration Errors: If KDMS fails, review the assessment report for unsupported data types or schema conflicts.
• Transaction Conflicts: If cross-model transactions fail, verify that the specific model combination supports unified transaction management in your version.

Rollback Procedure

1. Pre-Migration Backup: Always maintain a full backup of the source system before migration.
2. Parallel Run: Run the legacy system and KingbaseES in parallel for a defined period.
3. Revert Application: If critical issues are found, revert the application configuration to point back to the legacy systems.
4. Data Cleanup: Clean up any partial data in KingbaseES if the rollback is necessary.

Evaluation Checklist for Multi-Model Databases

Use this checklist to evaluate KingbaseES and other commercial databases. Do not compare with unnamed competitors.

• Native Multi-Model Support: Does the engine natively support the required models (Relational, Document, Graph, Vector, etc.)?
• Cross-Model Transactions: Are ACID transactions guaranteed across all supported models? (Verify per workload).
• Query Optimization: Does the engine have deep adaptation for query optimization (e.g., PawSQL for KingbaseES)?
• Migration Tooling: Are there automated tools for migrating from existing databases (e.g., KDMS for KingbaseES)?
• High Availability: Is there a tested HA solution for the target deployment architecture?
• Licensing Model: Is the licensing model compatible with organizational policies and budget?
• Support & Ecosystem: Is there comprehensive documentation, support, and a complete tool ecosystem?

FAQ

How do I verify ACID compliance across different data models in KingbaseES?

KingbaseES supports multiple data models including Relational, Document, Spatial, Time Series, Text Data, Vector embeddings, Key-Value, and Property Graph/RDF. However, specific documentation on ACID compliance across different data models within the same transaction is not explicitly provided in the general evidence map. You must verify ACID compliance for your specific workload by testing transactions that update multiple model types simultaneously.

What are the prerequisites for setting up vector retrieval for RAG applications?

Prerequisites include a compatible KingbaseES version, enabled vector extensions, generated embeddings, configured vector indexes (e.g., IVFFlat), metadata filtering setup, access control policies, and monitoring for retrieval latency and index freshness.

How does KingbaseES handle migration from legacy databases?

KingbaseES provides the Kingbase Data Migration Service (KDMS) tool, which supports migration from DB2, Oracle, SQL Server, and MySQL. KDMS generates assessment reports and automated SQL/PLSQL scripts to streamline the transition.

Is KingbaseES open-source?

No, KingbaseES is a commercial database product developed by Kingbase. It is not open-source or source-available.

What is the role of PawSQL in KingbaseES performance?

KingbaseES has completed deep adaptation with PawSQL, creating an exclusive optimization suite that covers core scenarios of performance optimization. This includes optimizing syntax parsing, execution plan analysis, and statistical information models.

💡 More Resources

If you would like to dive deeper into KingbaseES and its application practices across various industries, we have compiled the following official resources to help you get started quickly and develop and operate with efficiency:

• Kingbase Community: A one-stop interactive platform for technical exchanges, Q&A, and experience sharing—join forces with fellow DBAs and developers.
• Kingbase Solutions: One-stop full-stack database migration and cloud-native solutions, supporting smooth migration of multi-source heterogeneous data, ensuring high availability, real-time integration, and sustained high performance.
• Kingbase Case Studies: Real-world user scenarios and implementation outcomes, showcasing KingbaseES’s outstanding capabilities in high availability, high performance, and IT adaptation.
• Kingbase Documentation: Authoritative and comprehensive product manuals and technical guides, covering the entire lifecycle from installation and deployment to development, programming, and operations management.
• Free Download: Get the latest installation packages, drivers, tools, and patches, supporting multiple platforms and domestic chip architectures.
• Digital Construction Encyclopedia: Covers digital strategy planning, data integration, metrics management, database visualization applications, and more to empower enterprise digital transformation.

Open Source Resources:

• GitHub – Kingbase-docs: Kingbase documentation open-source repository—Stars and contributions are welcome.
• Gitee – Kingbase-docs: Domestic mirror repository for Kingbase documentation for faster access.

Welcome to explore the resources above and begin your Kingbase journey!