Accelerate database modernization using AI with the Database Modernizer Workshop

In this post, we show how you can use the Database Modernizer workshop to accelerate your database modernization journey from MySQL to Amazon DynamoDB. Database modernization is a complex, time-intensive initiative that calls for careful planning and participation from everyone involved. Developers, architects, and business users can redefine patterns and improve processes. Traditional approaches to migrating from relational databases to NoSQL solutions like DynamoDB can take several months, requiring extensive expertise in data modeling, application refactoring, and migration strategies. The Database Modernizer workshop, which can be scheduled by your account team as part of the Amazon DynamoDB Immersion Day program, uses AI to help you complete database modernization projects in days instead of months.

The Database Modernizer workshop provides a new, AI-powered approach that simplifies and accelerates database modernization. It is a multi-stage workflow covering from source database discovery to data movement. Each stage uses narrowly scoped prompts designed to provide outputs that will feed the next stage, reducing context windows and long, unnecessary re-analysis. It combines fine-tuned prompts, Model Context Protocol (MCP) servers, and human-guided decision points. The result is a structured, stage-based process that includes AWS modernization best practices that maintains the quality and reliability customers expect while reducing the time and expertise required for successful modernization projects.

Although this workshop currently focuses on MySQL to DynamoDB modernization, the underlying approach and architecture is designed to be extensible. Customers can adapt the modular MCP server framework to support additional source databases such as PostgreSQL, Oracle, SQL Server, etc. and target systems Amazon DocumentDB, Amazon Neptune, Amazon OpenSearch Service, etc. We’re actively expanding our capabilities and would love to hear from you about which database combinations would be most valuable for your modernization initiatives.

Solution overview

The Database Modernizer workshop provides a fully managed, self-contained environment through Workshop Studio. When you start the workshop, you will have an environment that is ready for your use with:

• A pre-configured Workshop Studio environment
• Required AWS service permissions
• Sample application and database
• Necessary development tools

No additional setup or an AWS account is required to complete this workshop.

Modernization workflow

The workshop guides you through seven stages that transform your MySQL database into a fully operational DynamoDB solution. Each stage builds upon the previous one, facilitating a systematic and controlled modernization process.The following diagram illustrates the seven stages of the modernization workflow, from initial analysis through final data migration.

Stage 1: Source database analysis

Using automated analysis tools, you first examine your existing MySQL database structure. The workshop automatically analyzes MySQL performance logs, table structures, and backend code to discover your current data architecture and relationships. Instead of manual cataloging, the system documents your database schemas and identifies access patterns.The analysis output shown in the following screenshot shows one artifact with the API access patterns documented, including its methods and API path.

Stage 2: NoSQL data modeling

In this interactive stage, you transform your relational model into an optimized DynamoDB design. You will experience how the data model takes shape in real time. This stage involves using the DynamoDB MCP server to examine access patterns, data velocity, and business requirements, recommending table structures, partition keys, and global secondary indexes (GSIs). When the data model is ready, the workflow generates a migrationContract.json file that maps MySQL tables to DynamoDB structures, including necessary joins and transformations. This file serves as the blueprint for subsequent stages.

Stage 3: Dual database abstraction layer

Following zero-downtime migration best practices, you create an abstraction layer that supports switching between MySQL and DynamoDB. This approach maintains your existing APIs and business logic while adding the flexibility to route database operations to both systems based on feature flags. The abstraction layer makes sure your application continues functioning throughout the migration process and provides a safe rollback option in case there are problems.

Stage 4: DynamoDB integration

This stage implements the core DynamoDB connectivity and data transformation logic. The workshop guides you through AWS SDK configuration and generates code that converts MySQL data structures into DynamoDB optimized formats. Following test-driven development principles, each component begins with a non-working unit test. You incrementally add functionality until the tests pass, supporting proper error handling and retry logic. As an additional safety measure, the system runs the entire test suite to verify that existing functionality remains intact throughout the integration process.

Stage 5: Feature flag control system

The workflow implements a comprehensive feature flag system that enables gradual, reversible transitions between MySQL and DynamoDB. These feature flags provide fine-grained control over database operations, supporting dual-write scenarios where data is written to both systems simultaneously and dual-read scenarios where responses from each database are evaluated against each other to check for consistency. This approach supports a thorough validation of data consistency and application behavior before proceeding with the full cutover. Through an intuitive admin console, you can monitor and control each aspect of the migration, achieving a safe and controlled modernization process.As shown in the following screenshot of the admin console interface, you can control feature flags and monitor each phase of your migration progress in real time.

Stage 6: Infrastructure deployment

This stage automates the creation of your DynamoDB infrastructure using AWS CloudFormation. The workshop converts your migrationContract.json file into infrastructure as code, so your deployed DynamoDB tables precisely match your data model design. Each table, index, and configuration setting is automatically provisioned according to the specifications developed during the modeling phase. This automated approach minimizes manual configuration errors and maintains consistency between your design decisions and the actual deployed infrastructure.

Stage 7: Data movement

The final stage orchestrates the data migration process using AWS Glue, implementing a carefully designed five-phase approach:

• Phase 1 (MySQL only) – The application runs exclusively on MySQL to establish a baseline and validate the abstraction layer’s functionality
• Phase 2 (dual writes and MySQL reads) – This step introduces DynamoDB writes while maintaining MySQL as the exclusive read source, allowing for write operation comparison without affecting user experience
• Phase 3 (dual operations with MySQL primary) – MySQL provides the official response while DynamoDB executes parallel queries for validation
• Phase 4 (DynamoDB primary with MySQL fallback) – This step promotes DynamoDB to the primary data source while maintaining MySQL as a safety net
• Phase 5 (DynamoDB only) – This step represents complete migration success, with DynamoDB handling the database operations independently

Implementation and technical architecture

The Database Modernizer workshop provides a preconfigured environment where you can practice database modernization using a sample ecommerce application. The following components are configured on an EC2 instance used to simulate an on-premise server:

• Workshop Studio environment – pre-configured VS Code with necessary tools
• Sample ecommerce application – Express.js backend with 48 access patterns and React frontend
• MySQL database – complete with sample schema and realistic data
• MCP Server integration – three specialized servers for modernization tasks

The following architecture diagram shows how users interact with the MCP servers throughout the modernization process, demonstrating the flow between source and target databases.

The workshop uses three specialized MCP servers:

• MySQL MCP Server – Analyzes source database and discovers schemas
• DynamoDB MCP Server – Handles NoSQL data modeling and provides best practices guidance
• Data Processing MCP Server – Manages ETL operations and AWS Glue integration

As shown in the following screenshot, the sample ecommerce application provides a realistic environment for practicing modernization techniques, complete with product catalog, shopping cart functionality and complex data relationships.

Prompt engineering

Each stage uses a structured three-document approach:

• Requirements document: Defines business objectives and acceptance criteria
• Design document: Specifies technical architecture and implementation approach
• Tasks document: Provides step-by-step execution instructions

The following diagram shows the three-document structure used in each stage to help provide consistent and reliable outcomes throughout the modernization process.

This structure minimizes generative AI hallucinations while providing comprehensive coverage of modernization aspects.

Getting started

The Database Modernizer workshop is available through the Amazon DynamoDB Immersion Day program. To begin your modernization journey, reach out to your account team to schedule an LGAM: GenerativeAI Application Modernization workshop. In addition, access the prompts used during this modernization workshop via the GitHub repository where you can find additional information.

By following the workshop’s steps and understanding the prompts, the sample application can be updated in approximately 4 hours. The complete workflow typically takes 11–12 hours to finish (starting from scratch), representing a reduction in time compared with traditional modernization approaches. Execute the workshop using your preferred AI coding agent, such as Amazon Q, Kiro, Cline, or Cursor—the prompts are publicly available in the repository.

For a deeper understanding of the concepts covered in the workshop, we recommend exploring the DynamoDB Labs. The workshop will teach practical skills in database modernization, DynamoDB data modeling, migration, feature flags, and code generation.

Conclusion

The Database Modernizer workshop represents a leap forward in making database modernization accessible to a range of customers and use cases. Using generative AI with established migration methods helps us give everyone expert guidance and simplify modernization.As we continue to enhance the workshop based on customer feedback, we’re expanding support for additional source databases and target systems. The extensible prompt-based architecture allows for rapid iteration and customization to meet diverse customer needs.

Ask your account team to host the database modernizer workshop for you. See for yourself how AI can transform one of the most complex aspects of cloud migration into a streamlined, guided process. For new features or improvements to the workshop, use the “Workshop feedback” template to create an issue in the GitHub repo.

For additional resources, refer to the Amazon DynamoDB Documentation.

About the authors