Migration & Modernization
Accelerate Your Mainframe Modernization Journey using AI Agents with AWS Transform
Introduction
Mainframe modernization has long been a challenging task for organizations, due to complex legacy codebases of critical applications, shrinking talent pool with mainframe expertise, and increasing need to embrace modern cloud capabilities.
AWS Transform for mainframe, a breakthrough solution for modernizing legacy systems that was previewed as ‘HAQM Q Developer transform capability for mainframe’ at re:Invent 2024 is now generally available. AWS Transform’s specialized AI agent can significantly accelerate the traditionally multi-year mainframe modernization journey, helping organizations complete their transformations in faster pace.
In this blog we will explore how AWS Transform for mainframe, using its specialized AI agent accelerates mainframe modernization from years to months by streamlining the entire transformation process.
The mainframe modernization challenge
Organizations during their mainframe modernization journey today encounter the challenges across three critical dimensions:
1. Speed and Agility: Legacy mainframe systems contain millions of lines of COBOL, PL/I, and Assembler code, developed across decades. These systems house complex business logic and data relationships critical to daily operations. The monolithic architecture of mainframe applications restricts rapid changes and innovations. Traditional modernization approaches demand thorough manual analysis and refactoring, resulting in multi-year implementation timelines. The contrast is clear: mainframe systems require lengthy development cycles and rigid change management, while modernized applications enable rapid changes, upgrades and deployments. This gap creates competitive disadvantages as organizations struggle to meet market demands with the speed required in today’s business environment. Organizations must balance preserving critical mainframe functionality while accelerating their systems’ ability to evolve and adapt.
2. Complex to execute: Mainframe applications typically entail tightly integrated monolithic architectures that don’t easily translate to cloud-ready architectures. Complex business logic is often buried deep within legacy code, which often lacks clear documentation. Additionally, these systems often employ nested and intricate coding patterns that were developed decades ago but do not translate well with modern architectures. This architecture resists modular transformation approaches, making the modernization process inherently complex and challenging to execute effectively. This complexity creates significant risk in modernization projects. Without proper analysis and understanding, critical business logic can be misinterpreted or lost during transformation, potentially resulting in costly business disruptions.
3. Skills Shortage: The mainframe talent landscape presents a significant strategic consideration for organizations. The current mainframe workforce consists of experts with decades of specialized knowledge in both applications and systems. As these professionals transition out, they take with them invaluable institutional knowledge about critical business systems. Organizations must compete for a limited pool of qualified personnel who understand both legacy and modern technologies. This skills dynamic makes workforce evolution a crucial factor in long-term technology strategy decisions.
Introducing AWS Transform
AWS Transform represents a breakthrough in modernization technology through its innovative multi-agent AI architecture. This specialized agentic AI system analyzes mainframe code bases, decomposes them into domains, and modernizes IBM z/OS applications to Java using its orchestrated AI agent. Mainframe agent specializes in specific transformation tasks to revolutionize the modernization process. The system’s architecture leverages deep learning models along with AWS extensive mainframe modernization expertise, accumulated from 19 years of enterprise migration projects. Users can interact with AWS Transform’s agents to create customized modernization plans through an interactive dialogue, defining target patterns and objectives that map their unique transformation journey from analysis through code conversion. This reduces modernization timelines from years to months while maintaining functional equivalence. The comprehensive solution enables organizations to modernize faster, reduce risk and cost, and optimize their applications for AWS cloud deployment. By encoding nearly two decades of AWS’s modernization expertise into an intelligent agent, AWS Transform for Mainframe delivers an efficient and reliable path to cloud modernization.
The below figure shows the phases of end-to-end mainframe modernization journey.
Figure 1: End-to-end mainframe modernization journey
As shown in preceding diagram, lets dive deep into the capabilities of AWS Transform and learn how it influences each phases of the modernization journey.
Key capabilities overview
AWS Transform for Mainframe offers a comprehensive suite of AI-powered capabilities designed to accelerate and simplify mainframe modernization. From initial analysis through transformation and deployment, AWS Transform addresses the core challenges of mainframe modernization while maintaining functional equivalence and reducing risk. Here are the key capabilities that our customers and partners are using to transform mainframe applications:
Code analysis for comprehensive application insights
Many organizations face challenges in understanding the scope and complexity of their established mainframe applications, which often support critical business processes. The AWS Transform agent during the analysis, performs a comprehensive examination of the mainframe codebase and create a detailed dependency graphs that map relationships between components. The agent analyzes program interactions, identifies related missing files, and generates key metrics including code complexity, lines of code, and file type distributions.
It automatically categorizes different types of code components including JCL, COBOL, and Copybooks, conducts dependency analysis to identify relationships between components, and flags missing artifacts that could impact modernization. By delivering these detailed insights, the analysis agent enables organizations to better understand their application landscape and make informed decisions during the modernization process, ultimately reducing risk and optimizing the path to cloud migration.
We are enriching the results provided in AWS Transform code analysis by identifying the following three file attributes across the code base:
1. Cyclomatic complexity – measuring how many different paths or decision points exist within a program flow.
2. Identically named files – identifying files types with same names
3. Duplicate IDs (same program ID) – detecting multiple programs using the same identifier
AWS Transform provides enhanced file analysis capabilities with detailed classification of source files in the codebase. Users can update the classification of unknown or .txt file extensions through an export/import function, giving them control over file type management. The AWS Transform user interface includes built-in file viewing and comparison features, allowing users to examine and compare source files directly within the platform. These features result in following business benefits:
Business benefits:
1. Save time and resources by automating complex analysis tasks
2. Improve decision-making based on application insights
Document generation for preserving application knowledge and extracting business logic
AWS Transform generates detailed technical and functional documentation of the mainframe applications to tackle the challenge of knowledge gap. The documentation describes key features, programs logic and functionality, data flows and dependencies, integrations, and more details. This ensures that both high-level summaries and detailed functional specifications are available during your modernization journey.
Business logic extraction
AWS Transform’s business logic extraction capability, provides comprehensive insights for both business and technical stakeholders during the modernization journey. For business users, it extracts and presents complex logic in plain language, offering clear visibility into business processes, calculations, and decision rules embedded within legacy applications. This enables business stakeholders to validate current rules, identify outdated processes, and make informed decisions about process optimization during modernization. Technical users benefit from detailed mappings of business logics to specific code segments, along with clear identification of core algorithmic patterns, computational logic, and dependencies between business logics and data structures.
Chat with documentation knowledgebase
At general availability, AWS Transform now features an intelligent assistant that learns alongside your modernization journey. Through an intuitive chat interface and natural language queries, users can engage with the comprehensive documentation generated by the agent, enabling dynamic knowledge discovery and informed decision-making. This AI-powered chat assistant proves to be valuable throughout your project, providing contextual insights and responses based on your application’s specific documentation, making the modernization process more collaborative and accessible.
These features result in following business benefits:
Business benefits:
1. Mitigate the risk of knowledge loss during modernization by preserving critical application insights despite employee turnover
2. Accelerate on-boarding of new team members to the modernization project
3. Improve understanding of applications for modernization initiatives
4. Enable faster and more accurate modernization decisions through comprehensive application understanding
5. Enable real-time knowledge discovery through contextual chat interactions with documentation
6. Bridge the gap between technical implementation and business requirements with business logic extraction
7. Empower non-technical stakeholders to participate in modernization decisions
Code decomposition for enhancing agility
Monolithic mainframe applications present significant modernization challenges due to their size and interconnected nature. AWS Transform’s capability for decomposition of large applications helps breakdown monoliths into smaller, maintainable domains based on customer guidance. AWS Transform addresses this by decomposing these complex applications into manageable domains during the modernization process, enabling organizations to achieve the agility and maintainability benefits of cloud architectures.
At general availability, AWS Transform introduces significant enhancements to the dependency mapping experience, including:
1. The ability for users to update dependencies through an export/import function
2. Tools for users to interact with the relationships (parent, child, neighbor) between components when creating domains
3. Support for importing domain details into AWS Transform, enabling customers to easily create logical domains
These features result in following business benefits:
Business benefits:
1. Increase business agility via alignment of application components with business domain
2. Enable more collaborative domain definition with import/export capabilities
3. Improve application understanding through interactive relationship exploration
4. Support customized modernization approaches with flexible domain creation
Planning modernization waves for efficient execution
AWS Transform’s planning capability creates prioritized modernization wave sequences based on multiple factors including code and data dependencies, code volume, and business priorities. Users can input their specific constraints and priorities in order to customize the proposed multi-wave plans. These features result in following business benefits:
Business benefits:
1. Align modernization efforts with business priorities and constraints
2. Enable data-driven decision making for transformation sequencing
3. Reduce risk through well-structured modernization phases
Code refactoring for mainframe application transformation
AWS Transform’s refactor capability automates the code conversion process, converting COBOL to java and JCL to groovy scripts modernizing the complete application stack. The specialized AI agent maintains functional equivalence while producing readable and maintainable code, refactoring business domains in a sequence defined by the human-in-the-loop. At general availability, AWS Transform delivers refactoring capabilities that accelerate the conversion process while maintaining application functionality.
Reforge – enhances refactored Java code for better maintainability (Public Preview)
In public preview, we’re introducing reforge, a new capability designed to optimize refactored code. AWS Transform’s reforge leverages Large Language Models (LLMs) to elevate the transformed code to a new level. This advanced feature restructures the code to closely resemble native Java, improving readability and maintainability. Reforge adds human-readable comments to enhance code understanding and introduces modern coding patterns and best practices. This evolution aims to align modernized applications closely with modern development standards, facilitating easier maintenance and future enhancements in the cloud environment. These features result in following business benefits:
Business benefits:
1. Accelerate the modernization of mainframe applications with millions of lines of code to AWS
2. Minimize errors and maintain functional equivalence, reducing business risk
3. Produce modern and maintainable code
4. Enable faster iteration and innovation with AWS application architecture
Code deployment for accelerated cloud migration
Organizations face slow, manual configuration processes and complex enterprise requirements when creating cloud environments for refactored applications. AWS Transform addresses this challenge by providing deployment templates that help create standardized environments and establish repeatable modernization processes, enabling a more structured approach to application transformation. At general availability, AWS Transform delivers deployment templates that provide infrastructure-as-code foundations for modernized application environments. These templates reduce the time and expertise needed to configure target environments. These features result in following business benefits:
Business benefits:
1. Reduce time-to-value by accelerating the deployment of refactored applications
2. Minimize configuration errors through standardized templates
3. Lower the technical barriers to completing the modernization lifecycle
4. Enable consistent, repeatable deployment processes across application portfolios
These integrated capabilities work together to provide our customers and partners with a comprehensive modernization solution that reduces risk, accelerates delivery, and maintains functional equivalence during the transformation journey.
Pricing of AWS Transform for mainframe
AWS Transform accelerates migration and modernization projects for mainframe workloads with agentic AI capabilities. Currently, we offer our core features—including assessment and transformation—at no cost* to AWS customers. This allows you to speed up your migration and modernization journey without upfront expenses.
*As we continue to expand AWS Transform capabilities, future add-on capabilities may be introduced as paid features.
AWS Transform for accelerating mainframe application transformation in action
Now, let’s explore how AWS Transform streamlines and accelerates the mainframe modernization journey through a collaborative web experience. After logging in to AWS console, navigate to AWS Transform and create a workspace to begin the transformation process.
Step 1: Comprehensive analysis
AWS Transform begins by analyzing your mainframe codebase stored in an HAQM S3 bucket. The analysis agents create detailed dependency graphs that map relationships between components, analyze program interactions, and identify related missing files. As shown in the below screenshot , the analysis provides key metrics including code complexity, lines of code, and file type distributions across the codebase.
Figure 2: AWS Transform code analysis with enhanced metrics
The ability to view and compare source files directly within the interface enables teams to quickly understand code characteristics while the export/import function allows for correction of file classifications, ensuring accuracy in the analysis phase.
Step 2: Application knowledge for modernization
After analysis, as you see in the following screenshot, AWS Transform generates comprehensive technical documentation for your selected programs. This can be viewed online or downloaded to your HAQM S3 bucket.
Figure 3: AWS Transform document generation with chat interface
The new chat experience allows team members to interact with the generated documentation, asking specific questions about program functionality and receiving contextual responses. For business stakeholders, the business logic extraction feature translates technical implementations into business language, bridging the gap between IT and business teams.
Step 3: Application decomposition
AWS Transform then helps break down monolithic applications into logical business domains by using user provided seed files (example programs that belongs to the domain). As per the following screenshot below, The enhanced dependency mapping features allow teams to interact with file relationships (parent, child, neighbor) when creating domains, and support for importing domain details enables more collaborative domain definition.
Figure 4: AWS Transform application decomposition with interactive relationship visualization
This visualization helps teams understand the complex inter-dependencies within their applications and make informed decisions about how to modularize their systems. Users can zoom-in to the dependency graphs and analyze multiple layers of dependencies for each components.
Step 4: Modernization planning
With domains established, AWS Transform creates prioritized modernization waves based on dependencies, code volume, and business priorities. The comprehensive planning tools allow you to customize wave sequences according to your specific constraints.
Figure 5: AWS Transform modernization wave planning
These visual planning tools facilitate clear communication with stakeholders and ensure a structured approach to the transformation process.
Step 5: Automated refactoring
Once the human-in-the-loop confirms the modernization plan as in step 4, AWS Transform begins the refactoring process, converting your COBOL to Java code and JCL to Groovy script.
Figure 6: AWS Transform automated refactoring with improved code quality
While standard refactoring preserves the functional equivalence of your COBOL applications, our new public preview feature – Reforge leverages large language models to enhance the readability and maintainability the refactored code. This capability transforms the output beyond simple translation, restructuring it to follow Java best practices and idioms. The resulting code is maintainable, with improved readability through human-like comments and documentation, making it easier for Java developers to understand and extend the application without requiring COBOL expertise.
Step 6: Streamlined deployment
The new code deployment capabilities provide default templates for setting up modernized application environments. These templates significantly reduce the time and expertise needed to properly configure target environments. Once transformed, the application can be deployed either on AWS Mainframe Modernization’s fully-managed runtime environment or in a self-managed environment. Both options support deployment on HAQM EC2 or HAQM EKS containers in customers’ own HAQM VPC securely in their respective AWS account. This ensures organizations have the flexibility in their modernization approach while maintaining application performance, security and reliability. This end-to-end approach ensures that organizations can efficiently transform their mainframe applications while preserving critical business logic, reducing risk, and accelerating their journey to modern cloud architectures.
What’s new in this release
1. Enhanced Code Analysis: Identification of cyclomatic complexity, identically named files, and duplicate program IDs
2. Export/import function to update classification of unknown or .txt file extensions
3. File viewing and comparison capabilities within the user interface
4. Interactive relationship exploration (parent, child, neighbor) between files when creating domains
5. Support for larger code bases (soft limit: 3M LOC per AWS account) with improved documentation generation performance
6. New chat experience aware of generated documentation content
7. Business Logic Extraction allowing business users to understand functionality and logic in source code
8. Flexible document viewing and interaction within the AWS Transform interface
9. Infrastructure-as-code foundations for modernized application environments
10. Available in US East-1 (N. Virginia) and Europe (Frankfurt) AWS Regions
11. Public Preview: LLM-powered code restructuring to optimize refactored java code with human-readable comments
Conclusion
AWS Transform for mainframe represents a significant leap forward in modernization technology, offering organizations a comprehensive, AI-powered solution to accelerate their journey to the cloud. By combining specialized AI agents with AWS’s proven mainframe migration expertise, organizations can now modernize their critical applications with reduced risk, cost, and complexity. To learn more about how AWS Transform can accelerate your modernization initiatives, visit the AWS Transform documentation or contact your AWS representative today.