Accelerating Mainframe Modernization: Combining IBM Analysis and Renovation Catalyst (ARC) and Kiro for intelligent workflow driven reimagining

The Evolution of Mainframe Modernization 

As per recent analyst studies, Mainframe remains core to Fortune 500 organizations mission critical business and part of modernization journey. As organizations increasingly explore hybrid patterns aligning with IBM-AWS joint strategy of “right workload on right platform”, reimagine with rearchitecting is one of the key modernization patterns as it provides a balanced and pragmatic approach bringing the best of both mainframe and cloud provider like AWS. This strategy enables organizations to maintain stable and efficient critical mainframe applications, while selectively modernizing applications needing enhanced agility, scalability, or integration with modern technologies.

Mainframe modernization still relies heavily on manual analysis, such as reverse engineering and dependency mapping, which is inherently labor-intensive and susceptible to inconsistencies. Although AI brings substantial acceleration, it cannot guarantee full accuracy without deterministic parsing and human interpretation. These realities reinforce the need for a hybrid model that combines deterministic parser-based techniques, human-in-the-loop validation, and AI-driven augmentation.

By choosing a hybrid approach, companies can derisk the modernization journey by avoiding "big bang" transformation, optimize costs by modernizing only where it adds business value, and leverage cloud-native capabilities for specific applications while keeping core mainframe strengths. This approach also helps organizations manage the pace of change according to their resources, skills, and business priorities.

Key Challenges in Mainframe Reimagining

Mainframe reimagining faces two critical challenges when it comes to legacy code complexity and skills gaps. Organizations face significant hurdles when dealing with understanding of decades-old undocumented code bases, complex business rules embedded within monolithic applications, and intricate interdependencies between programs that challenge modern transformation efforts. The presence of multiple programming languages and coding styles, along with the absence of modern development practices, further complicates the transformation journey. To address these critical challenges, organizations are now turning to innovative solutions that combine powerful tools and methodologies.

A New Approach to Modernization

IBM Analysis and Renovation Catalyst (ARC), Mainframe Application Modernizer (MAM) and Kiro Spec-based development approaches now accelerate mainframe application modernization through reimagining. IBM ARC tool provides automated capabilities to extract and analyze business rules from legacy mainframe applications, helping to understand and document core business logic that has evolved over decades. The MAM tool accelerates and improves the accuracy of incremental modernization rollout planning by increasing efficiency in analyzing static, dynamic, and user-defined data to define actionable increments.Teams can then leverage this extracted knowledge with Kiro Spec-based development to enable a specification-driven approach that rebuilds applications in modern architectures. Together, these tools create a more efficient pathway from legacy to modern systems - IBM ARC handles the complex task of understanding existing business rules, while Kiro Spec helps translate these rules into modern application specifications that can be rapidly developed using current technologies and cloud-native principles.

IBM ARC: Unlocking Legacy Business Rules

The Analysis and Renovation Catalyst (ARC) asset from IBM Consulting provides insights for modernizing mainframe and distributed applications. It can perform static code analysis to identify the program structure, unused and unfinished code, dependencies, and extract business rules. It can also monitor batch processes to predict job failures, improve process flows and help re-write application code in a standard format for microservices. Parsers are available for wide variety of programming languages including COBOL, PL/I etc.

ARC has different deployment options:

·      It can be run as a CLI tool by the developers, architects etc.

·      There is a web interface with augmented with generative ai which is available via IBM Consulting Advantage (ICA) for architects, business analyst, modernization SMEs etc.

ARC is container ready and can be run as a web UI on EKS or ROSA as well.

The IBM ARC tool produces the below mentioned deliverables

Mainframe Application Modernizer (MAM)

MAM builds a common graphical model by aggregating metadata from multiple sources and applies algorithms on top of it. It enables bounded context analysis, generates heatmaps, and derives disposition assessments using a range of quantitative and qualitative metrics to create a modernization roadmap.

Generate Heat Map to know the degree of coupling between different technical components across multiple Groups (Application / Business Area / Data Domain)

Carry-out Technical Assessment of the Groups to arrive at T-shirt sizing based on different parameters

Identify the interfaces and the upstream and downstream impacts for any logical grouping by doing Context Analysis 

Simulate multiple ‘what-if’ scenarios by creating ‘Increments’ by providing a seed artifact (such as transaction), which can automatically pull related components.

Iterate variations of Increments and utilize Heat Map, Technical Assessment and Context Analysis features to identify potential ‘to-be’ target design.

For scenarios requiring performance evaluation of Mainframe Application, utilize Operational Analysisfeature to see the operational metrics for transactions, jobs and datasets and see their associated call-chain

Kiro's Feature Set for Modernization Success

Kiro's comprehensive feature set offers significant advantages in mainframe modernization efforts through several key capabilities.

Knowledge Management

Kiro delivers a comprehensive feature set that streamlines mainframe modernization through several key capabilities. Model Context Protocols(MCP) provides teams with a robust framework to access documentation, knowledge bases, and best practices, ensuring they maintain current information throughout the transformation process.".

Prompt Management and Reusability

Kiro enables teams to save and reuse successful prompts through a slash command system. Users can store proven prompts as named templates and invoke them later using /promptname commands. This feature promotes consistency across projects, reduces repetitive work, and helps teams standardize their approach to common modernization tasks.

Interactive Documentation Translation

Vibe Kiro's Vibe coding interface provides an interactive way to transform technical IBM ARC documentation into business-friendly documentation. Through natural conversation-style interactions, users can input ARC-extracted mainframe code documentation and receive clear, business-oriented explanations and documentation. This helps bridge the gap between technical mainframe specifications and business requirements, making the modernization process more accessible to non-technical stakeholders. Additionally, Vibe coding serves as an interactive troubleshooting assistant, helping developers resolve issues during forward engineering by providing contextual guidance and solutions for modern architecture implementation challenges.

Governance and Quality Control

Kiro's steering documents serve as governance guardrails, maintaining consistency and quality throughout the modernization journey. These documents provide standardized templates and patterns, ensuring uniformity across different application components and teams.

The Power of Spec-Based Development

Kiro's spec-based development approach bridges legacy mainframe functionality with modern application development through a structured process. Beginning with a carefully crafted prompt, Kiro generates an initial requirements document that considers both technical and non-technical user personas. After stakeholder review and acceptance, Kiro produces a detailed design document and implementation plan. The implementation tasks can then be executed automatically to generate transformed code ready for testing. This process incorporates human oversight at each phase, ensuring proper review and validation of all changes.

This starts with a well thought out prompt.

Kiro then create an initial requirements document, that can be reviewed and updated. Kiro generates documents that incorporate requirements based on different user personas. Once the requirements document is accepted by the user, it creates a detailed design document and an implementation plan with detail tasks. Kiro then executes these tasks with a click of a button to generate the transformed code, allowing teams to deploy for further testing. Each phase requires human interaction to review, update and accept changes. Teams can then deploy the application on AWS using cloud native architecture.

The Four-Step Modernization Journey

Before embarking on the modernization journey, organizations should conduct a comprehensive discovery and assessment phase using tools like IBM ARC, MAM etc. This initial phase combines business and technical drivers to evaluate each application and determine the most appropriate modernization pattern. The assessment helps identify which applications are best suited for reimagining versus other modernization approaches, ensuring alignment with both technical feasibility and business objectives.

The reimagining journey consists of four strategic steps. Initially, the process begins with identifying specific components from the mainframe application suitable for business rule extraction. This identification requires minimal mainframe expert consultation. Once identified, as second step these components serve as inputs to the IBM ARC tool, which extracts the essential business logic. The third step leverages Kiro's spec-based development capabilities, following standard software development lifecycle phases from requirements definition through design. Finally, as fourth step the transformed application is deployed on AWS using cloud-native architecture. To execute this journey effectively, Kiro provides a comprehensive set of features designed specifically for modernization success.

Deployment possibilities

The below diagram shows one of the deployment possibilities keeping the data security and privacy concerns in mind. This considers Kiro IDE as the forward engineering tool.

To ensure security, KIRO and ARC CLI are being hosted in a EC2 running in private subnet, and acceded via a bastion host.

Here are some of the key points:

·      Get the code securely in the AWS EC2 instance running in a private subnet, to be ingested and analyzed by IBM ARC CLI.

·      Access the ARC output in KIRO and run the instructions via modes listed below:

·      Access ARC, KIRO securely and work with the EC2 environment using Bastion Host

Accelerating the Path to Modern Architecture

The combination of IBM ARC and AWS Kiro represents a significant advancement in mainframe modernization capabilities where a deterministic proven tool like ARC brings the context to augment the outcome of agentic ai powered IDE like Kiro.

By leveraging IBM ARC's powerful business rule extraction outcomes as context to Kiro's comprehensive spec-based development approach, organizations can now tackle mainframe modernization with greater confidence and efficiency. This integrated approach not only addresses the traditional challenges of legacy code complexity and skills gaps but also provides a structured, automated pathway to modern cloud-native applications. As enterprises continue their digital transformation journeys, this powerful combination of tools and methodologies enables organizations to modernize their mainframe systems practically and effectively, preserving valuable business logic while embracing modern technology platforms.

Business Value of Combined ARC and Kiro Solution

Contributed by:

1.    Jayanthi Rajaraman (AWS): Senior Solutions Architect, AWS

2.    Diptanu Roy (IBM): Lead Offering Architect, Global AWS Offering (Mainframe Application Modernization)

3.    Girilal Krishnapanickar (IBM): Principal Architect, Asset Engineering (Mainframe Application Modernization)