A contractor supporting a critical mainframe modernization project for a state government agency, faced a significant challenge. Tasked with migrating a legacy COBOL system, the contractor’s initial pilot migration attempt using AWS Blu Age could not fully meet customer requirements. The tool could only output Java and not meet the new requirement for Python or C#. Furthermore, the migration lacked traceability, making it impossible to validate. This contractor engaged Karsun Solutions, which leveraged its AI-powered ReDuX modernization platform. Using Agentic AI capabilities powered by Amazon Bedrock, ReDuX automatically analyzed the legacy system and generated the required C# programs, Summer Batch Core workflows, and MS SQL Server tables, delivering the 100% traceability needed to successfully complete the pilot.
About the Customer
The customer was one of three vendors awarded a contract for a cloud services project for a state government agency. This project’s goal is to modernize a core business system by converting legacy mainframe code to Python or C# while moving to a cloud-based platform.
Customer Challenge
A state government agency initiated a high-priority project to convert its legacy system to a modern, cloud-based platform. The state’s primary goals were to enhance security, improve flexibility, ensure data redundancy, and lower long-term maintenance costs.
The project was structured in two phases: a competitive part one pilot awarded to three bidders, followed by a part two award for the full system conversion given to a single winner from the pilot. The customer team was tasked with migrating a subsystem for the pilot, as it was the least entangled with other subsystems.
The customer’s initial modernization attempt hit two critical roadblocks. First, the agency’s requirements evolved. After initially preferring Java, the state later specified a preference for Python or C#. This immediately put the team’s approach in jeopardy, as AWS Blu Age can only create code in Java. Second, they were unable to show file-level COBOL program traceability to modern programs. This made validation and future maintenance impossible in the mind of the customer. Failure to solve these challenges would result in a failed pilot and disqualification from the lucrative part two of the contract.
The Partner Solution
The pilot contract was explicitly for cloud services. The state agency sought the enhanced security, flexibility, and redundancy that a cloud-native solution provides. The successful solution embraced this by using a suite of AWS-native services. Additionally, Karsun used its AI-powered modernization platform, ReDuX, to solve complex traceability problems in legacy systems and could output modern C# code, directly meeting the new requirements.
The ReDuX platform rapidly analyzed the system and generated a fully modernized, traceable equivalent.
- Agentic Analysis: First, Karsun used its ReDuX Blueprint tool. This tool, which is powered by Amazon Bedrock, applied GenAI models to analyze the system, including COBOL programs, JCL, and DB2 database tables. This process created a comprehensive and fully documented map of the legacy system’s architecture and business logic.
- Automated Transformation: With this detailed blueprint, Karsun’s ReDuX Code Companion used agents to automatically transform the legacy application. The modernized solution consisted of:
- Agentic Analysis: First, Karsun used its ReDuX Blueprint tool. This tool, which is powered by Amazon Bedrock, applied GenAI models to analyze the system, including all COBOL programs, JCL, and DB2 database tables. This process created a comprehensive and fully documented map of the legacy system’s architecture and business logic.
- Automated Transformation: With this detailed blueprint, ReDuX Code Companion used agents to automatically transform the legacy application. The new, modernized solution consisted of:
- Modern C# Application: C# programs were generated for every COBOL program. Verifiable, one-to-one mapping solved the traceability problem. This new application is hosted on .NET Core containers over the AWS EKS service or on EC2 instances.
- Cloud-Native Batch Processing: Summer Batch Core workflows were generated for every JCL, converting all mainframe batch processes into cloud-native, serverless workflows.
- File Storage: Application files are now stored in AWS S3.
- Angular Front-End: A new user interface in Angular precisely mimicks the flow of the original mainframe green screens, such as the menu and work order creation screens.
- MS SQL Server Tables: Every legacy DB2 database table was migrated to a new table in MS SQL Server.
Results and Benefits
Karsun’s AI-powered ReDuX platform successfully delivered a modernized pilot application that met all of the state’s requirements, rectifying the failures of the initial attempt. The primary benefit was the achievement of 100% traceability for the entire application stack.
The key success metric was the one-to-one mapping of legacy assets to modern, cloud-native code, all made possible by the GenAI analysis from AWS Bedrock:
- Every COBOL program was converted to a corresponding C# program, now running in .NET Core containers on AWS EKS or EC2.
- Every JCL was converted to a Summer Batch Core workflow.
- Every DB2 table was migrated to an MS SQL Server table.
This provided the customer with a complete, verifiable, and maintainable modern application, allowing them to successfully complete the pilot.
About Karsun Solutions
Karsun Solutions modernizes enterprise systems enabling agencies to make the next technological advancement their next opportunity to elevate mission capability. Solutions are tailored to meet agencies’ unique needs and optimize operations. These solutions adapt and stay relevant to current trends while using secure, digital architecture built to last. It is a proven modernization partner whose expertise elevates agency capabilities and ensures every next opportunity is within reach.