In today’s fast-changing digital world, modernizing mission-critical applications on IBM Z is essential for enterprise agility. Adopting hybrid cloud and DevSecOps requires a smooth, low-risk approach to updating legacy systems. watsonx Code Assistant for Z and IBM Test Accelerator for Z help meet this need through AI-powered automation and deep integration with IBM z/OS.
watsonx Code Assistant for Z uses AI to understand, refactor, and transform z/OS workloads. IBM Test Accelerator for Z supports automated, shift-left testing in flexible z/OS environments. Together, they provide a smart, unified path to speed up mainframe modernization, boost developer productivity, and shorten time-to-value within a secure enterprise pipeline.
IBM Z DevOps strategy
IBM’s Z DevOps strategy delivers a consistent and platform-agnostic experience for developers. It helps teams work across mainframe and non-mainframe environments with ease.
Key focus areas
-
Generative AI (gen AI) and automation first: Uses Granite LLM and automated testing to reduce complexity and risk throughout the development lifecycle.
-
Toolchain consistency: Supports popular tools such as Git, Jenkins, and VS Code to ensure smooth transitions across environments.
-
Integrated CI/CD pipelines: Promotes unified, Git-based pipelines across platforms to streamline delivery.
-
Shift-left testing: Introduces testing early in the development cycle to improve quality and catch issues sooner.
-
Continuous feedback loop: Encourages regular improvements in processes, skills, and product quality.
-
Secure hybrid cloud readiness: Enables stable, secure, and agile delivery across hybrid cloud setups.
Role of watsonx Code Assistant for Z and Test Accelerator for Z in the IBM Z DevOps pipeline
watsonx Code Assistant for Z and Test Accelerator for Z play key roles across different stages of IBM Z DevOps pipeline:
-
Analyze: watsonx Code Assistant for Z analyzes applications and provides deep understanding through its Understandcapability. This gives teams a clear view of the application structure.
-
Code: During development, watsonx Code Assistant for Z supports refactoring, optimization, and transformation of z/OS languages such as COBOL and PL/I, directly within popular IDEs.
-
Build: Modular code changes can be built consistently by using script or configuration based framework and running unit test cases.
-
Test: Test Accelerator for Z generates automated test cases and enhances the enterprise test suite by supporting unit, integration, and regression testing. It enables early testing by creating isolated z/OS environments that can run unit and early integration tests without requiring middleware.
Overview of watsonx Code Assistant for Z
IBM watsonx Code Assistant for Z uses generative AI through Granite large language model, to support developers, system programmers, and architects throughout the application lifecycle. It helps reduce the cost and complexity of modernization, makes it easier to onboard new talent, and speeds up delivery.
Key phases of watsonx Code Assistant for Z
-
Understand: Analyzes and documents the application landscape, creating a clear blueprint of how components are connected.
-
Refactor: Breaks down COBOL or PL/I applications into smaller, modular business services.
-
Optimize: Identifies performance issues in COBOL programs and suggests improvements.
Many clients and partners choose to stop after the preceding phases. However, if needed, the service can go further by transforming the application to Java for use on z/OS.
-
Transform: Uses generative AI to convert refactored and optimized COBOL code into object-oriented Java.
-
Validate: Checks that the transformed Java code matches the original COBOL code in function and behavior.
-
Explain: Uses AI to explain COBOL, PL/I, REXX, or JCL steps in plain language. These explanations can be added as comments or used for documentation.
Overview of IBM Test Accelerator for Z
IBM Test Accelerator for Z is an automated testing platform that is designed for z/OS workloads. It helps teams test earlier, faster, and more efficiently.
-
Testing framework: Supports different types of testing, from unit tests to full integration tests.
-
Isolated environments: Creates flexible, virtual z/OS environments for safe and repeatable testing.
-
Early testing support: Enables testing early in the development process to catch issues sooner.
-
Automated test data: Generates test data automatically to reduce manual work.
-
Code coverage and defect detection: Helps improve quality by measuring test coverage and identifying issues.
What sets watsonx Code Assistant for Z and Test Accelerator for Z apart
-
AI-powered features: Built on IBM’s Granite LLM for smart code understanding and automation.
-
Smooth code transformation: Supports COBOL to Java conversion with AI-based validation.
-
No middleware Needed: Makes testing setup easier and faster.
-
Developer-friendly tools: Works well with popular DevOps tools and workflows.
-
Scales for Enterprise use: Designed to handle large and complex environments.
How watsonx Code Assistant for Z and Test Accelerator for Z work together
watsonx Code Assistant for Z helps refactor and transform legacy code using AI. Test Accelerator for Z supports testing of the updated or transformed code and compares results with expected outcomes.
Example use case
User Story
As a new developer, I want to find and understand the business logic across multiple COBOL programs, so I can extract it and create a new, easier-to-maintain service.
Synergy use case 1: Discovering business logic with test recording
Understanding business logic that is spread across multiple COBOL programs can be difficult, especially for new developers. In this example, a senior developer uses Test Accelerator for Z integration test builder to record a key business transaction. The recording acts as a guide, showing which COBOL programs are involved.
Outcome:
The new developer quickly sees how the programs are connected, without needing to manually trace the entire application. This makes it easier to begin extracting the logic.
Synergy use case 2: Modularizing with AI and testing support
After identifying the right logic, the developer pulls code from different COBOL programs to create a new modular service. Further watsonx Code Assistant for Z code optimization advice can be used to improve performance and maintainability.
After the service is refactored, Test Accelerator for Z helps generate and run unit and integration tests to check whether it works correctly within the full application.
Outcome: The developer completes the modularization using AI suggestions and automated testing, ensuring the new service works well and does not introduce errors.
Synergy use case 3: Transforming to Java and validating with tests
Based on customer-specific needs and fit-for-purpose assessment, the new COBOL service can be transformed into Java using AI features in watsonx Code Assistant for Z. The developer then reuses the earlier Test Accelerator for Z integration tests and runs them on the new Java code in z/OS.
By comparing test results before and after the transformation, they confirm that the behavior remains the same.
Outcome:
The transformed Java service functions as expected within the COBOL application, ensuring seamless COBOL–Java interoperability on z/OS platform and stability through automated regression testing.
Conclusion
Modernizing mainframe applications is easier with tools such as watsonx Code Assistant for Z and Test Accelerator for Z. They help teams transform code and run reliable tests, making the process faster, safer, and more efficient.
References