Accenture Cloud First Leader Andy Tay discusses common cloud adoption challenges and how to overcome them
By Jennifer Goforth Gregory
By John Morell
When Dr. Grace Hopper helped give the world COBOL in 1959, the legendary computer science pioneer likely didn’t realize the computing language would still be used in mission-critical applications 65 years later. While a testament to the language’s utility, this staying power also underscores the need for modernization in IT environments. This is where tools like watsonx Code Assistant for Z can prove their worth.
COBOL is used by countless large companies worldwide in industries such as banking, insurance, airlines, government and more. Its long lifespan shows that the “common business-oriented language” has been a success in a variety of operating systems, from UNIX and Linux to Windows and z/OS.
Reveal to learn more ...
45 of the world’s top 50 banks use IBM Z for tasks like processing ATM withdrawals and credit card transactions.
Eight of the world’s top 10 insurers use IBM Z for tasks such as handling sensitive data and billing information.
Four of the top five airlines use IBM Z to administer tasks such as ticket purchases, seat assignments and flight changes.
Many large federal agencies use IBM Z to manage veterans’ benefits, social security payments, tax returns, federal employee paychecks, trade revenue and more.
Today, however, operators are looking to modernize their mainframes and migrate, sometimes fully, into the cloud. On top of that, COBOL-fluency among programmers is dropping as workers retire. This creates a need to transform COBOL business services into high-quality Java code that virtually any developer can use.
Generative artificial intelligence (genAI) is one of the solutions being used to easily create Java code. This technology is the basis for IBM’s watsonx Code Assistant for Z, which brings speed, automation and accuracy to the task of assessing, updating, validating and testing code. It also helps reduce developer workloads and aids organizations in efficiently modernizing large applications.
SPONSORED CONTENT
The tool uses AI to extract business rules and documentation, allowing you to target high-value refactoring opportunities. Your applications will operate on the cutting edge and streamline your developer resources. Save costs by running workloads on Ziip processors. Automate at scale and maintain your application within standard architectures. Just plug it into your standard DevOps pipelines and go. Your future developers will get up to speed quickly with minimal end-user training.
Java is an open-source system that runs on any platform. You can take Java and easily run it on the mainframe.
—Piyush Mishra, Associate Partner, IBM
When searching for information about converting COBOL into a modern language, many people have an obvious first question: Why Java over other languages? “Java is an open-source system that runs on any platform,” IBM Associate Partner Piyush Mishra said last month at the IBM Z Day virtual conference. “You can take Java and easily run it on the mainframe. You can use Java classes and objects and move from the transform phase to the AI-assist phase, where you’re feeding the COBOL code, and then you get high-quality Java code.”
With the coding assistant, Mishra noted, a project that used to take a year can now be completed in three to four months. “To give you an idea of the speed and productivity, I’ve seen 30,000 COBOL lines of code converted into Java in less than a week,” he elaborated.
To illustrate, Mishra pointed to a COBOL-to-Java project demo in which one team of developers used watsonx Code Assistant for Z while another team worked without the tool. “In just two days the developer-only team processed about 12 Java services from the COBOL application. Then we tried doing the same work using the GenAI tool. In just two hours the code assistant processed seven Java services,” Mishra said.
“Watsonx Code Assistant analyzes the complexities and plots similar functionalities into one service, which creates a better output. There’s nothing like it to accelerate productivity and quality.”
GenAI code conversion is more than a simple line-by-line translation. Watsonx Code Assistant for Z analyzes the application logic to produce high-quality Java code. This means the tool dramatically speeds up the COBOL conversion into Java without sacrificing the quality of the code output.
“We’ve fed the Java code produced by watsonx into SonarQube, the industry standard for code analysis. The results left the Java team happy about the code quality watsonx produced. It was much more manageable, which made future changes easier,” Mishra said.
He often hears a preconceived notion about mainframes lacking the agility to handle Java. “GenAI,” he said, “is helping challenge that myth with all of the new tools coming out under the umbrella of watsonx.”
Better-quality code results in Java classes and methods that are more structured, with improved variable naming and overall code readability. This makes modernized applications easier to maintain and enhance.
Because of COBOL’s longevity, many of the developers who have become experts in the language are reaching retirement age. “Leadership in many organizations see this as a real problem,” Mishra said. “‘Who will manage our mainframe workloads when the experts are gone?’ The answer is a solution like watsonx Code Assistant for Z, which quickly takes the COBOL code and converts it to Java while retaining all of the business logic and rules.”
The difficulty finding COBOL-fluent developers is often one of the arguments used to replace mainframe hardware. Using GenAI for code conversion helps mainframe users retain their legacy hardware, as younger developers use AI tools to help manage existing COBOL mainframe applications.
‘Who will manage our mainframe workloads when the experts are gone?’ The answer is a solution like watsonx Code Assistant for Z, which quickly takes the COBOL code and converts it to Java while retaining all of the business logic and rules.”
—Piyush Mishra, Associate Partner, IBM
Organizations recognize the need to modernize their applications, but they also need the flexibility to do so. One advantage of GenAI solutions is they allow users to modernize in place, negating the need to migrate from a mainframe.
Users can also decide which workloads to keep on their mainframes, while shifting others to the cloud. “I think that's the beauty of watsonx Code Assistant for Z,” Andy Bradfield, vice president of IBM Z Hybrid Cloud, said at IBM Z Day. “It gives clients the capability to meet their modernization needs for all or part of their journey. It's not an all or nothing approach, it's a hybrid.”
Users decide what stays on the mainframe and what goes to the cloud. “I think that's the beauty of watsonx Code Assistant for Z.
—Andy Bradfield, Vice President, IBM Z Hybrid Cloud
Even the most skilled developers make mistakes, which is something that can be avoided when using a GenAI code assistant. Automating code conversion improves code quality, reducing workloads and the risks associated with problem code.
Solutions like watsonx Code Assistant for Z examine the existing code and recommend improvements or optimizations that make the app more efficient and lower operating costs. They also automate common coding tasks, which reduces the workload on development teams, allowing them to focus on higher-value projects in addition to reducing risk for error.
Overall, the impact of IBM watsonx Code Assistant for Z and GenAI for code modernization is a game changer for IBM Z mainframe customers. It presents an end-to-end solution, facilitating application discovery, analysis, automated code refactoring and selective COBOL-to-Java conversion. Importantly, it shows that the pursuit of modernization doesn’t have to mean replacing legacy mainframe systems.