COBOL applications are often complex, extending 30 million to 100 million lines and running mission-critical workloads.
Finding COBOL programmers today is very difficult.
IBM designed the new watsonx Code Assistant for Z to help businesses refactor COBOL code into modern Java.
Despite its ancient provenance, the COBOL programming language remains critical to big business. However, maintaining COBOL is hard, as programmers skilled the aging language are retired or ... dead — and many younger programmers haven’t learned it at all.*
Nevertheless, IBM on Tuesday announced generative AI tools to help address that problem, modernizing antique mainframe software for the cloud by refactoring COBOL code to Java.
“Finding COBOL programmers is difficult, as only old gits like me have experience, and it’s very rusty these days,” said Omdia analyst Roy Illsley.
"If you can find a COBOL programmer, they are expensive. They command some of the highest salaries because so many mission critical apps are written in COBOL, and they need maintenance," he said.
The watsonx Code Assistant for Z, scheduled for release in the fourth quarter, is designed to help businesses refactor COBOL code into modern Java, which can be maintained by developers that entered the workforce after the disco era. The new code resulting from the refactoring can continue running on the IBM Z mainframe for performance, security and resiliency.
Modernizing mainframe apps will ease integrating those applications into hybrid cloud architectures, leveraging disciplines such as AIOps, DevOps, APIs and data integration, IBM said.
“By applying generative AI with watsonx to modernize the architecture of these COBOL applications and selectively translate COBOL to Java, organizations can tackle the talent gap and take advantage of the broad Java developer ecosystem,” Skyla Loomis, VP, IBM Z Software, said at a presentation for press and analysts last week.
A streamlined COBOL alternative
IBM is presenting the watsonx Code Assistant for Z as an alternative to other tools that convert COBOL to Java while preserving native COBOL syntax and architecture, resulting in code that’s difficult for modern developers to maintain.
Code generated with assistance from IBM’s watsonx tool results in native Java syntax and architecture that’s more easily maintained using DevOps practices and more readily interoperable with cloud applications, IBM said.
Refactoring COBOL to Java is a difficult process that can take decades and often fails. IBM expects the AI tool will speed the process by an order of magnitude. Although that’s still a long time, organizations will get incremental value while the process is underway, enabling innovation “immediately and continuously while the application is being modernized,” IBM claimed.
COBOL applications are often complex, extending 30 million to 100 million lines and running mission-critical workloads. “Being able to modernize these large, complex enterprise applications materially and confidently in a small number of years is groundbreaking and is simply not possible today,” an IBM spokesperson said in an email statement.
Over half of Fortune 1000 businesses rely on mainframes to process credit card and payment transactions, airline bookings and ensure that mission-critical data is always available, IBM said.
Mainframe applications are central to business and technology strategies, according to 70% of respondents to a survey by the IBM Institute for Business Value in conjunction with Oxford Economics. Approximately 68% said mainframe systems are central to their hybrid cloud, but nearly 70% of executives said mainframe-based applications need to be modernized.
Aging COBOL systems have become a big problem for business and government. During the early days of the pandemic, COBOL-based state unemployment systems were straining under the load of increased applications. That generated business for “COBOL Cowboys” (motto: “Not our first rodeo”), a Texas consulting organization comprised of retired COBOL programmers who ride to the rescue in mainframe crises. CEO and founder Bill Hinshaw’s LinkedIn profile shows work experience dating back to 1963.
Businesses love their mainframes
Nearly 80% of survey respondents plan to increase their IBM Z budget as a percentage of their total IT spend, IDC analyst Ashish Nadkarni said during the IBM presentation.
“By modernizing in place, people can make the most of their investments,” Nadkarni said.
IBM watsonx Code Assistant for Z is part of IBM’s revitalized AI strategy, announced earlier this year, to integrate the emerging technology with IBM’s traditional strengths, delivering multi-cloud enterprise solutions and services tailored to big business needs.
That strategy is built on foundation models for various business functions: fm.code for generating code through natural language to improve developer productivity and enhance IT automation; fm.NLP, a set of LLMs for specific industries, which can be quickly customized for client data; and fm.geospatial, built on climate data and sensing to help companies plan for natural disasters, sustainability and other geophysical processes that could affect business.
*Ed. Note: “younger” here means “pretty much everybody still in their normal working age” as COBOL was already uncool when Whitney Houston was warbling “I Will Always Love You.”