devopsy.kinsta.cloud

Tag: mainframe development

Transforming the Mainframe Developer Experience With Generative AI: A Day in the Life of a Developer

Step into the world of Alex — a seasoned mainframe developer facing the formidable challenge of navigating through the labyrinth of complex, multiple-module code. Every day, Alex is tasked with deciphering the intricate logic embedded within legacy systems, where decades-old applications coexist with modern demands for agility and innovation.

As Alex settles into her workstation, she is met with the familiar sight of lines upon lines of code stretching endlessly across the screen. Each line represents a piece of the organization’s legacy, a testament to years of development and refinement. But beneath the surface lies a tangled web of dependencies and a legacy burdened by its own complexity.

Today, Alex’s mission is clear — to make changes to a critical module that underpins a key business process. However, the task is far from straightforward. The codebase, a sprawling behemoth of interwoven logic, lacks documentation and comments, making it a daunting challenge for even the most experienced developers. It could take Alex at least a day to understand what she is even looking at, before starting any real work.

Moreover, the organization’s prevailing “fear of change” culture about mainframe code — where management and developers alike are hesitant to make any modifications to the monolithic code out of fear of inadvertently breaking something — looms large over Alex’s efforts. It also stifles innovation and progress, leaving an organization in stagnation.

Generative AI — Transforming the Mainframe DevOps Ecosystem

But Alex doesn’t back down from a challenge. Armed with determination and the latest tools, she sets out to tackle the seemingly insurmountable task. Enter generative artificial intelligence (GenAI), a revolutionary technology that promises to transform how developers like Alex interact with the mainframe applications and improve the developer experience.

GenAI is fast becoming the quintessential solution for augmenting the application developer journey in mainframe DevOps due to its innate capacity to comprehend, guide and enhance code-related tasks. Unlike traditional AI methodologies, which primarily focus on system optimization and anomaly detection, GenAI transcends these boundaries by actively participating in the software development lifecycle.

GenAI is Alex’s guiding light through the maze of complexity. With its advanced natural language processing (NLP) capabilities, it offers unparalleled insights into the inner workings of the codebase, unraveling its mysteries with ease. Gone are the days of painstakingly deciphering undocumented logic; with GenAI, Alex can now navigate the codebase with confidence and clarity.

One of the most powerful features of GenAI is its code explain functionality. A simple right-click lets Alex access detailed explanations of any line of code, revealing its purpose and functionality. It’s a game-changer for Alex, alleviating her fears about making changes to a monolithic codebase and allowing her to get to work right away instead of wasting a day analyzing the code.

GenAI’s impact goes beyond code comprehension, review, and testing. It also plays a crucial role in ensuring coding standards and best practices adherance. With real-time feedback and suggestions, GenAI helps Alex maintain consistency and compliance across the codebase, reducing the risk of introducing errors or vulnerabilities.

GenAI — Instilling Confidence in Both Mainframe Developers and Management

As Alex delves deeper into the codebase, she uncovers opportunities for optimization and improvement, with GenAI, an invaluable ally, providing insights and recommendations that help streamline the development process and enhance the performance of critical business applications.

Perhaps most importantly, GenAI empowers Alex to challenge the prevailing “fear of change” within the organization. With its ability to provide clear and concise explanations of code changes, GenAI gives Alex the confidence she needs to help the organization shift toward a culture of innovation and progress.

As the day draws to a close, Alex reflects on the transformative impact of GenAI on her work. What was once a daunting and arduous task has become a journey of discovery and empowerment. With GenAI by her side, Alex feels emboldened to tackle even the most daunting challenges, knowing she has the tools and support she needs to succeed.

Key Use Cases of GenAI to Improve the Mainframe Developer Experience

1. Explain What the Code Does: GenAI offers in-depth explanations of code snippets and programs. Leveraging sophisticated NLP methods, it seamlessly integrates comments into code, improving readability and fostering collaboration among developers. This automated process enhances understanding and streamlines the development workflow, ensuring that developers can grasp the intricacies of the codebase faster and more efficiently, which helps accelerate the development cycle.
2. Receive Immediate Feedback on Code Standards: GenAI plays a pivotal role in code reviews, assisting developers with real-time feedback, guidance, and remediation suggestions that promote consistency and compliance across applications.
3. Improve the Delivery of Your Changes: GenAI can significantly streamline the code delivery process. By facilitating quicker root cause isolation and enhancing the resiliency and performance of continuous integration and continuous delivery (CI/CD) toolchains, it expedites the mean time to resolution (MTTR).
4. Give Every Developer a Personal Assistant: GenAI functions as an indispensable virtual coding assistant, equipped with extensive knowledge of best practices, design patterns, and syntax. By seamlessly integrating into developer workflows, it provides contextual recommendations and promotes adherence to industry and organizational standards, enhancing developer productivity and code quality.
5. Efficient Testing with Minimal Viable Data: GenAI revolutionizes the testing paradigm by intelligently analyzing source code and devising optimal testing and test data practices.

Paving the Way for a Future of Mainframe Innovation

In conclusion, the integration of GenAI represents a paradigm shift in improving the mainframe developer experience. By providing unparalleled insights into complex monolithic codebases, GenAI empowers developers to navigate the complexities of legacy systems with confidence and clarity. Moreover, by challenging the prevailing culture of “fear of change,” GenAI paves the way for a future of innovation and progress within the organization. As Alex looks toward the future, she does so with optimism and excitement, knowing that GenAI will continue to revolutionize how she works for years to come.

June 21, 2024
IBM LLM AI: COBOL to Java ASAP ¦ ARM IPO is GO!

Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

This week: Translating legacy COBOL code to a slightly more modern language, and Arm will go public (again) next month. (more…)

August 23, 2023
Open Mainframe Project: Zowe Ready for Prime Time

There is a lot of interest in updating mainframe technology/interfaces across traditional enterprises. As development environments and toolsets have evolved outside the mainframe, there is a struggle to keep up—partially because backward compatibility requirements make wild changes difficult and partly because the very architecture of mainframes is different.

There have been several fragmented efforts over the last few years—some of them quite good—to update how a mainframe is accessed from external source and the ways that developers can code for the mainframe. After all, the number of developers proficient in mainframe toolsets continues to decline while the number of workloads does not, so something has to change.

The Open Mainframe Project’s Zowe attempts to address these issues in ways that are both more widely applicable and more technologically advanced. It takes advantage of Open Source REST systems to front-end hardcore mainframe functionality, so developers who are accustomed to application server development can call things such as Spring Boot-based interfaces and use languages including Node.Js to access their mainframe systems.

This week, OMP announced that Zowe 1.0.0 is production-ready. That means serious experimentation can be started at your organization to see if it addresses some or all of your mainframe modernization needs.

There will likely always be some (at least operational) differences in mainframe systems usage because the software infrastructure is designed a certain way, but increasing modern tool usage makes learning those different parts easier across the board. By implementing a vendor-independent architecture that has guidelines for users and vendors to add functionality from anywhere, Zowe is a powerful and expandable tool for getting the job done.

I sound a bit like a fanboy, but for most verticals, this is an issue that will continue to be a pressure point. It’s been more than a decade since companies started bringing back retirees to work on mainframe projects because new developers did not have the skills or interest to complete those projects. As more and more skilled CICS/RPG/PLI/etc. developers and mainframe operations specialists retire, this pressure will increase until it is resolved or unbearable. So I am a fanboy of any company/solution that is addressing this space instead of flocking off to the coolest new DevOps tool du jour. I’ve not yet used Zowe, so can’t directly recommend it, but reading over their documentation, I believe it’s a good idea to use tools we’re familiar with. So I’m rooting for it.

If you test it out, drop a comment and let the rest of us know what you think.

And keep rocking it. Whether in VS-2019 or VT-70, you are making the company engines hum.

— Don Macvittie

February 14, 2019
Culture Comes First in Bringing DevOps to the Mainframe
One of the most difficult challenges in engineering, oddly enough, isn’t technical: the challenge is culture transition management. In my 20-plus years of managing engineering teams worldwide, I’ve led teams through countless organizational adjustments and process improvements toward high-performance and quality, beginning with QA segregation to CMMI to ITIL, AGILE and SAFe.

Today, mainframe organizations have emerged as the focal point for enabling DevOps capabilities and practices. Why the mainframe? Well for starters, business critical applications rely on mainframe data and business components. These applications are mission-essential and that sustained innovation requires solution sets that deliver continuous testing, continuous delivery and continuous feedback on the mainframe platform.

But culture comes first.

Understanding Your Key Players

In conversations with our customers we’ve observed an interesting dynamic emerging within mainframe application development teams. Within these teams are experienced developers who were instrumental in creating the applications that helped make their companies so successful. Over time, many have become set in their ways, adopting a motto of, “If it’s not broken, why fix it?” Such developers are not interested in trying out new things, including programming languages and tools used by the modern-day programmer.

New to the scene are the millennials who have also taken up the mantle of enterprise software engineer. They are a bit more open-minded and bring a different programming language profile, including Java, .NET and C. These developers might be learning how to code APIs with the support of management, and have likely dabbled the basics of how to code iOS apps. Millennial developers are typically much more active online and, as a result, prefer to leverage Twitter and other social channels to stay informed with the development community.

United We Stand

Who, then, should be involved in the journey to realizing DevOps? The answer, of course, is both. Because we’ve heard it reinforced among our users we talk with in the CA Communities every day, at user groups and at prominent mainframe community events such as CA World and SHARE.

You know it, because you probably see it in your organization as well. There are experienced developers: people who have deep knowledge of the applications in place, business intelligence, database tables, cross-application dependencies, batch execution windows and the general experience of how things work on the z/OS platform.

Most importantly, they also know how to avoid breaking what’s working.

And there are the millennial developers: people who are curious to learn about new programming languages, platforms, people who, with the advent of Agile development, and open source have played with things, tested them and are using it to enhance their skill sets.

These people prefer to use modern user interfaces and tools to increase their productivity. Sometimes they might be tentative about making mainframe changes, not willing to risk their careers by breaking poorly documented code.

Culture – Have it Both Ways

While the interests of millennial and experienced developers may at times seem conflicting, both have pivotal roles to play in application development and will benefit greatly from the common sense practices of DevOps and agility that are coming to the mainframe platform. To survive and thrive, teams must unify to facilitate the change process. Most importantly:
- Code must be accessed and modernized at the speed of business. Enterprise developers, testers and architects cannot be afraid to go talk to the “black box” that is the mainframe team.
- Project managers can’t worry that a six-week turnaround for small changes will kill their hopes of meeting their delivery windows.
- The delivery team—QA, release management and operations—can’t afford to sit around waiting for simple deployment steps to happen: when we can automate and integrate to increase deployment frequency.
The promise of DevOps and agility on the mainframe is to fix this, and that’s why our engineering teams are actively building out solution sets that extend DevOps practices to the mainframe, including continuous testing, continuous delivery and continuous feedback.

Curious about what it will take to bring DevOps to the mainframe? Catch some tips and best practices from Forrester Research principal analyst Rob Stroud, or hear from the leaders driving change in their organization in our CA World 2016 Mainframe DevOps panel.

— Aline Gerew
January 31, 2017
Mainframe Apps and the Long View of DevOps Strategy
IT is constantly under intense short-term pressures. Line of business (LOB) executives demand a competitive advantage in digital customer engagement now. Marketing leaders demand better customer analytics now. Legal demands better audit reporting now.

In fact, much of the investment that companies are making in Agile, DevOps and continuous delivery is due to the urgent, short-term demands of the business.

But short-term actions alone are insufficient for business success. IT leaders must also strategically navigate the vagaries of the digital economy over the long term—or, as Wayne Gretzky’s father taught him, to “skate to where the puck is going to be.” Otherwise, today’s small wins inevitably will lead to tomorrow’s egregious failures and inefficiencies.

And integration of your mainframe apps and data into your enterprise DevOps tool chain is central to that long-term digital strategy.

Here’s why.

The Mainframe: Big, Nimble Iron

A DevOps tool chain is only as fast as its slowest link. For most large enterprises, that slow link is the mainframe.

Slow, waterfall-bound and underautomated mainframe dev/test processes are non-trivial—because mainframe applications and data are non-trivial. They typically are an organization’s most important apps and data. So if you can’t modify your mainframe application logic quickly or leverage your mainframe data in new ways easily, you’re uncompetitive toast in today’s fast-moving markets.

Also, mainframe applications and data typically serve as the back-end systems of record for front-end mobile and web-based systems of engagement. Your ability to innovate and remain competitive when it comes to digital engagement and digital perceived value is, therefore, almost universally contingent on lessons learned from frequent and timely mainframe code drops.

In other words, everything IT is under pressure to deliver—new mobile apps, better analytics, fulfillment of increasingly stringent compliance mandates, etc.—depends on bringing Agile, DevOps and (ultimately) continuous delivery to the mainframe.

Just about everyone realizes this. Sure, there are some holdouts who still advocate something they call “bimodal IT”—which prescribes allowing the mainframe to languish in obsolete waterfall and dev/test practices while the rest of the enterprise moves forward. But bimodal lT has proven to be a mistake. An Oct. 31, 2016, Forrester Research report titled, “Predictions 2017: CIOs Push For Speed Amid Volatility,” shines a light on the fatal realities of bimodal IT:

“In early 2016, many CIOs fell for the false promise of a bimodal strategy. They only pushed part of their team to adopt an accelerated pace — and left others unchanged. But those CIOs are already experiencing the shortcomings of operating at two speeds. Business peer frustration, polarizing and deteriorating cultures, and unsustainable operational complexities will continue, ultimately spurring most of these CIOs to course correct in 2017.”

Given the explosion of mainframe-related innovation that has taken place in the past couple of years, enterprise IT leaders and industry analysts agree that “mainstreaming” it into enterprise DevOps is the right and necessary strategy (see as examples here and here).

Some Good Things Take Time

While instant gratification may be the order of the day—and gain without pain may be an all-too-frequent vendor promise—the reality is that great IT achievements all require effort, investment and time.

Inclusion of the mainframe into your cross-platform DevOps tool chain is no exception. No CIO can achieve mainframe agility and cross-platform DevOps orchestration by simply waving a magic wand. But there is a clear, practical path to these goals for those with the requisite foresight and leadership.

That path includes:
- Modernizing the developer workspace. Most mainframe development is still performed in “green screen” TSO/ISPF environments that require specialized knowledge and limit staff productivity. A modernized, graphical workspace as part of a preferred DevOps tool chain will allow staff at all experience levels to perform dev/test tasks on both mainframe and non-mainframe applications.
- Graphical visibility into existing code and data structures. Mainframe applications typically are large, complex and poorly documented. These attributes are a major impediment to mainframe transformation. The right tools enable new participants in mainframe DevOps to quickly and easily “read” existing application logic, program interdependencies and data relationships.
- QA “shift left.” Early detection and correction of quality issues is a central principle of modern programming—and it’s particularly important on the mainframe, where there’s an especially low tolerance for error.
- Initial Agile training and adoption. Once you put the right development environment in place, you can start shifting your process from a traditional waterfall model with large sets of requirements and long project timelines to a more incremental model that allows teams to quickly collaborate on user “stories” and “epics.”
- Use operational data throughout the dev/test life cycle. By using operational data in the dev/test life cycle, your DevOps team can ensure that applications perform optimally in production. They also can significantly reduce your MIPS/MSU-related costs by avoiding excessive CPU consumption caused by inefficient code.
- Agile-enable source code management (SCM) functions. Traditional mainframe SCM environments are inherently designed for waterfall development and thus are incapable of supporting Agile requirements such as parallel development on user stories. The shift to Agile-enabled SCM is, therefore, absolutely essential for mainframe transformation.
- Automated release management. To keep pace with today’s fast-moving markets, you must be able to quickly get new code into production. That means automating and coordinating deployment of all related development artifacts across all targeted environments.
- Cross-platform DevOps. Ultimately, you need to fully de-silo your enterprise environment so the mainframe becomes “just another platform”—albeit an especially scalable, reliable, high-performing, cost-efficient and secure one—that can be modified quickly as needed to meet the relentlessly evolving needs of the business.
‘What’s My Motivation?’

With as much as IT departments have on their plates, it’s easy to continue putting off mainframe DevOps integration. But, ultimately, there are only two alternatives to doing so:
- Re-platform your mainframe applications. This alternative has proven to be a costly and extremely risky undertaking that delivers zero benefits while it consumes resources and potentially disrupts IT operations, or
- Continue doing nothing. This alternative puts your company at further competitive disadvantage and hastens the day when you’ll experience total mainframe paralysis due to the loss by attrition of all your veteran mainframe developers, QA specialists and operations staff.
Viewed in this context, genuine investment in the mainframe—whether you call it modernization, mainstreaming or “Millennializing”—is the only reasonable course of action. Such investment starts improving your digital competitiveness immediately, while paving the way for a future where you can make the most of all your resources and confidently transfer platform responsibilities to your next generation of IT professionals. Including the mainframe in your broader DevOps efforts (while extremely valuable) does require commitment, and there are resources available to help. Just be wary of those who may talk a good game, but whose actions don’t really convey a strong commitment to users’ best interests. Fortunately, there are truly dedicated experts who can help every step of the way, and have direct experience modernizing on platform within their own organizations.

It’s the smart, viable long-term DevOps strategy every CIO of a mainframe-enabled enterprise needs to adopt ASAP.

About the Author / Christopher O’Malley

Christopher O’Malley is CEO of Compuware. He has nearly 30 years of IT experience with past positions including CEO of VelociData, CEO of Nimsoft, EVP of CA’s Cloud Products & Solutions and EVP/GM of CA’s Mainframe business unit, where he led the successful transformation of that division. Connect with him on LinkedIn and Twitter.
November 21, 2016