The Business Case for Mainframe DevOps

Mainframe applications and databases are tried and tested, and operate at enormous scale and efficiency. At the same time, they can bring huge costs of maintenance and use and create challenges for enterprise organizations looking to leverage existing systems and data for new applications.

With the ability to deploy Linux and containerization solutions within a logical partition (LPAR) on a mainframe, the door is now open to build mainframe-hybrid applications, and port existing DevOps tooling into the mainframe environment. This makes it possible to leverage CI/CD practices to create new applications leveraging mainframe business logic and data, and the modernization of legacy mainframe code.

The result: Organizations can utilize new technologies and approaches without incurring the cost of revamping existing applications from scratch. By using old languages alongside new, and utilizing API/microservices-based systems, large organizations can continue to use old apps while building new apps on mainframe infrastructure.

Figure 1: Mainframe-Based Containerized Solution Architecture

Businesses that adopt mainframe DevOps experienced a 20% increase in developer productivity and realized 20% lower TCO for application deployments compared to the cost of increasing midrange server capacity to leverage data being extracted from the mainframe environment.

The efficiency savings enabled by CI/CD principles are significant, with testing proven to be up to 10x faster than with legacy mainframe processes. Code quality can increase 50% over three years while downtime is reduced by 70%. Other advantages include improved customer experience, future proofing of company platforms, and reduction of day-to-day costs (such as for data extraction).

Any organization with a sufficiently large application infrastructure built on mainframe technology will face a key challenge: how to leverage the value of the data in ways that fit modern expectations for flexibility, speed, and process.

Transactional-based processing scenarios are a powerful case in point. Finance institutions that depend on COBOL-based processes have been able to adopt applications such as fraud detection, currency adjustment, and loan processing, thanks to the drastically improved performance and lower costs enabled by agile approaches and container-based systems.

By taking advantage of agile techniques and utilizing existing code, data, and assets, organizations can directly interact with mainframe systems without confronting infrastructure bottlenecks and security risks that have made mainframe integration and modernization unviable in the past.

The alternative to adopting mainframe DevOps is to continue working with old, outdated languages supported by a shrinking pool of developers and using processes that don’t allow for swift deployment or for modern customer expectations (for new applications, services, and experiences) to be met.

The status quo makes agile development almost impossible, and customer experience will continue to worsen as competitors shift to mainframe DevOps.

Moving to mainframe DevOps processes presents a number of cultural challenges and risks. Working with existing development teams to incorporate new working practices must be a core area of focus in your change process.

Older developers who have built a career on legacy languages could see the shift to DevOps processes as “designing themselves out of a job.” It’s key to bring this group along for the ride, as their knowledge will be invaluable during the transition and beyond.

Don’t underestimate the paradigm shift developers face in moving from batch processing to transactional processing. Building these new techniques into established thinking can be a real challenge.

Consider how DevOps is evolving, especially in terms of measuring the generation of value and operational effectiveness. Other disciplines, such as value stream management, FinOps, and DevSecOps will all have a bearing on the approach that you take to mainframe modernization. The end goal is to align mainframe thinking and culture with the practices and tooling of DevOps as the latter matures.

• 30 days: Evaluate the potential avenues for legacy code modernization, identifying key areas in which a modern application with full access to the mainframe data could drive business value. Based on these focus areas, begin to identify tooling options.
• 60 days: Assemble a team, under the direction of the mainframe business process owners, including skilled DevOps as well as experienced mainframe personnel. Put together a roadmap for how your new code will hook into existing mainframe processes.
• 90 days: Deliver the initial proof of concept, prove value, and adjust the roadmap in order to leverage maximum value based on discoveries.

• Achieving the Strategic Vision of API-Based Innovation – Gigaom
• Key Criteria for Evaluating Financial Operations (FinOps) Tools