Taking a comprehensive perspective to mainframe application modernization with a disposition strategy

Introduction

Mainframe customers face a myriad of options to modernize. Organizations now face urgent imperatives to modernize due to talent shortages, high and often rising costs, and restrictions on business agility imposed by the legacy environment. Customers also find themselves navigating a multitude of modernization patterns, tools, and strategies.

The disposition strategy includes guiding principles that help you address complex mainframe monolithic applications and large code bases. Customers decompose legacy applications into manageable chunks, develop a target migration pattern, sequence the migrations, and build integrations to remaining mainframe workloads. This is iterated on until the application is fully decoupled.

A disposition strategy is a multi-pattern approach to mainframe modernization. The right pattern is selected for the right workload, based on prioritization of business and IT objectives and workloads characteristics. Rather than view mainframe modernization as the collection of individual projects and disjointed migration phases, the disposition strategy advocates for the creation of a comprehensive perspective. This includes a roadmap that defines the entire journey from mainframe to cloud native from the beginning. This approach serves to accelerate migrations, reduce risk, and help customers achieve business and technology objectives in acceptable timeframes.

Define a North Star

Mainframe estates are varied across geographies, industries, and customers, with no two mainframe applications truly being alike. There is great diversity in legacy technologies, business objectives, future state requirements, and risk appetites across customers. Many large enterprises have multiple lines of business supported by applications running on the mainframe. These business lines also have a wide number of business leaders, application owners, and stakeholders across the organization dependent on mainframe processing. This dynamic often results in a scenario where organizations lack a clear strategy for the future state of the mainframe. We often see customers early in their modernization journey who have a mix of strategies actively being implemented or planned by different mainframe stakeholders. This results in a disjointed approach to modernization. The modernization may not have a north star vision articulated that guides the organization on their modernization journey.

The first step for any successful mainframe modernization program is to define a north star. This north star is shared by the C-Suite and often the Board of Directors of an organization. Customers recognize the growing risks, costs, and competitive disadvantages associated with remaining on the mainframe. With executive leadership guidance for modernizing legacy applications, we see customers operate with greater speed, urgency, and successful delivery. Without a clear mission, we see that customers engage in a series of disjointed, tactical modernization programs. Disjointed programs may succeed in moving workloads of the mainframe platform yet struggle to deliver the full benefits from the modernization. In some cases, this may even result in MIPS usage increasing due to constraints put on the mainframe. To avoid this situation, we advise customers to define a north star by answering three primary questions:

Why are we modernizing?
Where are we going?
When will it happen?

Answering these questions sets the foundation for a successful mainframe migration program and helps the organization define common business and technical goals shared by the organization.

Business Criteria for Modernization: Balancing business requirements with technical realities

Business objectives can vary dramatically across divisions within organizations. Some business units have urgent requirements to change the functionality of their legacy applications, while others resist modification to established processes or user experiences. Broadly, we see priorities from the business falling into two categories:

Category 1: Business functionality remains the same, but technology modernization is essential for business agility.

No desire for functional change. In this scenario, the business users are quite happy with the existing business functions supported by the legacy applications. The business functions and workflows that exist today do not need to be altered, and the business users are opposed to the idea of functional changes. This can even be true for customers who have users interacting with greenscreens. Users that have been using greenscreens may be operating at such a high efficiency level that replacing this with a modern UX could result in productivity declines.

Customers should generally expect that the majority of their mainframe workloads will fall into this category. Mainframe applications have persisted in organizations for decades. These applications may be well-suited to the business with years of custom business logic. They could be tailor-made for large enterprises and their differentiated way of doing business in their respective industries. Not every application requires functional transformation. For systems where stability and predictability are paramount, it is important to consider:

Refactor options that convert the legacy applications to a modern programming language and relational database. For example, using AWS Transform for mainframe, customers can refactor COBOL applications to Java on AWS using specialized AI agents from AWS.
Re-platforming options that maintain existing functionality and preserve legacy technology stacks while moving to more modern infrastructure. For example, AWS Mainframe Modernization capabilities provide options to re-platform to mainframe compatible runtimes in the cloud.

For these applications, focus on communicating the non-functional benefits of modernization: resiliency, reduced impact radius, availability, agility.

Category 2: Business functionality requires changes to remove technical debt, add new features, and decompose monoliths for product alignment.

Requirements for functional enhancements. This is where the business requires functional enhancements to be made to the applications. Generally, customers expect only a portion of their mainframe workloads to fall into this category. In this situation, the business may have desires for modern UIs, real time capabilities, or faster batch processing. Customers may also have goals to break legacy monoliths into product aligned business functions. This approach creates a microservices architecture for loose coupling, which facilitates agility and innovation.

Customers are increasingly met with growing end-customer expectations for near real-time capabilities. These are challenging to introduce to mainframe monoliths. Additionally, we meet with customers who have ambitions to expand into new markets, industries, or otherwise grow their customer base. Often, customers with aggressive growth objectives find that maintaining mainframe applications hinders their ability to grow and win new business.

Growth enablement: Which applications, if modernized, would unlock new revenue streams or support business expansion?
Customer experience impact: Which applications directly affect customer interactions and satisfaction?
Market responsiveness: Which systems currently limit your ability to respond to market changes?
Innovation potential: Which applications would benefit most from modern development practices and integration with cutting-edge technologies?

Business units with clearly articulated functional requirements for enhancement should typically receive higher prioritization. Their specific needs, whether for new features, improved user experiences, or integration capabilities, provide concrete objectives that can drive modernization efforts and demonstrate tangible value.

Reimagine pattern is defined as the rewriting and rearchitecting of a mainframe application to a modern technology stack. The goal of this modernization is to introduce functional changes to the application. When the business requires new functionality, the reimagine pattern is the preferred approach.

Technical Criteria for Modernization

The disposition strategy for mainframe modernization should also incorporate technical criteria evaluated at both the organizational and application levels.

Evaluating Technical considerations at the organization level: Data center exit strategy or urgent migration deadlines

Organizations facing mandates for a data center exit have urgent migration deadlines that must prioritize speed and risk mitigation. The rehost pattern, meaning moving mainframe workloads to a partner-managed data center, with no code changes, might be a pragmatic first step. The rehost approach can often be completed in months, bypassing the lengthy cycles of alternative migration patterns. Rehosting can provide business continuity. It also lays the foundation for future modernization and helps organizations incrementally adopt more advanced patterns. These patterns can address modernization needs like refactor or reimagine within time and resource constraints.

Niche Mainframe technologies: programming language, transactional monitor, database

Specific programming languages, transactional monitors, and database technologies used by existing mainframe applications can have a significant impact on the feasibility and complexity of modernization efforts. This is an important technical consideration when evaluating a mainframe modernization disposition strategy.

Certain mainframe technologies, like Natural/Adabas, IDMS, and others may not be directly supported or fully supported by some modernization patterns like re-platforming or refactoring. The availability and maintainability of skills for these legacy mainframe technologies is also a factor. This can limit the modernization options available, and patterns like replace or reimagine might be the only viable choices.

Organizations must carefully assess the mainframe technology stack being used and how that aligns with the feasibility and complexity of different modernization approaches. This technical assessment is a key input into determining the right disposition strategy.

Vendor renewal timelines

Vendor renewal timelines can be an important technical consideration when evaluating a mainframe modernization disposition strategy. Organizations often have various mainframe vendors with different software licensing contracts. Renewal timelines raise risks when commercial terms are being evaluated. In this case, some customers determine the best value outcome includes an exit from those vendor technologies as quickly as possible. This timeline velocity can also influence the modernization approach to choose.

For example, if there is an urgent deadline to exit a license agreement, a refactoring pattern may be more suitable than a replacement or reimagine approach. Refactoring helps the organization to modernize the application while preserving the core functionality, which can often be done more quickly than a full rewrite or reimplementation.

However, it’s important to note that not all refactoring solutions support all mainframe technologies. An assessment of the right refactoring solution for the specific technologies in use must be completed. In some cases, there may not be an obvious or proven refactoring solution available. A reimagination or replacement approach may be the only viable option to exit the vendor technology by the required deadline.

It is important to evaluate mainframe vendor contracts and renewal timelines as part of the overall technical assessment when determining the best modernization strategy. This helps align the selected approach with the urgency to exit particular vendor technologies.

Available mainframe skillsets

When companies face a resource talent risk due to constrained mainframe skillsets, it is important to choose mainframe modernization options that have less reliance on those skills. In such cases, strategies like refactoring and reimagining the mainframe applications can be effective approaches.

Conversely, when companies have a capable mainframe talent pool within their organization, a re-platforming approach can be a suitable strategy for modernization. They can use existing expertise to migrate the workloads to a more modern platform.

Evaluating Technical Complexity and Dependencies of the application/workload

The selection of an appropriate pattern should be based on both business considerations and technical requirements. It should consider the specific characteristics of each workload or application.

It is critical to conduct a thorough technical assessment of the various applications and workloads to determine the best modernization approach for each one. During this assessment phase, consider the following factors:

Source technologies: Evaluate the programming languages and the volume of the existing source code. Some languages and frameworks are more amenable to automated transformation and modernization than others. This can impact the feasibility and complexity of certain modernization patterns.
Data considerations: Assess the data store technologies used on the mainframe (Db2, IMS DB, VSAM, etc.). Evaluate the volumes of data, the complexity of the data structures, and the relationships between data entities. The nature and complexity of the data can influence the appropriate modernization approach.
Degree of coupling: Identify the level of coupling between the different applications and workloads. For example, an application that includes transactional context propagation likely has tight coupling. In this case, it would pose more modernization challenges than those with loose coupling or clear service boundaries. This is because the interdependencies of the tightly coupled functions must be addressed in order and be specifically managed in the modernization journey.
Integration complexity and dependencies: Assess the various integration points between the applications and workloads. Identify shared resources, data dependencies, and the overall complexity of the integration landscape. This can help determine the appropriate modernization pattern that can preserve the existing integrations or help to provide a transition with less risk.
External interfaces: Depending on the chosen modernization pattern, some client applications running outside of the mainframe but accessing it through external interfaces may also be changed. Verify that the selected pattern supports the required interfaces for all external connection points, API operations, and data exchange mechanisms with external systems.

This detailed technical assessment should consider factors such as:

Grouping applications that access the same data in read/write mode together and choosing the same patterns for those groups.
Selecting the same pattern for workloads with a high degree of coupling
Considering the impact of the source programming languages on the feasibility of different patterns
Selecting a pattern that will minimize changes to external interfaces and integrations where possible.

An application and workload analysis is a key input into the overall disposition strategy. It can add the right modernization patterns and solutions for workloads based on their unique technical characteristics and dependencies.

Strategy Development

Building a Business-Outcome-Driven Program

Rather than treating modernization as a purely technical exercise, develop a program that:

Works backwards from the organizations north star: As mentioned at the start, customers need an organizational strategy and approach for their mainframe estate. Successful mainframe migration projects operate within the parameters set by the company’s leadership. Why are they modernizing, where are they going, and when will they achieve it.
Aligns with strategic business objectives: Modernization should support specific business outcomes such as increased agility, improved customer experience, or new capabilities.
Considers the entire portfolio from the start: Even if modernization is defined as a phased approach, planning should cover the complete application landscape to avoid creating new technology silos.
Balances tactical wins with strategic goals: Design the program to deliver incremental value while working toward comprehensive modernization.
Establishes clear metrics for success: Define how you’ll measure progress in both business and technical dimensions.

Without a strategy set at the C-suite level, individual teams may adopt divergent approaches or fall into a “wait and see” mindset. This could delay modernization and create additional complexity.

Avoid the “rebuild everything” pitfall

Our experience shows that the 80/20 principle generally applies to mainframe estates: roughly ~80% of the mainframe applications do not require functional changes, and ~20% of the applications must be reimagined.

We advise customers to consider modernization approaches that include substantial mainframe exits. Customers like Transamerica and Goldman Sachs have successfully used refactor and re-platform patterns to move mission-critical mainframe workloads to AWS. Taking individual application-by-application approaches may be too slow to meet business imperatives. Consider incorporating multiple modernization patterns, based on business and technological objectives.

Refactoring at scale: AWS Transform for mainframe provides refactor capabilities that can help you to modernize the legacy application to modern, Java frameworks. This pattern can be used when you want to reduce dependency on legacy technologies while benefiting from the accelerated migration timeline offered by deterministic tooling.
Re-platforming: Re-platforming uses emulation technology to achieve a like-for-like migration of mainframe applications. This is often referred to as a “COBOL to COBOL migration”. In this case, the re-platforming pattern can address situations with COBOL talent shortages, and accelerate mainframe exits.

Mass modernization approaches combined with strategic reimagining provide customers the opportunity to align technical and business outcomes while still driving towards a platform exit. Customers who consider multiple patterns in their strategy are able to address more varied objectives within their organization. This is done while delivering operational cost reductions goals within the same time period.

Conclusion: The Time for Action is Now

Today, the imperative for mainframe application modernization is strong. Beyond the commonly cited challenges of talent shortages, increasing software costs, and organizational inefficiencies, a new driver has emerged: the growing impact of generative AI on software development. As generative AI coding assistants revolutionize productivity for modern languages, the development speed and productivity gap between modern and mainframe technologies will compound. Organizations with applications in COBOL, Assembler, or PL/1 languages face a growing competitive disadvantage regarding the speed to value. Their peers may operate core systems written in modern technologies that can move at ever increasing development speeds.

There is no “silver bullet” for mainframe modernization. Success requires a business-driven, multi-pattern approach that aligns IT and business objectives around concrete outcomes. By using automation and iterating incrementally, you can focus on the value beyond cost savings.

The disposition strategy provides a framework for this journey, one that recognizes the nuances of each application portfolio. By modernizing mainframe applications with this approach, organizations can preserve the valuable business logic built over decades while positioning themselves for future innovation demands.

Migration & Modernization