Many businesses continue to rely on legacy applications running on mainframes because those systems are stable, proven, and continue to deliver measurable ROI. However, as organisations shift more workloads to cloud platforms, they must develop a clear strategy for which mainframe applications to migrate, how to migrate them, and how to preserve the business logic that underpins their competitive advantage.
The right approach lets a business modernise its mainframe estate and benefit from cloud economics—elasticity, pay-as-you-go pricing, and modern tooling—while retaining the intellectual property embedded in decades of proven application code.
Core Requirements for Any Mainframe-to-Cloud Solution
Before evaluating specific migration paths, organisations should establish non-negotiable baseline requirements. A viable cloud solution for mainframe applications must be:
- Cost-effective — ideally shifting capital expenditure (CapEx) to predictable operational expenditure (OpEx).
- Secure — meeting or exceeding the security posture of the existing mainframe environment.
- Reliable — preserving the high availability and resilience that mainframe users typically expect.
- Modernisation-ready — enabling the business to adopt current and future development practices, tools, and languages.
The Three Mainframe Migration Strategies
When planning a mainframe-to-cloud migration, there are three primary options. Each carries a different risk profile, cost, and level of disruption to existing business logic.
1. Rewrite
The existing application functionality is recreated from scratch in a language more typical of cloud-native development (Java, Python, Go, etc.). This addresses concerns around the shrinking pool of COBOL and assembler developers, but it comes at a significant cost. A rewrite requires substantial investment in engineering time without necessarily adding new features or functionality. The risk of regression—failing to replicate complex, battle-tested business logic faithfully—is high. Rewrites frequently overrun budgets and timelines.
2. Replace
Organisations can implement a Software-as-a-Service (SaaS) solution to replace the current system entirely. This creates an opportunity to review and modernise business processes simultaneously. However, for all but the most commodity workloads, off-the-shelf SaaS products rarely replicate the depth of functionality built up over years in a custom mainframe application. Replacing a highly customised system often means accepting reduced capability or surrendering the competitive differentiation that the original application provided.
3. Rehost (Lift and Shift)
Rehosting moves mainframe applications to a cloud platform with minimal or no code changes. Specialised platforms can emulate the mainframe runtime environment—including COBOL execution, JCL job scheduling, VSAM data access, and CICS transaction processing—on standard cloud infrastructure. This approach preserves existing competitive advantage and lets the business continue to run proven code while incrementally modernising over time. The main challenge is organisational and political: mainframe decisions often involve entrenched stakeholders, vendor relationships, and legacy procurement contracts.
Combining Strategies: Replace Plus Rehost
For most organisations, a hybrid approach—replacing commodity components with SaaS where appropriate, and rehosting the core business-logic applications unchanged—delivers the greatest practical benefit. This allows the business to deploy modernised application components to the cloud or an alternative platform based on business and operational requirements, without a high-risk big-bang rewrite.
Benefits of Moving Mainframe Workloads to the Cloud
When executed correctly, migrating mainframe applications to a cloud environment can unlock a broad set of operational and financial benefits:
- Reduced risk — Retaining existing intellectual property and competitive advantage keeps change incremental and risk manageable. There is no need to recreate decades of proven business logic.
- Flexibility — Cloud platforms offer a wide range of virtual environments, and resources can be provisioned or decommissioned on demand rather than according to hardware procurement cycles.
- Faster time to market — Moving test and development environments to the cloud allows teams to provision mainframe-compatible instances in minutes rather than days. This enables genuine mainframe DevOps: faster iteration between development, test, and production teams.
- Cost optimisation — Shifting development, test, and production workloads to the cloud reduces dependency on proprietary mainframe hardware and software licensing. Costs become OpEx rather than CapEx, improving budget predictability.
- Reliability — Major cloud providers offer highly resilient infrastructure with automated failover, geographic redundancy, and strong SLA commitments.
- Scalability — Cloud environments support automatic horizontal and vertical scaling and load balancing, enabling workloads to grow without manual infrastructure intervention or performance degradation.
- High performance — Cloud platforms offer multiple instance types optimised for different compute-to-I/O ratios, allowing organisations to match infrastructure to workload characteristics.
- Security — Rehosted applications can be integrated into a standard cloud security stack, including identity and access management, encryption at rest and in transit, audit logging, and multi-factor authentication—in some cases improving on the security controls available in legacy mainframe environments.
Practical Considerations: Mapping Mainframe Services to the Cloud
A successful migration requires a solution that maps core mainframe services—transaction monitors, job schedulers, data access methods, and security subsystems—to their cloud equivalents, while allowing the application code itself to run unchanged. Getting this mapping right is what makes rehosting viable and what distinguishes it from a full rewrite. Organisations that achieve this can unlock significant savings through reduced operating costs, improved developer productivity, and better system performance.
Moving to a platform such as Microsoft Azure or a comparable cloud provider also frees businesses from the constraints of proprietary mainframe hardware refresh cycles, giving them faster access to infrastructure, greater scalability, higher availability, and improved business continuity—all hosted in highly secure, certified data centres.
Case Study: Migrating COBOL Payroll Processing to the Cloud
A German public-sector IT services organisation provides payroll processing to its clients, handling salary payments for approximately 420,000 individuals every month. The payroll system relied heavily on COBOL for core business logic, running on mainframes. Constantly changing salary rules, tax rates, and regulatory requirements meant the team needed to update software frequently, but the mainframe environment slowed development cycles.
To respond more quickly to client requests and adopt modern development tooling, the organisation chose to migrate its COBOL applications from the mainframe to a cloud environment using a rehosting approach. By moving the COBOL code without rewriting it, the team could migrate the most critical business logic incrementally and with low risk—one component at a time.
The migration was completed quickly, giving the business a stable, future-proof platform. Development teams could now iterate faster using modern IDEs and CI/CD pipelines while the proven COBOL logic continued to execute correctly. As the organisation’s public-sector clients continue to adopt digital workflows, the modernised platform allows the team to deliver high-quality, reliable services while protecting the integrity of critical HR and payroll processes.
Summary: Choosing the Right Path
There is no single correct answer for mainframe modernisation. Rewriting is high-risk and expensive. Replacing with SaaS may sacrifice competitive differentiation. Rehosting preserves proven code but requires organisational buy-in and a platform that genuinely supports mainframe runtime semantics. For most enterprises, a pragmatic combination—rehosting core applications and replacing commodity functions—provides the best balance of speed, risk, and long-term flexibility. The goal is not to abandon what works, but to give proven applications a modern, scalable foundation from which the business can continue to evolve.