Choose a migration strategy for each application or component—not once for the whole portfolio. Rehost when a stable, compatible workload needs a low-disruption move; refactor when code changes have a clear business case; replace when a suitable alternative meets the need and its transition costs make sense. If none fits, replatforming, rearchitecting, rebuilding, retaining, or retiring may be better.
The right choice starts with the business reason for change, then weighs the workload’s condition, dependencies, risks, and the value of modernization. The framework below draws on Azure-oriented guidance from Microsoft Learn; implementation details should be checked against your target cloud and workload.
As an Amazon Associate I earn from qualifying purchases.
Start with the reason for changing the application
Before comparing technical approaches, state what the business needs to achieve. A move intended to reduce disruption calls for a different decision than one driven by high maintenance costs, architectural limits, a SaaS alternative, or the fact that the application is no longer needed. Microsoft’s migration strategy guidance maps these kinds of drivers to different workload choices.
Then check whether the proposed change actually addresses the problem. Moving an application does not, by itself, fix poor performance, reliability problems, technical debt, or an architecture that cannot support the intended business goals. Modernization may provide longer-term value, but it also adds work and risk; Microsoft advises grounding decisions in business value rather than adopting new technology for its own sake in Plan your cloud modernization.
#1 Best Overall
Compare the three strategies
| Strategy | What changes | A reasonable fit when | Main caution |
|---|---|---|---|
| Rehost | Move the workload with minimal changes. | The application is stable and compatible, disruption needs to stay low, and near-term modernization is not expected. | Existing performance, reliability, or architectural problems move with it and may require later rework. |
| Refactor | Change existing code while retaining the workload’s functionality. | Technical debt, maintenance costs, or cloud optimization needs justify code-level work, and the team has the time and skills. | Development and testing effort should be justified by expected benefits; unclear value or limited team readiness are reasons to pause. |
| Replace | Adopt a SaaS or other suitable solution instead of continuing to run the custom application. | An alternative covers the required capabilities with little customization, and integration and total cost of ownership support the change. | Account for data migration, user training, process changes, and gaps in features or integrations. |
Rehost: move first, with eyes open
Rehosting is a like-for-like move with minimal changes. It can suit a stable, compatible application when limiting disruption is the priority and there is no near-term plan to modernize. A move can also help a team build cloud operations experience, but it should not be mistaken for remediation: known defects and architectural limits remain unless addressed separately.
Refactor: change code for a defined benefit
Refactoring modifies the application’s code to improve maintainability, performance, or alignment with cloud practices while preserving its functionality. Consider it when technical debt or high maintenance costs are material, or when code changes can deliver needed cloud optimization. Plan for development and testing, and check that skills, time, and resources are available; Microsoft’s modernization planning guidance cautions against modernization without a reasoned benefit.
Rank #2
Replace: fit the work to an alternative
Replacement is not simply a verdict on an application’s age. It depends on whether another solution can meet the business need, how well it integrates with surrounding systems, and whether the full transition is worthwhile. Evaluate data migration, training, process changes, required customizations, and total cost of ownership before deciding.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Consider the choices beyond the title
Rehost, refactor, and replace are not the entire option set. Microsoft’s migration framework also describes these alternatives:
- Replatform: Move components to a managed platform with minimal code changes when reducing operational burden or improving reliability is valuable without full redevelopment.
- Rearchitect: Redesign the architecture when limits on scalability, agility, service orientation, or component-level scaling block business goals.
- Rebuild: Develop a new cloud-native workload when the legacy system is obsolete or modernization is not feasible.
- Retain: Keep a stable, compliant workload in place when it still meets needs and there is no near-term reason to move.
- Retire: Decommission a workload that no longer provides sufficient business value, after checking for critical dependencies.
Use a workload-by-workload decision process
- Define the driver. Record the intended outcome—such as lower operational burden, less disruption, reduced technical debt, different architecture, an available SaaS alternative, or ending support for an unnecessary workload.
- Assess the application’s condition. Review stability, compatibility with the target, performance and reliability issues, maintenance costs, technical debt, and whether the current architecture can meet business goals.
- Map constraints and dependencies. Identify integrations, business criticality, security and compliance requirements, skills, timeline, and available resources. Sequence work according to the needs and priorities of each workload; Microsoft’s migration preparation guidance emphasizes planning for organizational readiness.
- Compare value with effort and risk. Estimate what the change is expected to improve, what it will take to deliver, and what could disrupt users or operations. Do not modernize beyond the benefit the business needs.
- Select and validate. Choose the strategy that fits the workload, then review the assumptions with business and technical stakeholders. Revisit the decision if readiness, dependencies, or future modernization plans change.
Resolve close calls with the right comparison
When two approaches appear plausible, compare them against the same criteria rather than relying on labels such as “legacy” or “cloud-ready.” The Microsoft framework supports assessing business fit, workload stability and compatibility, disruption and migration risk, operational burden, technical debt, architecture limits, expected modernization value, effort, team capacity, dependencies, security, and compliance.
For replacement specifically, add functional fit, integration, total cost of ownership, data migration, training, and process change. A candidate that looks inexpensive to license may still be a poor fit if it requires substantial customization or creates costly operational changes.
A portfolio need not use one strategy throughout. A particular application may contain components with different needs, so one component could be replatformed while another is refactored or retained. Microsoft’s modernization planning guidance recommends matching the approach to each component and combining approaches where appropriate.
Recommended Free Tools
Scope and evidence
The cited Microsoft guidance is Azure-oriented and provides a decision framework, not a cross-cloud comparison or evidence that one strategy produces a particular cost saving, success rate, or delivery timeline. Validate platform-specific requirements against the target cloud and the application’s actual operating context.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
