Your next hardware refresh is approaching, or rising cloud bills and licence costs are putting pressure on your budget. Dependence on one provider may be limiting your options, while regulators and clients want clarity on data location and legal access. Your leadership team also needs to demonstrate that risks are being managed. Could a migration give you greater control, and which systems would benefit from moving? This guide to cloud migration strategy helps you work through those questions.

The guide follows the cloud migration process from the first business questions to reviewing the results. Along the way, we explain the choices that shape the plan and link to detailed articles where you need a closer look.

Whether you are leaving on-premises infrastructure or reconsidering an existing cloud environment, the starting point is the same: establish what the move needs to achieve. In practice, planning is iterative: findings from assessments and pilot migrations may change earlier assumptions about the target environment or migration approach.

What a cloud migration strategy needs to address

A cloud migration strategy defines which workloads should move, where they should run, and how much they need to change. It explains how those choices support the business case and what must be in place before migration begins. A cloud migration roadmap then turns the decisions into a sequence of work with clear responsibilities and acceptance criteria. These choices form part of your wider cloud strategy, which also covers how cloud adoption supports future business needs.

Cloud capabilities and business value

For an application with variable demand, assess whether the target platform’s scaling options suit the workload. These may include autoscaling for virtual machines or containerised workloads, as well as serverless services. Weigh the expected business value against implementation costs and provider dependencies.

Read more: Our cloud computing guide explains the deployment and service models behind these decisions >>

Business goals and readiness for migrating to the cloud

Before comparing cloud platforms, establish what needs to improve. An expiring hosting contract makes timing a priority; slow releases or unpredictable cloud costs raise different questions. Choose a measure that reflects the problem, such as environment provisioning time or cost per service, and record its current value so you can compare results after migration.

That baseline helps you compare credible alternatives, including keeping suitable systems where they are. The business case needs to include migration costs, training, and the period when both environments run. Compare total cost of ownership (TCO) over the same period for each option. Include data storage and transfer charges alongside computing power, licences, support, and the people needed to operate the environment.

Next, check whether your existing applications support the proposed plan. Mapping applications and infrastructure, their dependencies, data volumes, licences, and resource use may reveal work missing from the original estimate. Legacy systems may need changes before they can run in the target cloud environment. Existing virtual machines can simplify parts of the move, but do not establish whether an application will suit the target cloud.

Read more: Understand the difference between virtualisation and cloud computing >>

Application owners need to assess the impact on business operations, while finance, security, and operations need to understand the proposed change. Involve them alongside a business sponsor so it is clear who can approve the plan, including any cost increases or changes to service availability.

Check whether the team has the skills and capacity to deliver cloud migration projects while supporting existing services. Cloud adoption may require training or specialist support – plan this early and prepare affected users for changes to their applications or workflows.

If this assessment leaves important questions unanswered, we can help resolve them through migration and discovery with NoviArc – Spyrosoft Cloud Foundry. AI-assisted analysis, reviewed by our specialists, supports the assessment of your environment and the migration plan.

Choosing the right cloud provider and deployment model

The assessment may show that some legacy systems should stay where they are while others move. Your cloud computing environment needs to accommodate those choices and the business needs behind them:

  • Public cloud services are offered for general use on infrastructure operated by a cloud service provider.
  • Private cloud provides cloud infrastructure for one organisation’s exclusive use, either on its own premises or hosted elsewhere.
  • Hybrid cloud connects distinct environments, typically a public cloud with a private cloud solution or on-premises infrastructure.
  • Multicloud uses cloud services from more than one provider.

These arrangements can overlap: a hybrid environment can also use several cloud providers. The benefit needs to justify the additional integration and management work.

Alongside the choice of cloud environment, decide how much of the software stack your team should manage:

  • Infrastructure as a service (IaaS): the provider runs the underlying cloud infrastructure, while your team manages operating systems and applications.
  • Platform as a service (PaaS): the provider manages the underlying platform, such as the operating system and runtime, while your team manages its code and data.
  • Software as a service (SaaS): the provider operates a finished application that your organisation configures and uses.

Check the responsibility split for each service. Even with SaaS, your organisation retains responsibilities for its data, users, and access settings.

The provider shortlist should follow from these decisions. Public cloud platforms include AWS, Microsoft Azure, Google Cloud Platform, and IBM Cloud. Compare their services and latency, as well as data handling requirements, support, and security assurances. Check how you could leave a cloud vendor later. A managed service may save enough work to justify relying on that cloud provider, as long as you understand what switching would require.

Read more: Compare public, private, hybrid, and multicloud environments against your business requirements >>

Types of cloud migration: the 7 Rs

Even within the same cloud environment, existing applications can need very different treatment. The widely used 7 Rs framework helps you select an appropriate migration strategy for each workload:

  • Rehost (lift and shift): move an application with minimal changes to its code or architecture.
  • Relocate: move workloads at platform or hypervisor level to a compatible environment without redesigning the applications.
  • Replatform: make targeted changes to optimise an application for the cloud while keeping its core architecture, for example by moving a database to a managed cloud service.
  • Refactor or re-architect: change the application’s code or architecture to take advantage of cloud capabilities.
  • Repurchase: replace the application with another product, often a SaaS service.
  • Retain: keep a workload in its current environment, with a reason and a review point.
  • Retire: decommission a workload that no longer provides sufficient value.

Organisations can combine migration strategies across their applications and refine the target design as they learn more. Separate changes needed for compatibility from improvements that can wait, then assign responsibility and budget for the work planned for later.

Types of cloud migration: choosing an approach for each workload

Learn more

When lift and shift fits the plan

A hosting deadline often makes rehosting attractive. Limited code or architectural changes mean less development work before migrating to the cloud. For a stable application, this may be enough; for another, it may be the first stage of a longer modernisation plan. The same limited scope explains why rehosting can disappoint. An oversized server may remain oversized, while restrictive licences or tightly coupled components can make cloud-based infrastructure expensive to run.

Assess these costs and dependencies to decide whether a faster move is worth the potential expense and constraints afterwards. Security checks and cloud migration testing remain necessary even when code barely changes.

Identify which changes are needed to achieve the expected improvements in scalability or release speed, and decide whether to make them during migration or afterwards. Some improvements may come from infrastructure and delivery automation; others may require replatforming or refactoring. Reflect that decision in the business case and cloud migration plan, including when the expected benefits can be realised.

Assess when lift and shift pays off

Learn more

Data security and other challenges of cloud migration

The approach chosen for each application also changes the risks. Rehosting can reduce the cloud migration effort while carrying existing inefficiencies into the cloud; redesigning an application adds testing and schedule uncertainty. Before the migration process begins, agree who is responsible for the risks and what they will do if problems arise.

A risk assessment should cover service disruption, data integrity, security and compliance, unexpected costs, and supplier dependency. Recording the likelihood and impact of each concern helps the team prioritise safeguards and decide which unresolved cloud migration challenges should delay a workload’s move. Classifying sensitive data and confirming your team’s and cloud provider’s security responsibilities helps establish the required controls. Check encryption and privileged access management against the regulatory requirements that apply to your workloads.

Use the risk assessment to plan the checks and recovery arrangements for each cloud migration wave. Agree how application behaviour and data integrity will be tested, how service would be restored if problems arise, and who can authorise or stop cutover. Recovery planning must account for any data written after the switch, so that returning to the previous environment does not introduce a new loss.

Give cloud spending a clear owner and track it against the forecast, including the period when both environments run. Reviewing risks and costs after each wave helps control costs and lets the team adjust the remaining work before small deviations become larger problems.

Explore key migration risks and ways to mitigate them

Learn more

Planning the cloud migration process from on-premises infrastructure

The cloud migration roadmap can now reflect your starting point. For an on-premises migration, dependencies help determine which systems move together. Migration plans must also reflect the needs of critical systems and agreed change windows. For a data centre migration, also plan when the old infrastructure can be decommissioned. Contract expiry dates set deadlines, but the plan must allow enough time to prepare and test the systems included in each wave.

Before production workloads arrive, prepare the cloud infrastructure, often organised as a landing zone, with access controls, networking, logging, and recovery arrangements. The source and target environments then guide the choice of cloud migration tools for discovery, replication, infrastructure automation, and testing. Where practical, use infrastructure as code and automated checks to make environments and migration steps repeatable.

Plan data management and transfer alongside the application move. Check whether data volumes and available connectivity fit the cloud migration window. Where supported, ongoing replication can keep the target up to date while the source remains in use, helping reduce the risk of data loss. Plan a final synchronisation and data consistency checks before cutover.

Choose a pilot workload that limits business risk while representing important technical characteristics of the migrations planned for later waves. Comparing results with the baseline helps validate cloud migration tools and refine estimates before later waves. Each wave then follows an agreed transfer and validation process, with cutover and recovery matched to the application’s dependencies and data.

Plan the move from on-premises to cloud

Learn more

Cloud-to-cloud migration and portable cloud architecture

For workloads already in the cloud, pay particular attention to dependencies on provider-specific services, such as managed databases, identity services, or serverless functions. Alongside the application and infrastructure dependencies assessed in any migration, establish which services have compatible alternatives and which substitutions require changes to the applications that use them.

Mapping current services and their dependencies to proposed replacements shows what your engineers will need to change. It also reveals work that a comparison of service prices alone would miss. An open-source component may improve portability while transferring maintenance responsibilities to your team or partner, and the business case needs to include that work.

Review contracts and data export arrangements for both providers to understand exit constraints and identify ways to limit future vendor lock-in.

Cloud-to-cloud migration: from a hyperscaler to EU infrastructure

Learn more

Cloud migration decisions in practice

Our work on TakTo’s migration to the AWS Cloud shows how these decisions work in practice. The financial services company needed greater infrastructure flexibility while keeping its customer application available. That requirement shaped both our choice of lift and shift and the way we divided the migration process into stages.

Our team started with smaller, non-customer-facing environments and tested them before moving the larger databases. The cloud migration was completed without affecting customer activity.

Read more: How TakTo moved to AWS while maintaining continuity for its customers >>

Cloud operations and measurable cloud benefits

After cutover, service owners need to support the new cloud environment. Handover is easier when responsibilities and skills have been addressed during planning, including who maintains the platform and applications. Agree how incidents and changes will be handled, and which duties remain with an external partner.

During stabilisation, confirm that monitoring and support arrangements work in practice. The previous environment can be retired once acceptance and recovery requirements are met, with records and backups preserved according to policy. Systems that remain in a transitional state should retain an owner and visible costs.

A successful cloud migration also needs to deliver the improvements set out in the business case. The measures established before migrating to the cloud give the organisation a consistent way to assess progress:

  • Cost: actual operational costs against the forecast, with changes in demand accounted for.
  • Delivery: time to provision environments or release application changes.
  • Service quality: response times and error rates under representative load.
  • Recovery: results of disaster recovery tests against the agreed recovery time objective (RTO) and recovery point objective (RPO), covering downtime and acceptable data loss.

Underused cloud resources may need right-sizing; an architectural bottleneck may require modernisation. Use these results to decide which postponed improvements are still needed to achieve the original cloud migration goals and where further cloud investment is justified.

From strategy to an actionable cloud migration plan

What you learn about existing systems helps you choose the target environment, while the migration strategies selected for your applications determine the work and risks involved. A successful cloud migration strategy explains why these choices were made, what must be checked before systems move, and how the organisation will measure the results.

At Spyrosoft, we help turn these decisions into a migration plan through discovery with NoviArc – Spyrosoft Cloud Foundry. We use AI-assisted analysis to map your environment and its dependencies, with outputs reviewed and approved by our architects and security engineers. Working with your team, we identify portable alternatives to existing services and use the findings to shape the target architecture and phased migration plan. Find out more about NoviArc migration and discovery to see how it could help with your cloud migration project.

A cloud migration strategy defines which workloads should move, where they should run, and how much they need to change. It connects these decisions to business goals, costs, and acceptable risk. A cloud migration roadmap then turns the strategy into a sequence of work with clear responsibilities and acceptance criteria.

At Spyrosoft, we recommend connecting the business case with an assessment of applications, dependencies, and operational requirements. A cloud migration strategy should define the target environment and migration approaches for individual workloads, alongside risk controls and delivery priorities. It should also establish responsibilities after the move and measures of success.

The 7 Rs are rehost, relocate, replatform, refactor or re-architect, repurchase, retain, and retire. These migration strategies describe different approaches to individual workloads, from moving applications with minimal changes to replacing or decommissioning them. Most portfolios combine several approaches because applications have different requirements and business value.

Lift and shift can suit stable applications that need to move quickly, particularly when a hosting contract or hardware lifecycle creates a deadline. It limits changes to application code and architecture, but can carry existing inefficiencies into the cloud. Its suitability depends on compatibility, licensing, and expected operating costs, as well as the timing of future modernisation.

An on-premises migration starts with systems hosted in your own facilities or a colocation environment. Cloud-to-cloud migration moves workloads between cloud environments, where they may already depend on provider-specific managed services or APIs. Both require an assessment of compatibility and application dependencies, but replacing those provider-specific services can add significant work to a cloud-to-cloud move.

It makes risks and responsibilities visible before execution. Assessing dependencies, sensitive data, and potential disruption to business operations helps the team decide which safeguards and checks are needed before systems move. Confirm applicable regulatory requirements, and use phased delivery and testing to inform cutover decisions. Recovery planning must also account for data written after the switch.

No. The business case needs to account for cloud migration effort and ongoing operations, including licences, support, and the period when both environments run. Carrying oversized resources or unsuitable architecture into the cloud can undermine expected savings. Ongoing cost management includes comparing actual spending with the forecast and adjusting resources where needed.

Compare results with the business goals and baseline established before migration. Relevant measures may include operating costs, environment provisioning time, release speed, and service performance. Recovery tests should also demonstrate whether agreed downtime and data loss objectives can be met. Use the results to decide what to optimise next and whether postponed modernisation is still needed. This review also helps keep wider cloud adoption plans grounded in measured results.