Prepare and test your skills

Prepare and test your skills

Rehosting is a migration path that moves a specialized legacy application to cloud infrastructure with little or no change to its code or architecture. The business drivers for rehosting are speed, reduced operational burden, cost control, compliance preservation, and avoiding the risk of a rewrite. Google Cloud supports rehosting through Google Cloud VMware Engine (opens in a new tab) for VMware workloads and Google Compute Engine for standard virtual machines (VMs).
Rehosting is most viable when an application is tightly coupled to its current platform, must meet strict compliance constraints, or needs immediate cloud scalability and reliability without redesign. Rehosting preserves critical workloads with minimal disruption because the operating system, dependencies, and structure stay the same. Other migration paths may fit better when a workload no longer satisfies requirements, can be replaced by a similar application, depends on end-of-life third-party technology, or carries uneconomical license fees.
The business case for rehosting combines faster moves, less management work, and better control over cost and compliance:
These drivers matter most when refactoring before migration is not feasible or not worth the cost.
Google Cloud VMware Engine is a private cloud service for migrating VMware VMs to Google Cloud with minimal change. VMware Engine is a strong fit for specialized legacy applications that depend on VMware features such as vSphere, vCenter, HCX, and NSX-T. Migration is planned in waves, with application dependencies and mappings factored in, and VMware HCX is used for migration with bulk migration preferred.
Before migration, teams assess the source environment, check version compatibility for vSphere, vCenter, HCX, and NSX-T, and identify VMs with resource needs that exceed the current node type. Content that cannot be migrated, such as mounted ISOs, NSX-T tags, passthrough devices, multi-writer disks, and physical RDMs, needs pre- and post-migration strategies. For example, physical RDMs can be converted to virtual compatibility mode.
VMware Engine offers configuration choices that affect cost and reliability. CPU core overcommit ratio ranges from 1.0 to 8.0 with a default of 4. Memory overcommit ratio options are 1.0, 1.25, 1.5, 1.75, and 2.0 with a default of 1. Hyperthreading weight defaults to 25%, and HCX mesh is enabled by default. NSX-T Edge size Large is recommended for most production deployments, while Extra Large is for the most demanding throughput and latency. Storage optimization recommends Compression for most use cases, while Deduplication and compression may save more but has performance implications. vSAN storage policies define failures to tolerate (FTT) and RAID; higher FTT requires more nodes per cluster, from a minimum of 3 nodes for FTT=1, RAID=1 up to 7 nodes for FTT=3, RAID=1. VMware Engine Protected adds backup and disaster recovery but increases total cost of ownership (TCO) and may not be available in all regions.
Google Compute Engine is the target for rehosting VMs that do not need a VMware-specific environment. Compute Engine offers two tenancy modes that affect isolation and licensing. Multi-tenant is the default, where VMs share a physical server with VMs from other projects and only pay-as-you-go (PAYG) licenses are supported for Windows Server. Sole-tenant gives VMs exclusive access to a dedicated physical server, supports both bring-your-own-license (BYOL) and PAYG for Windows Server, and lets teams configure Simultaneous Multithreading (SMT) to balance performance and licensing costs.
Choose sole-tenant when security or compliance requires physical separation or when Windows BYOL needs per-core or per-processor licenses. For sizing, Migration Center offers three strategies that affect the TCO report and workload cost. No rightsizing matches the current machine size, Moderate is the default and balances performance and cost, and Aggressive applies aggressive rightsizing based on current performance.
Reliability and scalability for rehosted workloads come from adding cloud services around the application without changing its code. Load balancing routes traffic to available resources and prevents individual resources from being overloaded. Health checks ensure load balancers send traffic only to responsive backends; if backends are managed instance groups (MIGs), autohealing deletes unavailable VMs and creates new ones. Rate limiting protects applications from sudden or sustained load spikes, such as distributed denial-of-service (DDoS) attacks, and can also limit billing through project-level quotas.
Network Service Tiers let teams optimize connectivity for cost and performance. Premium Tier is for applications that serve users globally and have backends in more than one region. Premium Tier routes traffic through the Google network from the closest point of presence (PoP) to the resource and outbound to the closest PoP to the destination.
The two main Google Cloud rehosting targets differ in how much of the existing platform they preserve:
A decision tree for rehosting a legacy workload on Google Cloud: it first checks whether the app can run as a standard VM or is tightly coupled to VMware, then branches on compliance and Windows licensing needs to pick VMware Engine, sole-tenant Compute Engine, or multi-tenant Compute Engine.
| Target | What it preserves | Choose when |
|---|---|---|
| Google Cloud VMware Engine | VMware environment, vSphere, vCenter, HCX, NSX-T, vSAN policies | The application is tightly coupled to VMware and needs minimal change |
| Google Compute Engine | Individual VMs, with multi-tenant or sole-tenant tenancy | The application can run on standard VMs and does not need VMware-specific features |
Exam tip: Rehosting keeps the application largely unchanged; workloads that no longer satisfy requirements, can be replaced by similar applications, or depend on end-of-life technology may be better candidates for other migration paths.
Rehosting, also called lift and shift, moves a workload to the cloud as-is or with minimal changes to its code or configuration. The operational benefit of rehosting specialized legacy applications is that it reduces data-center overhead, simplifies infrastructure management, and preserves existing configurations while the organization prepares for later modernization. Rehosting is often the first step in a phased migration because it transfers server management from on-premises resources to cloud infrastructure.
Rehosting, also called lift and shift, is the lowest-change migration path for workloads that must keep running during the move. Rehosting is best for workloads that have a relatively small number of dependencies, are not worth refactoring or cannot be refactored before migration, or are based on third-party software. Rehosting is different from refactoring, which changes the workload to take advantage of cloud capabilities, and from rebuilding, which replaces the application.
The modernization path can be grouped into three ordered phases:
Rehosting optimizes applications from an infrastructure perspective first. After applications run in Google Cloud, it becomes easier to integrate cloud services and further optimize them. Rehosted applications can still communicate with other environments over a hybrid network connection.
Rehosting reduces data-center overhead by moving workloads from an environment where the organization owns physical infrastructure to a public cloud where the provider manages more of the stack. In an on-premises environment, the organization is responsible for cooling, physical security, hardware maintenance, power, network cabling, and the virtualization platform. In a public cloud environment such as Google Cloud, the service provider handles physical security and safety, power and network cabling, hardware and maintenance, and the virtualization platform.
| Resource | On-premises environment | Public cloud environment |
|---|---|---|
| Physical security and safety | You | Service provider |
| Power and network cabling | You | Service provider |
| Hardware, including maintenance | You | Service provider |
| Virtualization platform | You | Service provider |
| Application resources | You | You, eventually leveraging fully managed services |
When calculating total cost of ownership (TCO), include hardware and software plus operational costs for power, cooling, maintenance, and support services. Google Cloud elastic scalability can make costs easier to reduce than in a rigid on-premises data center. One commonly overlooked cost is cloud networking: in a data center, routers and switches are often a one-time cost, while in the cloud, network utilization can be billed, so data-intensive workloads may need new architectures and network flows.
Google Cloud compute services simplify infrastructure management by letting teams run existing workloads on managed cloud infrastructure instead of maintaining physical servers. A lift and shift of Windows-based .NET applications typically starts by moving on-premises Windows Server and SQL Server instances to Compute Engine (opens in a new tab) VM instances. This move can be done manually or automated with a migration tool.
Manual migration can use Compute Engine Windows Server images, SQL Server images, or ready-to-deploy Google Cloud Marketplace solutions such as the ASP.NET Framework solution. Tool-based migration can use Migrate to Virtual Machines or VMware Engine to move on-premises VMware VMs to Google Cloud. After VMs are configured, teams can create custom VM images to recreate instances on demand. If domain services are needed, teams can deploy a fault-tolerant Microsoft Active Directory environment on Compute Engine with a Virtual Private Cloud (VPC), or use Managed Service for Microsoft Active Directory. For SQL Server, teams can use SQL Server images or move to Cloud SQL for SQL Server.
Rehosting also provides initial operational benefits. Custom machine types let teams choose CPU, memory, and storage and pay for what they use. VMs and disaster recovery environments can be started and stopped at will and paid for only when running. Persistent disks can be attached to VMs, snapshots simplify backup and restore, live migration can remove the need for maintenance windows, and rightsizing recommendations can be applied before migration.
Rehosting preserves existing application configurations because it migrates conventional applications as-is or with minimal changes to code or configuration. The Windows and .NET example shows this pattern: on-premises Windows Server and SQL Server instances move to Compute Engine VMs, and the application continues to run with its existing setup. This approach reduces the need for extensive rewriting and lets specialized applications keep their existing behavior.
Rehosting can also limit disruption by keeping some workloads on-premises when they require legacy hardware or software or must meet compliance and local regulatory requirements. A VPN or Interconnect solution securely connects on-premises and cloud resources, so rehosted applications can communicate with other environments over a hybrid network connection. For mission-critical workloads that cannot tolerate downtime or data loss, zero or near-zero downtime migration strategies are available, but they are usually more costly and complex than strategies that require downtime.
Choosing between migration paths depends on the workload's dependencies, condition, and business value:
A decision tree that routes a legacy workload based on its dependencies and value: rehost when it moves as-is, refactor when dependencies must be untangled, and rebuild when it no longer satisfies requirements.
| Migration path | What changes | Choose when |
|---|---|---|
| Rehost (lift and shift) | Workload moves as-is or with minimal code or configuration changes | Workload has few dependencies, is not worth refactoring or cannot be refactored before migration, or is based on third-party software |
| Refactor (move and improve) | Workload is modified to take advantage of cloud capabilities | Dependencies must be untangled, workload relies on systems that cannot be accommodated in the cloud, or it does not use compute or storage efficiently |
| Rebuild (remove and replace) | Application is decommissioned and rewritten as cloud-optimized | Workload no longer satisfies requirements, can be replaced, uses end-of-life technology, or has uneconomical license fees |
Exam tip: Rehost (lift and shift) changes where the workload runs, not how it is built; refactor (move and improve) changes the workload to use cloud capabilities.
Rehost migration, also called lift and shift, moves an application to Google Cloud largely as it is without rewriting code or changing its architecture. The cost and risk value of rehosting is that it shifts spending from capital expenditures to predictable operational costs while keeping architecture stable for applications that cannot be easily refactored. Virtual machines running Windows Server and SQL Server, for example, move to Compute Engine (opens in a new tab) VM instances, where Google manages the physical infrastructure.
Rehost stability means the application keeps the same operating system, dependencies, and structure after it moves from on-premises to the cloud. Data flows from the existing servers to matching cloud VMs, and the application continues to run the same way because no major code changes are required. This stability matters for specialized legacy applications that cannot be easily refactored, including third-party software and workloads tied to specific hardware or licenses.
Choose rehost (lift and shift) when one or more of the following is true for the workload:
Rehost is the lowest-change path when the goal is to keep the workload running while removing the burden of managing physical servers.
Rehost cost change means moving from owning data-center assets to paying for cloud resources as they are used. In an on-premises environment, the organization owns and manages cooling, physical security, hardware maintenance, power, and network cabling. In Google Cloud, the service provider manages physical security, power and cabling, hardware, and virtualization, while the customer manages application resources and can use fully managed services.
After the move, teams can pay for what they use with custom machine types for CPU, memory, and storage, start and stop VMs and disaster recovery environments and pay only when they run, and use rightsizing recommendations before and after migration. Compute Engine also supports autoscaling groups of instances and sizing recommendations for running VMs, which helps match capacity to demand instead of buying fixed on-premises capacity. A complete TCO comparison should include hardware and software plus operational costs for power, cooling, maintenance, and support services, plus cloud network utilization billing, which is often overlooked because on-premises network cabling is a one-time cost. The Google Cloud price calculator can be used to estimate the cost of Google Cloud resources.
Rehost risk reduction comes from avoiding code rewrites, untangling of dependencies, and new deployment toolchains during the initial move. Because the architecture stays intact, there are fewer chances to introduce bugs or security issues, the migration takes less time than refactor or rebuild paths, and teams can focus on learning Google Cloud operations. Programs such as the Rapid Migration Program provide best practices to lower risk, control costs, and simplify the path to cloud success.
Rehost also works as the first step in a phased approach. Lifting and shifting first optimizes the workload from an infrastructure perspective, and after the application runs in the cloud, it is easier to integrate cloud services and then replatform, refactor, or rearchitect. Applications running in Google Cloud can still communicate with remaining on-premises systems over a hybrid network connection using VPN or Interconnect, which is important when legacy hardware, compliance, or regulatory requirements keep some workloads on-premises. Initial operational gains include simplified backup and restore with persistent disks and snapshots, and increased reliability through live migration that avoids maintenance windows for host upkeep.
Deeper migration paths trade higher upfront time and skill requirements for greater cloud optimization:
| Migration type | What changes | Time and skill impact |
|---|---|---|
| Rehost | Application moves as-is to cloud VMs with minimal configuration change | Shortest time; focuses effort on the move rather than code changes |
| Refactor | Workload is modified to improve performance, cost, scalability, and availability | Takes longer than rehost; requires learning new cloud skills |
| Rebuild | Existing app is decommissioned and rewritten as fully cloud-optimized | Takes longer than rehost or refactor; requires new toolchains and is not suitable for ready-made apps |
Exam tip: Choose rehost when refactoring is not feasible or worthwhile; choose refactor when dependencies must be untangled or the operating system, hardware, or database cannot run as-is in the cloud.
Choose rehost (lift and shift) when the workload has a relatively small number of dependencies, is not worth refactoring or cannot be refactored before migration, or is based on third-party software. Rehosting moves the application with little or no change to its code or architecture, while refactoring modifies the workload to use cloud capabilities and rebuilding decommissions and rewrites the application as cloud-optimized.
Google Cloud VMware Engine preserves the VMware environment including vSphere, vCenter, HCX, NSX-T and vSAN policies, so choose it when the application is tightly coupled to VMware and needs minimal change. Google Compute Engine hosts standard virtual machines (VMs) with multi-tenant or sole-tenant tenancy, so choose it when the application can run on standard virtual machines (VMs) and does not need VMware-specific features.
Rehosting shifts responsibility for physical security and safety, power and network cabling, hardware maintenance, and the virtualization platform to Google Cloud while the customer continues to manage application resources. It shifts spending from owning data-center assets to paying for what is used with custom machine types, the ability to start and stop virtual machines (VMs) and disaster recovery environments and pay only when running, rightsizing recommendations, and autoscaling.
Last updated