Rehosting on Azure virtual machines is usually the fastest first move for IIS and SQL Server applications. Replatforming to App Service and a managed SQL service reduces server upkeep but requires configuration and code changes. Choose based on how much the app depends on Windows-specific features and how much change your team can absorb.
Key takeaways
- Rehosting moves IIS sites and SQL Server databases to Azure with few code changes, which makes it the usual first step.
- Replatforming trades configuration and code work for less server maintenance and managed platform services.
- Windows-specific dependencies, licensing and database size decide which path is realistic for a given application.
- Both paths need a full inventory, a test plan and a cutover rehearsal before production traffic moves.
What does rehosting involve?
Rehosting, often called lift and shift, means recreating the IIS sites on Windows Server virtual machines in Azure and moving SQL Server to a virtual machine with the same engine version. The direct answer is that rehosting keeps the runtime, configuration model and database engine nearly unchanged, so it is usually the quickest route to the cloud.
Microsoft's Azure Migrate documentation describes discovery and assessment tools that help map servers and dependencies before a move. Those tools find what runs on a server, but they do not see every hidden dependency, so a manual review remains necessary.
The work usually falls into a fixed set of items:
- IIS bindings, certificates, application pools and the identities they run under.
- Connection strings, DNS entries and any hard-coded server names or IP addresses.
- SQL Server logins, SQL Agent jobs, linked servers, database mail and maintenance plans.
- File shares, scheduled tasks, Windows services and COM components that the application calls.
When does replatforming make sense?
Replatforming keeps most of the application logic but moves hosting components onto managed services, such as Azure App Service for the web tier and a managed SQL offering for the database. It makes sense when the team wants less patching and backup work and can accept configuration changes in exchange.
Azure App Service is a platform service with its own deployment and configuration model, described in the App Service overview. Azure SQL Managed Instance offers a near-complete SQL Server surface as a managed service, and Microsoft lists its supported features in the Managed Instance overview. Check that list against every feature your application uses, because support differs between managed options.
Replatforming is a stronger choice when the application has stateless web tiers, predictable traffic patterns and a database whose features fit the managed service. It is a weaker choice when the application depends on local file paths, custom COM objects or SQL features that the target does not support.
How should you assess the application first?
Assessment answers the question of which path is realistic before any tool moves data. Work through the following checklist with the people who own the application, not only the infrastructure team.
- List every external dependency, including third-party APIs, file shares and mail relays.
- Record the SQL Server version, edition, database sizes and the use of CLR, linked servers or SQL Agent.
- Identify which components write to local disk and whether that data must persist across restarts.
- Confirm which licences apply to SQL Server and Windows Server, and whether existing licences can move to Azure.
- Define acceptable downtime for the cutover and the rollback steps if the new environment fails.
Once the inventory is complete, run a test migration of a copy of the environment. The test shows where configuration assumptions break and gives a realistic estimate of the cutover window.
Hypothetical examples
Here are two hypothetical scenarios:
- Scenario one: an internal ASP.NET application writes generated reports to a local folder, calls a COM component installed on the server and runs SQL Agent jobs for nightly imports. A rehost to a virtual machine keeps these behaviours intact, while a replatform would first require redesigning the file storage and rewriting the COM dependency.
- Scenario two: a public-facing site renders stateless pages, stores sessions in the database and has a single SQL database with few scheduled jobs. This profile suits replatforming, because the web tier can move to App Service and the database to a managed service after compatibility checks.
Neither scenario is a recommendation for a specific organisation. They show how the same platform question leads to different answers depending on the dependencies that the application already has.
How does moving IIS and SQL Server to Azure differ by path?
Rehosting gives the fastest move with the least change, while replatforming gives lower long-term operations effort at the cost of more preparation. The table compares the two paths on qualitative criteria.
| Criterion | Rehost on virtual machines | Replatform to managed services |
|---|---|---|
| Code changes | Minimal | Configuration and some code changes |
| Server patching and backups | Remains with your team | Largely handled by the platform |
| Fit with Windows-specific features | High | Depends on the target's feature support |
| Preparation effort | Moderate | High |
| Best first step when | Speed and stability matter most | The application can change and long-term upkeep is the goal |
Where does each option fail?
Rehosting fails when the virtual machine inherits problems that cloud hosting was meant to remove, such as unpatched servers and manual backup routines. Replatforming fails when a feature gap or a hidden dependency is discovered after the cutover date has been announced.
- Rehost failures: oversized virtual machines kept running without review, backups not tested, and licence costs left unexamined.
- Replatform failures: unsupported SQL features, session state stored on local disks, and secrets hard-coded in configuration files.
- Shared failures: DNS and certificate changes that were not rehearsed, and no rollback plan for data written during the cutover.
For help with either path, you can review the Cloud and DevOps service or read the guide to securing cloud infrastructure before you finalise the target design. Integration work between the migrated application and other systems is covered on the Integrations page.
Frequently Asked Questions
Is rehosting always cheaper than replatforming?
Rehosting is not always cheaper, but it usually needs less preparation and testing time at the start. Running virtual machines keeps the server management work with your team, so the total cost depends on operations effort, licensing and how long the environment stays in its current form.
Can I move SQL Server to Azure without changing the application?
Often yes, if the database uses only features supported by the target and the connection strings are updated. The application may still need changes to configuration, file paths and authentication. A test migration against a copy of the database is the reliable way to confirm this.
Should I rehost first and replatform later?
Rehosting first is a common sequence because it gets the workload out of an aging datacentre with limited change. The second step can then be scheduled separately, once the team has stable monitoring and a tested rollback plan.
What should I test before a production cutover?
Test authentication, scheduled jobs, file access, certificates, DNS resolution and data reconciliation against a recent copy of production. Also rehearse the rollback. A cutover is ready only when these checks pass together in one full run.
Who should decide between rehosting and replatforming?
The decision should involve the application owners, the operations team and whoever owns the database and licensing. Each group sees different risks, and the choice is clearer when all three review the same dependency inventory.
Conclusion
Moving IIS and SQL Server to Azure is less a single decision than a sequence of checks. Rehosting offers speed and fewer code changes, while replatforming offers lower operations effort once the application fits the managed services. The inventory, the test migration and the rollback plan matter more than the label. If you want a second opinion on your dependencies, request a project consultation at https://ruicodex.com/Contact or email info@ruicodex.com to discuss your project. You can also review related delivered work at https://ruicodex.com/CaseStudies.
Sources
- Azure Migrate overview - Microsoft Learn
- Azure App Service overview - Microsoft Learn
- Azure SQL Managed Instance overview - Microsoft Learn