Guides ·
Azure to AWS migration: service mapping, .NET code conversion and a cutover plan
How to map Azure services to AWS, what AWS Transform's Azure SDK to AWS SDK code conversion covers since 16 October 2026, and an eight-step cutover plan with a rollback window.
By Upvantix team
An Azure to AWS migration has three parts. Map each Azure service to its AWS equivalent. Move each workload: rehost virtual machines with AWS Transform MGN (AWS Application Migration Service), migrate databases with AWS Database Migration Service (AWS DMS), and refactor cloud-native application code. Then cut over wave by wave, with a rollback plan.
On 16 October 2026, AWS made Azure SDK to AWS SDK code conversion generally available in AWS Transform for .NET applications. It converts Azure Functions to AWS Lambda, Azure Blob Storage to Amazon S3 and Azure Queue Storage to Amazon SQS. Azure Cosmos DB to Amazon DynamoDB conversion is still in preview.
Conversion is automatic for eligible modern .NET applications. It changes application code only, not infrastructure. AWS Transform workspaces can be created in Asia Pacific (Mumbai), among other Regions.
What AWS Transform now converts, and what it doesn't
The AWS Transform .NET modernization agent finds Azure SDK use in your submitted source code and shows conversion options in the assessment report and modernization plan. You can accept or adjust those options before the modernization runs. AWS Transform then converts the identified Azure service code to equivalent AWS SDK calls and validates that the changes are incorporated into the transformed code.
Through agent chat, the modernization plan or a steering document, you can opt out of some or all conversions, set naming conventions, change comment style, or redirect a default target. AWS's example is converting Blob Storage code to Amazon DynamoDB instead of Amazon S3. Code that can't be converted is noted in TODO comments, stubbed code, or a generated Next Steps markdown file.
- Code only: AWS says the feature "converts application code only and does not handle migration of cloud infrastructure."
- Eligible modern .NET applications only, and only the four services listed. For other Azure services, AWS suggests AWS Transform custom or an AI code companion such as Kiro.
- Cosmos DB to DynamoDB is a preview feature. DynamoDB uses a different data model, so plan data-model and query testing.
| Criterion | AWS target | Status |
|---|---|---|
| Azure Functions | AWS Lambda | Generally available (16 Oct 2026) |
| Azure Blob Storage (Azure.Storage.Blobs) | Amazon S3 (Amazon.S3) | Generally available |
| Azure Queue Storage (Azure.Storage.Queues) | Amazon SQS (Amazon.SQS) | Generally available |
| Azure Cosmos DB (Microsoft.Azure.Cosmos) | Amazon DynamoDB (Amazon.DynamoDBv2) | Preview |
Azure to AWS service mapping, beyond code
These are nearest equivalents, not one-to-one guarantees. Verify each mapping per workload.
| Criterion | AWS equivalent | How to move it / notes |
|---|---|---|
| Virtual Machines | Amazon EC2 | Rehost with AWS Transform MGN, which performs continuous block-level replication of physical, virtual and cloud servers running supported 64-bit Windows and Linux versions. Ubuntu servers on the Azure-tuned kernel must switch to a standard kernel or the linux-aws kernel first. |
| Azure SQL Database | Amazon RDS for SQL Server | AWS DMS supports Azure SQL Database as a source, but doesn't support change data capture (CDC) with it. Plan a full-load cutover window or another sync method. |
| Azure SQL Managed Instance | Amazon RDS for SQL Server | AWS DMS supports it as a source in the same way as SQL Server. |
| SQL Server (optional later step) | Amazon Aurora PostgreSQL | AWS Transform for full-stack Windows modernization runs in US East (N. Virginia) only. It works from SQL Server hosted on Amazon RDS or EC2, or from an uploaded DDL export. It is not a direct Azure SQL connection. |
| Azure Database for PostgreSQL flexible server | Amazon RDS for PostgreSQL or Aurora PostgreSQL | AWS DMS supports it as a source; ongoing replication needs logical replication enabled on the Azure server. |
| App Service | AWS Elastic Beanstalk, or Amazon ECS / Amazon EKS | Elastic Beanstalk serves a similar role; choose containers where they fit better. Decide per app. |
| Azure Functions / Durable Functions | AWS Lambda / AWS Step Functions | AWS Transform converts eligible .NET function code. |
| Blob Storage, Queue Storage, Service Bus | Amazon S3, Amazon SQS, Amazon SQS / Amazon SNS | Use SQS FIFO queues where you need ordered processing. |
| Cosmos DB | Amazon DynamoDB | AWS Transform conversion is in preview. |
| Key Vault | AWS Secrets Manager / Parameter Store | Parameter Store is a capability of AWS Systems Manager. |
| Managed Identity | IAM roles | For Lambda, the function's IAM execution role. |
| Entra ID | Amazon Cognito / API Gateway authorizers; AWS IAM Identity Center | App authentication moves to Cognito or API Gateway authorizers. Workforce single sign-on can federate Microsoft Entra ID into IAM Identity Center with SAML 2.0 and SCIM. |
| Azure Monitor, Application Insights, Log Analytics | Amazon CloudWatch, AWS X-Ray, CloudWatch Logs Insights | Rebuild alerts as CloudWatch alarms. |
| VNet / Private Endpoints | Amazon VPC / private connectivity in the VPC | AWS maps Private Endpoints to a VPC with a private load balancer or private API Gateway; use VPC endpoints and AWS PrivateLink for private access to AWS services. |
Source
- AWS .NET Blog: From Azure Functions to AWS Lambda
- AWS Solutions: Migrating from comparative platforms
- AWS DMS: Azure SQL Database as a source
- AWS DMS: Azure SQL Managed Instance as a source
- AWS DMS: Azure Database for PostgreSQL flexible server as a source
- AWS Transform MGN: supported operating systems
- AWS News Blog: AWS Transform full-stack Windows modernization
- IAM Identity Center with Microsoft Entra ID
A cutover plan that survives contact
- Inventory and dependencies: apps, data stores, identities, network paths and scheduled jobs. Tag each workload rehost, replatform or refactor.
- Landing zone first: accounts, identity, network and logging on AWS.
- Waves: start with low-risk apps, and group them by shared data and identity dependencies.
- Code conversion for in-scope .NET apps with AWS Transform. Review every diff, then run unit and integration tests. Treat Cosmos DB conversions as preview.
- Data sync: AWS DMS full load plus CDC where the source supports it. Azure SQL Database needs a full-load window or an alternative.
- Rehearsal: a dress-rehearsal cutover with timings, plus agreed rollback criteria.
- Cutover: freeze, final sync, a DNS switch with a low TTL set in advance, smoke tests, then hypercare. For serverless functions, AWS's .NET migration playbook calls a parallel run followed by a weighted cutover the safest pattern, with Lambda aliases for rollback.
- Decommission on Azure only after the rollback window closes. Watch for egress and duplicate-run costs while both environments run.
Region and data-handling considerations
AWS Transform workspaces can be created in Asia Pacific (Mumbai) and US East (N. Virginia), among other Regions. The workspace you create a job in determines the job's Region. AWS also notes that your data might be processed in a different Region from the one where you use AWS Transform, so check its cross-Region processing page against your code-handling policy.
AWS's Supported Regions list for AWS Transform doesn't include a Middle East Region. Teams in the UAE should choose a workspace Region that fits their code-handling policy.
Target Regions for the migrated workloads are a separate decision. AWS Transform MGN, for example, supports Asia Pacific (Mumbai), Asia Pacific (Hyderabad) and Middle East (UAE).
How Upvantix can help
As an AWS Select Tier Services Partner, Upvantix can assess your Azure estate, build the AWS landing zone, run AWS Transform and the migration waves with your team, and support cutover. No funding, credits or savings are promised.
Azure to AWS migration: FAQ
Can AWS Transform convert Azure code to AWS?
Yes, for eligible modern .NET applications. Since 16 October 2026 it converts Azure Functions, Blob Storage and Queue Storage SDK code to AWS Lambda, Amazon S3 and Amazon SQS. Cosmos DB to DynamoDB conversion is in preview.
Does AWS Transform migrate Azure infrastructure?
No. Azure-to-AWS code conversion changes application code only. Move virtual machines with AWS Transform MGN and databases with AWS DMS.
Is AWS Transform available in the Mumbai Region?
Yes. AWS lists Asia Pacific (Mumbai) among the Regions where you can create AWS Transform workspaces. AWS notes that some data may be processed in another Region.
What happens to Azure code AWS Transform can't convert?
It is noted in TODO comments, stubbed code or a generated Next Steps file, for your team to finish.
Can I migrate Azure SQL Database with AWS DMS?
Yes, as a source. AWS documents that DMS doesn't support change data capture with Azure SQL Database, so plan the cutover window accordingly.
How should we sequence an Azure to AWS cutover?
Inventory and dependencies, landing zone, migration waves, code conversion and testing, data sync, a rehearsal, then cutover with a rollback window before decommissioning Azure.