Assess your workload

Completed

In the Assess phase, you evaluate the readiness of your workload and plan for the migrated state. After you complete this phase, you can deploy the workload for migration.

Diagram that shows the steps of the Migrate methodology.

Your cloud adoption team should evaluate technical compatibility, the required architecture, performance and sizing expectations, and dependencies. Use this information to ensure that you can deploy the migrated workload to the cloud effectively.

Classify workloads

Classify your workloads to help clarify governance, security, operations, and cloud-scale analytics requirements. Classify data based on how a potential leak of that data affects a business or customers. Highly sensitive data increases the security risk. Workload criticality is based on how significantly an outage affects a business.

After you classify workloads, share those details with your supporting teams. Low or unsupported workloads are likely to have little influence on your supporting teams. But when workloads approach mission-critical or unit-critical classifications, they have more operational dependencies.

Evaluate workload readiness

When you want to migrate a workload, your cloud adoption team must ensure that all assets and associated dependencies are compatible with your deployment model and cloud provider. Your team should document any required efforts to remediate compatibility problems.

Design your workload architecture

Before you migrate, you must design the intended migrated state of your workload.

Consider common design assumptions to design your:

As you finish your architecture design, revisit your cloud estimate to make sure that you're still within the planned budget.

Migrations tend to focus on maintaining an existing architecture and transitioning it to a cloud platform. But there are times when you might need to rearchitect your workload, even for migration. You might need to make architectural changes before you migrate if you need to:

  • Pay for technical debt.
  • Improve reliability.
  • Optimize high-cost workloads.
  • Meet performance requirements.
  • Secure applications.