Migrating AWS EC2 to GCP with Google Cloud’s Migrate to Virtual Machines

As businesses evolve, many organizations reassess their cloud strategies — whether to optimize operational costs, consolidate data across platforms, or respond to new organizational directives. In such scenarios, executing a cloud migration with precision and minimal disruption becomes critical. **Google Cloud’s Migrate to Virtual Machines** offers a compelling solution, providing a highly automated lift-and-shift capability that facilitates the seamless transition of virtual machine workloads from AWS and other environments to Google Cloud.

In this blog post, I’ll walk through the process of migrating Amazon EC2 instances to Google Cloud Platform (GCP) using Migrate to Virtual Machines, highlighting its key features, architectural components, and step-by-step workflow.

Why Choose Migrate to Virtual Machines?

Migrate to Virtual Machines enables frictionless, enterprise-grade migrations suitable for businesses of all sizes. Key benefits include:

• Agentless migrations: No need to install client-side agents.
• Automated replication: Continuous background syncing of VM disk data using change block tracking.
• Minimal downtime: Built-in test-cloning and short maintenance windows make the cut-over phase predictable and safe.
• Scalability: Migrate individual VMs, grouped workloads, or build a migration factory with Cloud APIs.
• Cost and operational efficiency: Reduce labor, eliminate on-prem software/hardware costs, and use analytics to rightsize migrated VMs.

Architecture Overview

Migrate to Virtual Machines is designed to migrate virtual machine (VM) instances and their disks from various sources — including On-premises vSphere environments, AWS EC2 instances, Azure VMs, and Google Cloud VMware Engine — to VM instances or Persistent Disk volumes on Google Cloud.

The tool uses continuous replication to synchronize data from source VMs to GCP without causing downtime. Once replication is in place, you can test clones before initiating the final cut-over to production.

The general architecture for migrating Amazon EC2 instances to GCP involves setting up communication between your AWS environment and GCP, allowing Migrate to Virtual Machines to manage the replication and migration process. Source VMs in AWS have their data replicated to Google Cloud Storage, which is then used to create Compute Engine instances.

High-level architecture for migrating AWS EC2 instances to GCP

Migration Workflow

Here’s a high-level walkthrough of the migration process from Amazon EC2 to GCP:

Task 1: Enable Required APIs

The first step is to enable the necessary APIs in your Google Cloud project. Primarily, you’ll need to enable the VM Migration API to allow the service to function.

Task 2: Onboard VMs in a Group

Migrate to Virtual Machines allows you to organize your source VMs into groups to perform bulk migration tasks on multiple VMs simultaneously. This approach streamlines the planning and execution process, enabling you to migrate VMs in manageable batches.

Before starting replication, you add one or more source VMs from your AWS environment to a group. You can then perform migration operations on the entire group or a subset of VMs within it.

Task 3: Start Replication

This phase initiates the replication of disk data from the source AWS EC2 instances to Google Cloud. This is a continuous background process:

• First Replication: Migrate to Virtual Machines creates an initial snapshot of the source VM data disks and replicates this snapshot data to Google Cloud. The duration of this first replication can vary from minutes to hours, depending on the volume of disk data.
• Incremental Replication: After the successful completion of the first replication, incremental replications occur at set intervals (defaulting to every two hours). In each step, a new snapshot is taken for each data disk. Only the data that has changed since the previous replication (identified using a Change Block Tracking — CBT — mechanism) is replicated to Google Cloud.

Task 4: Configure Group Targets

With replication underway, you can configure the target details for the group. These settings define the landing zone for your migrated VMs on Google Cloud during the test-clone and cut-over phases.

Target details include the Google Cloud project, Compute Engine instance type, network settings (VPC, subnet), and more. Migrate to Virtual Machines will use these details to create the Compute Engine instance that will host the migrated VM. These target details can be modified at any time before initiating a test-clone or cut-over.

Task 5: Test a Clone

In the test-clone phase, you deploy a clone of a migrated VM to a Compute Engine instance within your testing environment on GCP. While optional, this step is a highly recommended best practice to validate the migration setup and application functionality before moving to production.

The test-clone Compute Engine instance is created from the latest available replication data. You can test any VM that is in an “Active” state, indicating the initial replication was successful.

Task 6: Cut-Over to GCP

The cut-over phase is when you transfer operational control to your migrated VMs now running as Compute Engine instances in your production environment on Google Cloud. This phase involves a short period of VM downtime and should be scheduled during a planned maintenance window.

During a planned maintenance window:

• Shut down source EC2 instances.
• Initiate cut-over, launching the final VMs in GCP production.
• Redirect traffic to the new instances.

The entire cut-over typically completes within 15–20 minutes.

Task 7: Finalize the Migration

After a successful cut-over, the replication data used to create the Compute Engine VM is retained. This allows for the possibility of creating additional instances if needed. However, to complete the migration and free up resources, you must finalize it.

Finalizing a migration deletes all replication data and any other storage resources associated with that migrated VM. It also changes the state of the VM in the Migrate to Virtual Machines console to “Finalized.”

Conclusion

In this blog post, I demonstrated the migration of Amazon EC2 instances to GCP using this powerful service. At a high level, the steps involved defining settings and initiating continuous data replication, performing thorough testing by launching instances for non-disruptive validation, and finally, executing the migration by launching cutover instances on GCP.

If you’re exploring multi-cloud strategies or seeking cost and operational advantages, this tool can significantly simplify your migration journey from AWS (or other environments) to Google Cloud.