Azure Migrate: Simplifying VMWare to Azure Migrations

Dale Kirby, Cloud Solution Architect

Companies with large virtual footprints often have less-than-ideal documentation on what is actually running in their environments. These organizations can have lots of virtual hosts, each of which have multiple VMs. Given that this infrastructure has been changing and growing over many years, it’s inevitable that insufficient details are available covering what’s what in the environment.

The complexity of mapping out each VM, cores allocated, memory and storage used, and app dependencies is only part of the challenge. We also have to ask, “Will this VM easily migrate to the cloud, or will there be some work to get it migration-ready?” Factors like OS version and boot type are critical to know before attempting migration. We often need to consider “right-sizing” VMs by examining utilization of CPU, memory, and storage, and ensuring that when we move a VM from on-premises to the cloud we aren’t over-allocating cloud resources (and therefore paying for more than what’s needed).

The top questions you need answered before moving these complex environments to the cloud are:

  1. Which VMs can be moved easily, and which will require some additional prep?
  2. Where can I “right-size” VMs for cloud based on their actual usage patterns (CPU, memory, storage)?
  3. Where do VMs running multi-tier apps have dependencies on one another?
  4. What are the costs to run my infrastructure once migrated?

Azure Migrate is a free migration planning tool, and can help answer these questions. Azure Migrate uses agentless or agent-based discovery to assess a VMWare environment, identify Azure readiness for VMs, make suggestions for right-sizing, and show dependencies among VMs supporting tiered applications. Let’s explore how it works.

Step 1 – Create the Azure Migrate project

First, log in to the Azure portal and search for “Azure Migrate” in the top search bar. Select Azure Migrate under Marketplace and create your project by selecting a project name, resource group, and location, as pictured below.

Step 2 – Configuring the Discovery virtual appliance and initiating discovery

Once our project is created, click Discover Machines to begin the process of discovery. Discovery starts with installing a virtual appliance in the vCenter environment.

Once installed, a wizard walks you through configuration and initiating discovery. By default, discovery is agentless and doesn’t require any installation on individual VMs. An agent-based discovery is also available, which can be used to map VM dependencies.

Note: If you’re using the hands on lab environment linked below, you will need to enable copy/paste on the collector VM in vCenter before you power up the VM.


Step 3 – View Assessments and Migration Recommendations

Once discovery is complete, it’s time to create assessment reports that show Azure readiness, right-sizing recommendations, application dependencies, and cost estimates. Most machines will likely be ready to migrate without issues. Others will be ready with conditions, meaning there may be some work required beforehand, such as moving from EFI to BIOS boot or upgrading older OS versions. In the example pictured below, a large majority of VMs are ready, while others have conditions that will need to be checked before migrating. You can also see helpful information including the best migration path for a VM, which may be Azure Site Recovery for migration.

You can also use agent-based discovery to view dependencies between VMs. This feature leverages Azure Service Map to show the dependencies in graphical format. In the example below, see how DataTierVM01 is dependent on a few other VMs.

Azure Migrate will also show CPU and memory utilization, and make a “right-size” recommendation for the corresponding Azure VM series and size. Pictured below, you can see that this VM is not utilizing much CPU or memory, so it can potentially be downsized for Azure and save on cost. You can also configure Azure Migrate to plan for “as-is” migration without right-sizing if you plan to keep original CPU and memory ratios. Azure Migrate also shows an estimated monthly cost for this VM.

Azure Migrate has other great features, including the ability to create multiple assessment reports from a single discovery, use vCenter tags and groups for organizing the migration, and make recommendations on database migration paths and VM storage allocations. There’s plenty to check out, so try it out for yourself today with this free lab.

Cloud migrations can be complex, but with the right tools they don’t have to be intimidating. With Azure Migrate, you get a jumpstart on your migration and maximize visibility into existing infrastructure. And if you’re looking for additional hands-on support for your migration, Microsoft has several partners that offer dedicated migration services. Whether you’re already an Azure veteran or you’re just getting started, we’re here to help. Check out Azure Migrate today and continue your journey towards digital transformation.

Applications & Infrastructure Community call

Excited about what you’ve read? Want to hear more? Register for the Applications and Infrastructure community call scheduled for Friday, April 20, 2018 at 8:30 am PT. We’ll talk about these innovations in a conversational dialogue and go in-depth about Azure Migrate.

Applications and Infrastructure Technical Community