Azure Migration Strategies
1. Plan your migration.
Your first task is to plan
your company's migration to Azure. You need to present a plan to your
leadership team to get their support and approval.
2.
Azure migration
Framework.
2.1.
ASSESS
A.
Discovery and evaluation
Start with a full assessment of your current environment.
Identify the servers, applications, and services that are in scope for
migration. You can then bring in the IT and business teams that work with those
services. By bringing these teams in as early as possible in the process, you
ensure that they can provide guidance, feedback, and support for the migration.
Next, produce a full inventory and dependency map of
servers and services that are in scope for migration. The inventory and map
determine how those services communicate with each other. A modern business
might have hundreds of applications spread across the estate. Each application
must be fully investigated before any work takes place.
For each application, there are multiple migration options:
·
Rehost: Recreate your existing infrastructure in Azure.
Choosing this approach has the least impact because it requires minimal
changes. It typically involves moving virtual machines from your data
center to virtual machines on Azure.
·
Refactor: Move services running on virtual machines to
platform-as-a-service (PaaS) services. This approach can reduce
operational requirements, improve release agility, and keep your costs
low. Small enhancements to run more efficiently in the cloud can have large
impacts on performance.
·
Rearchitect: You might be forced to rearchitect some systems so that
they can be migrated. Other apps could be changed to become cloud-native
or to take advantage of new approaches to software such as containers or
microservices.
·
Rebuild: You might need to rebuild software if the cost to
rearchitect is more than that of starting from scratch.
·
Replace: While you're reviewing your estate, it's possible you'll
find that third-party applications could completely replace your custom
applications. Evaluate software-as-a-service (SaaS) options that can
be used to replace existing applications.
Review each application to determine which of the
above option is the best fit.
B.
Involve key stakeholders.
Applications are used by specific sections of the business. The
owners and superusers of applications have a wealth of experience on which to
call. Involving these people in the planning stage increases the chance of a
successful migration. These resources can offer guidance in areas where the
person running the migration project might have knowledge gaps. Often, each
area of the business will appoint business owners to manage specific sections.
C.
Estimate cost savings.
Part of the business's plan to migrate to Azure could be to
reduce costs because moving to the cloud offers cost savings over running your
own on-premises estate. After you complete the initial scoping exercise, use
the Azure Total Cost of Ownership (TCO) Calculator to
estimate the real costs of supporting the project in light of the company's
longer-term financial goals.
D.
Identify tools.
Several tools and services are available to help you plan
and complete the four stages of migration. In some migrations, you may only
need to use one or two of these tools.
|
Service
or tool |
Stage |
Use |
|
Assess
and migrate |
Performs
assessment and migration of VMware VMs, Hyper-V VMs, cloud VMs, and physical
servers, as well as databases, data, virtual desktop infrastructure, and web
applications, to Azure. |
|
|
Assess |
Maps communication between application components on
Windows or Linux. Helps you identify dependencies when scoping what to
migrate. |
|
|
Assess |
Estimates
your monthly running costs in Azure versus on-premises. |
|
|
Migrate |
Uses the Data Migration Assistant and the Azure portal to
migrate database workloads to Azure. |
|
|
Migrate |
Migrates
existing databases to Azure Cosmos DB. |
|
|
Optimize |
Helps you monitor, control, and optimize ongoing Azure
costs. |
|
|
Optimize |
Helps
optimize your Azure resources for high availability, performance, and cost. |
|
|
Monitor |
Allows you to monitor your entire estate's performance.
Includes application-health monitoring via enhanced telemetry, and setting up
notifications. |
|
|
Monitor |
Provides
intelligent security analytics for your applications. |
2.2.
MIGRATE
A.
Deploy cloud infrastructure targets.
You'll need destination systems and services on Azure to
which you can migrate. The scope of your migration has been defined as your
company's current VMware machines and existing relational databases. In this
scenario, you don't need to create the resources in Azure beforehand. The two
tools you'll use to do the migration, Azure Migrate, and the Azure Database
Migration Service, will create the required Azure resources for you.
In other situations, you may need to set up resources in Azure
to have them available as a migration destination.
B.
Migrate workloads.
It's often best to start with a small migration
instead of migrating a large, business-critical workload. This approach lets
you become familiar with the tools, processes, and procedures for migration. It
can reduce the risk of issues when you migrate larger workloads. As you become
more comfortable with the migration process, you can progress to larger and
more business-critical workloads.
Each tool will guide you through the migration. The steps to
complete them are covered in later parts of this doc. At a high level, the
steps are:
·
Prepare the source (vCenter Server) and
target (Azure) environments.
·
Set up and start the replication
between the two.
·
Test that the replication has worked.
·
Failover from the source servers to
Azure.
For the database migrations, the high-level steps are:
·
Assess your on-premises databases.
·
Migrate the schemas.
·
Create and run an Azure Database Migration
Service project to move the data.
·
Monitor the migration.
C.
Decommission on-premises infrastructure.
After all migrated workloads have been tested and verified as
successfully migrated to Azure, you can decommission all your on-premises
systems. Even after you decommission them, it can be useful to keep backups and
archive data from the migrated systems. This practice gives you a historical
archive of data in case it's needed. This data could be stored on-premises, or
in a cloud-storage service such as Azure Blob storage.
2.3.
OPTIMIZE
After your services are migrated, it's important to
optimize them to ensure that they're running as efficiently as
possible from a cost and performance standpoint.
A.
Analyze running costs.
Use Microsoft
Cost Management to start analyzing your Azure costs at different
management scopes. For example, by choosing a subscription in the portal, you
can see a breakdown of all the resources for that subscription. You could also
view a resource group to see all the costs associated with all the resources in
just the selected group:
B.
Review opportunities to improve.
Microsoft
Cost Management shows you cost-reduction advice
from Azure Advisor. The advice includes suggestions like reducing
the performance of underused VMs, making use of additional discounts, or
reserving resources instead of paying as you go. Azure Advisor also shows you
recommendations for network security, high availability, and performance.
Review the recommendations that Advisor presents to further optimize your
environment.
2.4.
MONITOR
A.
Integrate health and performance
monitoring.
Azure
Monitor can capture health and performance information from
Azure VMs if you install a Log Analytics agent. You can install the
agent on machines running either Windows or Linux, and you can then set up
alerting and reporting.
You can set up alerts based on a range of data sources, such
as:
·
Specific metric values like
CPU usage
·
Specific text in log files
·
Health metrics
·
An auto scale
metric
It's also important to have event logging and visibility into
security events across your enterprise. Microsoft Sentinel provides security information
and event-management (SIEM) capabilities, along with
artificial intelligence to help you protect against, detect, and respond to
security events. This information helps security operations (SecOps) teams
triage critical alerts and prioritize work effectively.
EmoticonEmoticon