A beginner’s guide to cloud migration

Matt Barlow

icon of server moving up into the cloud

 

Unless your business was born in the cloud, you likely rely on some IT applications or legacy infrastructure that you own, host or manage yourself. These systems may have fueled your growth in the past, but as you move toward newer technologies — such as AI, machine learning and automation — your legacy, non-cloud-based systems can hold you back. They simply aren’t built to take advantage of cloud native technologies.

In this article, we’ll work to set you on the right path to migrating your workloads, applications and processes to the cloud. You’ll learn the basics about moving workloads to the cloud — as well as how to still get value out of your legacy investments. We’ll also show you how to get help when you need it. In the end, you’ll be one step closer to transforming your business. 

 

What is cloud migration?

A cloud migration is the process of moving digital assets — including workloads, data and applications — to a public or private cloud environment. It also involves decisions around how you plan to use, maintain, optimize and govern your cloud once the digital migration is complete.

 

What are the benefits of cloud migration?

The specific benefits of a cloud migration are closely tied to the benefits of the platform you choose. For example, if you migrate to a managed private cloud platform, you’ll experience substantial security and performance gains. And if you migrate to a public cloud platform, your benefits will center more around microservices and flexibility. But regardless of which platform you choose, moving from an on-premises environment to a hosted one will generally result in:
 

Cost efficiency

By moving away from expensive, legacy infrastructure, you can realize instant savings on IT operations, as you move from a capex to an opex model. This lets you hold on to more cash or reinvest your capex budgets into business-critical initiatives.

 

Productivity improvements

In the cloud, your IT can be freed of its operational burdens and turn its expertise toward building what’s next. For end-users, the cloud provides enhanced functionality — allowing them to do their jobs faster and more efficiently.

 

Innovation enhancements

Modernized cloud infrastructure gives your IT team greater agility so that they can deliver new functionality to users faster. Leveraging the cloud also allows them to implement cutting-edge technology like AI, IoT and machine learning to drive innovation. Attempting to execute such resource-intensive technologies on legacy hardware is not only expensive but, in some cases, impossible.

 

Assessing your applications before migrating

The first step in a cloud migration is workload selection. If your project is too large, you’ll likely face scope creep and long timeframes — so start with a small, impactful workload and move on to increasingly complex workloads later, after you’ve gained some cloud migration experience.

Use application profiling to gather and organize information about your workloads and applications. Follow these steps to assess and prioritize workloads for cloud migration:
 

  1. Audit your existing environment’s metrics around compute needs, performance output, response times and other factors important to business operations. This will help you establish a baseline and develop KPIs for the incoming platform.
     
  2. Itemize and collect key information about your workloads, such as physical and virtual server configurations, network topology, compliance requirements, data and application dependencies, geographic considerations and user needs. This will help you set requirements for selecting the right cloud platform to support your environment.
     
  3. Based on the audit and information gathering, categorize your workloads in order of migration complexity. Identify which workloads can be easily migrated without the need to replatform or refactor. Prioritize these easy-to-migrate workloads for cloud migration. 

 

What is the right cloud deployment model?

Once you’ve identified your candidate workloads, align their requirements with their best-fit cloud platform. Though we refer to “the cloud” as a singular entity, there are multiple cloud types to consider. This is why assessing your applications and workloads is so critical. That effort will help you make informed cloud platform choices based on what you need and what the platform can provide.
 

  • Public cloud. In a public cloud, infrastructure is shared by multiple businesses and owned and operated by a service provider. Because it allows you to easily scale resources up and down to meet demand and pay-as-you-go, it’s a great option for managing unpredictable traffic and maximizing cost savings.
     
  • Private cloud. In a private cloud, infrastructure is dedicated entirely to your business. This gives you the ability to customize your compute, storage and networking and, as a result, achieve greater levels of control and security. Depending on your workload requirements and resource utilization, private cloud may also result in more cost savings than public cloud infrastructure.
     
  • Hybrid cloud. Some workloads require a hybrid cloud that connects both public and private cloud environments. A hybrid cloud can give you the control of the private cloud for your sensitive, business-critical assets, plus the flexibility and cost savings of the public cloud for your public-facing operations.
     
  • Multicloud. Basically, multicloud means what it sounds like: multiple clouds. From your on-premises data center and private clouds to hyperscale clouds like Amazon Web Services (AWS), Microsoft® Azure® and Google Cloud Platform™ (GCP), cloud-based SaaS applications and even colocation environments — it all comes together to create your unique multicloud.

 

And the lines are starting to blur a bit between some of these definitions. With solutions like AWS Outposts, now you can bring a public “hyperscale” cloud into your own data center, on dedicated hardware. And you can run a private VMware Cloud on AWS. This just means you have more options to choose from, as you look for the right platform for your cloud migration.

 

Choosing your cloud migration strategy

After you’ve decided which workloads belong on which clouds, you need to select the best path for transitioning from here to there. An organization will likely use multiple migration strategies across workloads as there is no one-size-fits-all approach. For example, an organization’s monolithic ERP system may use a lift-and-shift strategy for technical or licensing reasons, while an HR system is completely replaced by a SaaS option.

There are six common cloud migration strategies for making the move:

 

1. Replace

In this approach, you completely decommission and replace your existing legacy components with a cloud-based alternative. This creates a fast route to the cloud, but it requires a lot of planning — and you still have to tackle the job of migrating data from one system to the other, or choosing to leave it behind.

 

2. Rebuild

You can choose to fully rebuild your legacy elements, creating a fully modernized, cloud native solution. This is considered the longest and most-expensive migration type, but the end state delivers the most benefits. Since it’s built in the cloud and for the cloud, you can integrate the latest technologies, like containers, service meshes, microservices, immutable infrastructure and declarative APIs. As a result, you can achieve more flexibility, better performance and lower operating costs in the long term.

 

3. Replatform

If you’re looking for a migration solution that gets your workload onto the cloud, with minimal code changes, consider replatforming. This involves porting your application’s components to a new runtime platform — such as moving a COBOL-based system from a UNIX or mainframe system to a LINUX or Windows environment. Your application’s features and functions stay the same, and you can take advantage of the cloud’s inherent cost savings and scalability, with minimal effort. Plus, you can continue to derive value from your legacy solution.

 

4. Rehost

When you rehost, you essentially “lift and shift” your digital assets to a new environment (physical, virtual or cloud infrastructure) — without any code changes or functionality adjustments needed. This approach gets you to the cloud faster, but it doesn’t take full advantage of cloud native tools, performance or cost benefits. And with emulators, such as the COBOL mainframe emulator for Windows, you can essentially recreate your legacy environment, but on modern infrastructure.

 

5. Refactor & rearchitect

This option allows you to migrate portions of your application to the cloud, while leaving other elements in your legacy environment. For example, you might continue to host your monolithic application in-house, while moving its database to the cloud, where it can achieve better performance and take advantage of cloud-based analytics tools. You may still need to make some backend adjustments to legacy elements, but this approach can help you move your monolithic application piece-by-piece to the cloud.

 

6. Retain

Of course, sometimes it makes sense to simply retain your current situation and not make any changes or updates. For example, if you’re anticipating an upcoming merger or end-of-life announcement, there might not be a compelling enough reason for modernization. Or perhaps you need to maintain certain elements that serve as “connectors” or “bridges” to other modern elements in your organization. But there should still be a long-term strategy in place, since the longer you hold on to inefficient, resource-heavy infrastructure, the less budget and resources you’ll have when you’re ready to innovate.

 

7. Retire

Sometimes, the best route is simply to retire a legacy system and move your users to other systems that are already in place. This often requires you to redesign some processes, but it also can be an opportunity to improve and optimize those processes.

 

Four-step cloud migration process

Every migration is different based on a variety of factors, like application maturity, level of infrastructure complexity and skill level of the IT team. As a general guide, every migration will include some form of the following four steps. It’s important that these steps are taken to avoid making the wrong technology and methodology choices or encountering surprises during the migration process.
 

Step 1: Plan and assess

Map your entire environment, including dependencies, services, applications, and physical and virtual server configurations. Capture everything including shadow IT implementations and third-party resources that could complicate migration if not properly considered.
 

Step 2: Design

Identify cloud-ready candidate applications, databases, storage, and physical and virtual servers. Compile SLAs, dependencies, user needs and compliance and security requirements. Prepare contingency and roll-back plans to answer worst-case scenarios or undo the migration if needed. Assemble resources for migration and ongoing maintenance. Document a preliminary migration path for candidate elements.
 

Step 3: Pilot migration

Resolve issues that surfaced during design. Finalize the migration team, including both technical and business members, and establish activities around communication and training. Schedule a pilot migration, based on your design, in a non-production environment and address any issues that arise during the pilot. From this dry run, develop a runbook that documents the process (pre-migration requisites, production environment requirements, post-migration testing protocols and rules establishing when to cut over and when to go live).
 

Step 4: Migrate

Schedule the migration. Consider scheduling during a slow period, like on the weekends, overnight, during holiday breaks or other time that will generate the least impact should there be an outage or performance issue during the migration. Follow your runbook to execute the migration. After you cut over and go live, perform a post-migration validation of applications, data and network. Troubleshoot failures, service outages, data anomalies or performance hiccups. If you are unable to quickly remediate major issues, minimize impact by instituting a rollback to investigate what went wrong and re-schedule the migration.

 

What are some of the challenges of cloud migration?

Because every migration is different, each business will face its own set of challenges. Below are common challenges that businesses face during cloud migration projects.

 

  • Cultural shift: Cloud migration will bring new efficiencies and processes around the organization. These new efficiencies will have accompanying processes that users will have to adapt to. Transforming the business means re-acquainting the organization with new business processes and methods that might be met with resistance. Ensure that users understand the changes that will result from the cloud migration, plan for adequate training and expect a learning curve.
     
  • Legacy assets: Traditional IT assets, real estate and data center commitments will need to be decommissioned. As you plan your migration, be sure to include any contract penalties, building leases or hardware removal costs into your plan. For legacy elements that cannot be fully migrated to the cloud, consider hybrid environments that allow you to keep the core system and supplement storage and compute with the cloud.
     
  • Skills gaps: Managing a cloud environment requires different skills than managing on-premises or data center environments. Ideally, you’ll want to reskill your existing team. If this isn’t feasible, you’ll need to recruit resources or reach out to a managed services provider to oversee the day-to-day operations. Also, ensure that users are prepared with new skills that may be needed to perform their jobs using new tools and processes.
     
  • Disruption: Even with the best planning, things can go wrong with migration. Common pitfalls include data loss, outages or performance degradation. Plan the move during low-traffic business periods. Communicate with the business to inform users of the migration window and the possibility of unexpected behavior.

 

Why migrate to the cloud with Rackspace Technology?

As you take steps on your cloud migration journey, partner with experts who know how to create successful migrations, across diverse industries and geographies. With over 20 years of cloud expertise, Rackspace Technology can help you not just migrate, but also transform your IT operations and improve your business processes.

We’re with you through every stage — from selecting which workloads to migrate, to ongoing management and optimization on your new cloud platform. We get to know you and your organization’s unique goals and challenges, and work with you to build a roadmap leading to your optimal business outcomes. Put our cloud migration expertise to work for you.

 

Adopt the right cloud migration strategy for your business.